Gender effects in university mathematics education:

an exploratory study

Elizabeth Jean Pascoe Fraser

THESIS SUBMITTED FOR THE DEGREE OF MASTER OF PHILOSOPHY

UNIVERSITY OF EDINBURGH

1992
 Acknowledgements

I would like to express my thanks to all those who contributed to this work:

to Geoff Cohen and Sheila Cormack, for trying to steer me between the red
herrings.
to John Sean, who made it all possible.
to the Christian Salvesen Trust and the Scottish Education Department, for
funding this project.
to a certain member of the Physics Department, for demonstrating just how
rational and objective solid-state physics can be.
and to all the students who participated in the surveys, for obvious reasons.
Except where otherwise stated, the research undertaken in this thesis
was the unaided work of the author. Where the work was done in
collaboration with others, a notable contribution was made by the
author.
 Abstract

This thesis is an account of work carried out at Edinburgh University on the nature
and extent of possible differences between women and men in university
mathematics education. In order to ascertain the situation at Edinburgh University, a
series of small-scale attitude surveys were carried out. These surveys involved
mathematics undergraduates and mathematically students who were not taking a
mathematics degree.

The aim of the study was two-fold. I wished to see whether some widely-reported
differences in attitude between the sexes were replicated in the chosen samples. I
also wished to examine the attitudes of the students towards their courses to see
whether mathematics students showed the same motivation regarding their degree
subject as did non-mathematics students.

In conjunction with the surveys, data describing more wide-spread patterns of

women's participation and achievement in university mathematics education were
analysed.
Not that anything I wrote about them is untrue, far from it. Yet when I
wrote, the full facts were not at my disposal. The picture I drew was a
provisional one, like the picture of a lost civilisation deduced from a
few fragmented vases, an inscribed tablet, an amulet, some human
bones, a gold smiling death mask.
From Clea, by Lawrence Durrell

I suggest that gender difference is not absolute, abstract, or

irreducible; it does not involve an essence of gender. Gender
differences, and the experience of difference, like differences among
women, are socially and psychologically created and situated. In
addition, I want to suggest a relational notion of difference.
Difference and gender difference do not exist as things in themselves;
they are created relationally, that is, in relationship. We cannot
understand difference apart from this relational construction.

From Feminism and psychoanalytic theory, by Nancy Chodorow

 Table of Contents

1 Introduction
1.1 Overview 1
1.1.1 Mathematics education 2
1.1.2 Biology and destiny 4
1.1.3 Girls in the science and mathematics classroom 6
1. 1.4 Women in society 9
1. 1.5 Women in science 13
1.2 The present study 21
1.2.1 How it started: a different picture 21
1.2.2 The Edinburgh University surveys 21
1.2.3 Methodology 23

2 Primary and secondary mathematics education 26

2.1 Primary school 26

2. 1.1 England and Wales 27
2.1.2 Scotland 27
2.1.3 Comments 28
2.2 Secondary school 29
2.2.1 The APU surveys: findings on attitude 29
2.2.2 National patterns for participation and achievement at the
public examinations 30
2.2.3 Comments 33
2.3 The Edinburgh surveys: school experience 35
2.3.1 Gender differences in attitude to mathematics at school 35
2.3.2 Gender differences in the perception of encouragement 37
2.3.3 Possible question order effects 37
2.3.4 Reasons given for studying mathematics at school 38
2.3.5 Conclusions 42

3 Choice of degree and university 43

3.1 The Edinburgh surveys: choice of degree subject 44

3.1.1 Surveys 1 and 2: questions on choice of degree 44
3.1.2 Survey 3: questions on choice of degree 46
3.1.3 Patterns of response 47
3.1.4 Survey 1: reasons for choice of degree 49
3.1.5 Survey 3: reasons for choice of degree 49
3.1.6 Survey 2: non-mathematics students' reasons for degree
choice 5.1
3.1.7 Comparison between mathematics and non-mathematics
students 53
3.1.8 Conclusions 55
3.2 Choice of university 56
3.2.1 The Edinburgh surveys: choice of university 56
3.2.2 Entry to Mathematics degree courses in Scotland and England
58
3.2.3 Conclusions 61
3.3 Summary 62
4 Attitude and performance at university 64

4.1 Attitude towards degree subject 65

4.1.1 Background information 66
4.1.2 Attitude of mathematics students towards their course 68
4.1.3 Comparison between mathematics and non-mathematics
students 70
4.1.4 Attitude of mathematics students towards the components of
their mathematics courses 72
4.1.5 Non-mathematics students' attitude towards mathematics at
university 76
4.1.6 Conclusions 77
4.2 Mathematics as a 'masculine' subject 78
4.2.1 Perception of the non-mathematics students 79
4.2.2 Perception of the mathematics students 80
4.2.3 Conclusions 81!
4.3 The university learning experience 82
4.3.1 Affiliation and the personal dimensions of attitude 84
4.3.2 Attribution and motivation 87
4.3.3 Learning styles 90
4.3.4 Conclusions 92
4.4 Achievement and attitude at Edinburgh University 94
4.4.1 Achievement of the mathematics students 94
4.4.2 Achievement and attitude: Survey 3 97
4.4.3 Conclusions 101
4.5 Performance at Scottish and English universities 103
4.6 Career aspirations among the Edinburgh mathematics students 109
4.7 Summary 112

S Conclusion 114

5.1 Confidence 114

5.2 Motivation 115
5.3 Mathematics students' attitudes towards their course 117
5.4 Achievement and participation in university mathematics 117
5.5 Synthesis and outlook 118
References 120

I Survey 1 . 125

1.1 Methodology 125

1.11 Results 127
1.111 Questionnaire 1 134

Il Survey 2 138

II.! Methodology 138

11.11 Results 140
11.111 Questionnaire 2 145

Ill Survey 3 148

111.1 Methodology 148

111.11 Results 149
 ffl.11I Questionnaire 3 160

IV Entrants to Mathematics degrees 168

V Degree results for Scotland and England 170

VI Published papers 171

 Chapter 1.
Introduction

In the past two decades or so there has been an upsurge of interest in the subject of
gender and science education. The sciences would appear to be of increasing
importance to employment, and in this context many researchers have expressed
concern about the position of women in science and mathematics. There are two
facets involved here: one is that girls and women are underrepresented in these
subjects (apart from biology), the other is that there are achievement differences
favouring boys and men (although in the latter case there is some debate about the
extent and importance of such differences). The actual figures on participation and
achievement patterns are presented and discussed at greater length in Chapter 2. In
this chapter I will describe some of the previous work in the field of gender and
science (Section 1.1) and the context of the present study (Section 1.2). While my
current research project is specifically concerned with mathematics, many of the
findings regarding gender effects in other sciences apply equally to mathematics.
Therefore I will also consider examples of research done in the field of the physical
sciences. I have deliberately excluded studies on gender differences in the fields of
engineering and computing in the interests of conciseness and clarity. There have
been many interesting publications in these fields (Rothschild 1983; Siann et al.
1988; Durndell et al. 1990), but it was felt that the interests of this study would be
better served by restricting the areas of review to those more directly relevant to the
subject under consideration.

The issues involved are exceedingly complex and the following discussion only
aspires to present the bare bones of the 'problem' of women in mathematics. I use
the word 'problem' with some caution because many of the implications and
assumptions present in its use are themselves part of more widespread, pervasive and
problematic ideology systems involving women and society as a whole.

1.1. Overview

Much of the research on gender effects in mathematics has been done at primary and
secondary school level and I will begin with a brief description of the current

II
 2

situation of girls in school mathematics and the problems of interpretation involved

in Section 1.1.1. Further details of some of the large-scale studies can be found in
Chapter 2. Section 1. 1.2 deals with some of the more traditional theories concerning
gender differences in mathematics. These involve biological and cognitive aspects of
perceived mathematical ability. For various reasons, which will be elaborated, the
arguments above are presented solely in the interests of completeness. The focus of
this study is on the potential sociological and psychological factors involved in girls'
and women's experience of mathematics and these are the bsis for the subsequent
discussion. Section 1.1.3 examines some of the findings on girls' experiences in the
science and mathematics classroom, and these are discussed in the light of differing
social expectations for females and males in Section 1.1.4. Section 1.1.4 also
examines some of the findings of studies on psychological differences between the
sexes which could affect their mathematics education, and looks at the situation of
women in higher education. This unholy mix is made necessary by the fact that the
ideas involved are conceptually intertwined and intimately related to social
definitions of the abstract notions of 'femininity' and 'masculinity'. Section 1.1.5
attempts to relate some of the issues discussed in the previous section to the
particular context of women's involvement in science and mathematics (or lack of it,
as the case may be). At this point, the question is no longer one concerning the
whats and hows of the situation, but one of the whys. These are possibly the most
contentious considerations as well as the most fundamental to any study of gender
differences. In order to provide an insight into the whys of the matter, I will give an
account of some of the models of psychological development which might explain
certain aspects of such differences in mathematics.

1.1.1. Mathematics education

The extent of girls' participation in mathematics at secondiry school and the

performance patterns observed are described in some detail in Section 2.2, and the
data on women's participation in university mathematics are presented in Section
3.2. Therefore I will merely give a general indication of the current situation at this
point in order to set the discussion of previous research findings in context.

Presently,girls studying mathematics at Certifibate of Secondary Education (CSE) A

level comprise approximately one third of the total number of A level Mathematics
entrants. The position in Scotland is less extreme since girls make up about one half
 3

of Higher Mathematics entrants. The differences between Scotland and England

regarding girls' participation are discussed at greater length in Chapter 2.

Performance patterns at the A level and Scottish Certificate of Education (SCE)

Higher examinations show some differences in favour of the boys. A higher
proportion obtain grade A, but the pass rates for grades A-C (Higher) and A-E (A
level) are virtually identical. Other studies have found similar patterns in
mathematical performance as well as differences in attitude, with girls finding
mathematics more difficult and less interesting (Assessment of Performance Unit
1985).

Not surprisingly, women are also underrepresented in university level mathematics

education. Again, there are differences between Scotland and England in this respect
with women constituting 40% or so of the Scottish universities' intake to
Mathematics degrees, compared with 30% for the English universities (Section
3.2.2). Differences in achievement at this level, as measured by degree results, are
very small (Section 4.5).

One might remark that the phenomenon of women's underrepresentation in

mathematics would appear to be due to differences in achievement in the top
performance band, which could be interpreted as signs of differing mathematical
ability between the sexes. It is certainly true that there is some similarity between the
proportion of girls or women in the top achievement band at one level and their
representation at the next. Differences which might otherwise be regarded as small in
absolute terms (the percentage of each sex obtaining grade A) become important in
relative terms (the proportion of mathematics students at a certain level who are
female).

However, attempts to explain differences in mathematical achievement by referring

to innate biological differences in actual ability have not proved entirely satisfactory.
For instance, Scottish girls did marginally better than the boys in the Certificate of
Sixth Year Studies Calculus paper (Section 2.2.2), which would imply that there are
other factors involved apart from immutable biological ones. There is a considerable
overlap in mathematical performance between the sexes (APU 1985), and it is
doubtful whether the magnitude of any average difference is sufficient to account for
women's relatively low participation rates. International studies of mathematical
achievement provide further evidence of a large overlap in performance, with girls
from some countries outperforming boys in others (H'usen 1967).
One might also question what it is that is being measured in a mathematical test. Is it
'really' mathematical ability or are there confounding factors such as shifting
perceptions of what constitutes 'real' ability and how it is manifested? This idea and
some of its implications will be further examined in the following sections. But in
the next section, I would like to briefly present a couple of the arguments which have
been advanced explaining observed gender differences in mathematics in terms of
biological differences.

1.1.2. Biology and destiny

The underlying ideology of arguments evolving around biological differences is the

concept of two 'natures', one for women and one for men. In terms of the tests used
to assess cognitive styles, women's 'nature' is verbal while men's is numerical and
spatial. That is to say, from adolescence on, boys and men have higher scores on
tests purporting to measure quantitative and space perception ability, and girls and
women score higher on tests assessing verbal skills (Maccoby & Jacklin 1974).

Spatial visualisation is one area where measures show consistent differences in

favour of males (Maccoby & Jacklin 1974; Sherman 1975). Tests used include
variations on the Rod-and Frame test where the subjects are asked to judge whether a
rod in a tilted frame is tilted or not. Some other tests aim to assess the subjects'
ability to visualise rotating solids. The interest of spatial visualisation is that it is
sometimes considered to be associated with mathematical ability. The observed sex
differences in the former domain have been used to explain girls' relative
underachievement in mathematics and science. Some researchers have suggested
that spatial visualisation ability might be a genetically determined characteristic and
postulated that the relevant gene might be X-linked and recessive. This would imply
that boys would be more likely to acquire the characteristic since they would only
need to inherit the gene from their mothers. On the other hand, girls would require
two copies of the gene, one from each parent, for the characteristic to be expressed.

Another theory concerned with biological reasons for differences in mathematical

achievement postulates differences in cerebral dominance (Maccoby & Jacklin
1974). In this case, the basic idea is that the areas of the brain concerned with verbal
functions become more developed in women, to the detriment of those involving
spatial perception. These two types of functions tend to be localised in different
hemispheres, the former in the left and the latter in the right. There are several
hypotheses using the idea of brain lateralisation (the extent to which one hemisphere
dominates the functioning of the other) and it is beyond the scope of this study to
deal with them adequately. The reader will find a more detailed review in Maccoby
and Jacklin. However, in view of conflicting evidence on this particular subject,
all that is clear is how little is known either about how the brain works or about
any reliable sex differences in its functioning. It is equally clear that, despite the
inconsistencies, biases and ambiguities of the data, the project of all the
research to date had been precisely to seek out sexual differences. Without an
ideology of sexual difference we could never have imagined the supposed sex
differences in the brain in the first place. (Segal 1990,p.63)

In addition, Maccoby and Jacklin point out that it is not always obvious exactly what
aspect of cognitive functioning is being measured and Sherman claims that the
average difference in such measures is usually quite small. There is also the problem
of the cultural and social context of mathematics education, which are not adequately
taken into account by arguments such as those mentioned above. Girls and boys
appear to show different levels of interest in mathematics at school (APU 1985), and
one would imagine that such differences affect performance as well as participation
rates. This idea is explored further in the following sections. However, as far as
possible biological influences go, it is practically impossible to accurately judge the
extent of their impact on an individual's mathematical ability in the absence of other
factors.

Hyde (1981) argues that under 5% of the total variance in a population is due to sex
and therefore differences in spatial perception abilities alone cannot account for the
small proportion of women in fields were these may be important (in some branches
of engineering, for example), even considering the differing proportions of females
and males amongst the highest scorers. The link between spatial visualisation and
mathematical achievement is somewhat tenuous, especially considering the decline
of geometry in the secondary school curriculum. There are also different ways of
approaching problems in mathematics, the most obvious alternative methods for
many problems at school level being algebra and geometry. It might be interesting
to note that Skemp (1986) associates algebraic reasoning with the verbal aspects of
cognitive functions, as opposed to the visual. One might therefore argue that
mathematics incorporates both aspects of cognition and that the observed superiority
of one sex in one area does not adequately explain differences in mathematical
performance and participation. There is some evidence that spatial visualisation
improves with practice, and the type of play associated more with boys than girls
might contribute to the development of their spatial perception (Sherman 1975;
 311

Ben-Chaim et al. 1988).

The idea that social and cultural forces, such as those which influence what is
deemed appropriate and 'normal' play activity for each sex, have an impact on
seemingly fundamental processes like cognition is an important one. In the following
sections, I will therefore examine how some of these forces have been seen to affect
girls' and women's positioning in mathematics with respect to boys and men.

1.1.3. Girls in the science and mathematics classroom

There have been some studies which suggest that girls may be marginalised in
science and mathematics classes. Kelly (1987) gives some vivid examples of how
gender is 're-contextualised' in the science class situation at secondary school.
Practical experiments can give boys the opportunity to exert control over apparatus,
reinforcing the stereotype of boys being 'tougher' than girls, more able to deal with
potentially dangerous chemicals and machinery. Teacher attention is also a resource
which can be unevenly distributed between the sexes (Crossman 1987). In these
cases, science is, in a sense appropriated by the boys, confirming its image as a
masculine domain. This idea is further developed in the next sections. Thus the
ideological baggage in terms of gender identity which pupils bring into the
classroom, has the potential to adversely affect girls' inclination and opportunity to
participate in the sciences.

In the context of boys monopolising resources to the detriment of girls' education,

there is an ongoing debate concerning the benefits of single-sex science and
mathematics classes for girls. The results of comparisons between single-sex and
mixed schools are not clear-cut (Hhsen 1967; Dale 1974). But there have been some
intervention strategies involving putting girls into single-sex classes for mathematics
(Eales 1986) and running courses and conferences on mathematics and science for
girls (Burton & Townsend 1986; Smail 1987) which have shown positive effects.

The obvious comment here is that it is maybe the extra attention involved which has
the positive effects rather than the absence of boys per se. There is also the
influence of obvious positive expectations concerning girls in mathematics and
science which such strategies generate. After all, if the organisers did not believe in
the girls' abilities, then they would not have made the effort in the first place.
 7

Another example of the marginalisation process is illustrated by Walden and

Walkerdine (1986) in their account of interactions in a mathematics classroom. They
present a detailed analysis of some of the social relations involved in girls'
mathematical experience at school and suggest that
there is a particular combination of classroom practices and and an
understanding of mathematical learning which produces failure in girls, and that
in consequence girls are being put in the position of being successful but not
succeeding (Walden & Walkerdine 1986, p.124).

That is to say, even when girls are successful in mathematics, this is not always seen
as indicating a 'teal' understanding of the subject. They have 'failed' in the sense
that they have not convinced the teacher of their mathematical ability.

This 'failure' in interpreted in the context of an incompatibility between what is

considered appropriate and normal behaviour for girls and behaviour indicating
'real' understanding. The conflict is between the concepts of rule-following and
rule-challenging, with the latter being seen as a sign of mathematical 'flair'. On the
other hand, girls' perceived tendency to be better behaved in the classroom may be
taken to indicate passivity and rule-following behaviour, and thus a lack of 'natural'
ability. It might be of interest to note that what teachers sometimes characterise as
girls' 'problems' (rule-following rather than rule-challenging for instance) are also
considered characteristic of low-achievers generally (Trickett & Suilcie 1988).

One implication of the above observation is that some teachers appear to be

considering girls as low-achievers in mathematics almost by definition: what was a
low-achiever's problem has now become a 'girls' problem'. Such a redefining of the
situation is hardly surprising in the context of Walkerdine's argument that science
and mathematics were historically defined as male preserves, as were the qualities of
reason which the disciplines demanded (Walkerdine 1989). This idea is further
discussed in Section 1.1.5.

The notion that teachers sometimes interpret a pupil's behaviour differently

according to the sex of the pupil is supported by several studies. Spender's account
of an exercise involving a class of PGCE students is particularly worrying
considering that the class was taking a course on sex discrimination in education
(Spender 1984). The students were given identical 'report cards' to assess; one half
had 'Jane Smith' written on them and the other half 'John Smith'. Though the class
only consisted of ten students, the differences between the assessments of 'Jane' and
'John' indicated fairly clearly that the students considered the characters attributed to
 I
'John' as signs of ability. Comments from the students assessing 'Jane' included the
suggestions that she would make a good secretary and that the probably would not
want to stay on at school; 'John' would do well in the Civil Service.

Spear (1987) obtained similar results with science teachers' evaluations of samples
of work attributed to 11 year-olds. 'Boys' were consistently rated higher for
'richness of ideas', 'interest' and '0 level suitability'.

Of course studies of this type are somewhat artificial and can only give a limited
picture of how individuals are assessed and the criteria used in the evaluation.
Certainly they imply that gender affects the assessment process in some way and in
some situations, particularly in the absence of other influences such as personal
knowledge of the individual concerned. They also suggest that there is probably no
'true' scale of gender difference, since in some situations the assessment may be
biased (as can be seen from the above studies) and most large-scale studies show
small overall differences (Hyde 1981; APU 1985). Walkerdine (1989), for one,
argues that there is a complex play of situations and individuals which combine to
mitigate or exacerbate the essentially negative effects of being female in a 'man's
world'. It is this aspect which I wish to examine next. The term 'man's world' may
appear extreme, and it is certainly not a term I would use without qualification.
Nonetheless, it illustrates many of the social and psychological processes potentially
involved in determining women's and men's positions relative to each other,
particularly regarding the way knowledge and learning are constructed and validated.

Before I continue, I would like to clarify some of the assumptions I am working

with. The first one is that knowledge, and by implication the learning experience at
all levels, is fundamentally a social construct. This idea, which will be elaborated on
in the next two sections, is implicit in accounts of the history and philosophy of
science (Klemke et al. 1980; Kline 1980; Davis & Hersh 1981) and made explicit in
much feminist writing which argues that the problems women encounter in being
equally represented in various non-traditional spheres are due to their merits being
judged on the basis of their gender: the operating framework would appear to be that
what men do is considered somewhat better and more important than what women
do (Mead 1962; Kolodny 1981;Spender 1986). Therefore women are marginalised
initially almost by definition.

Although this concept appears simplistic and it is beyond the scope of this study to
explore all the arguments and implications involved, it brings up the principal point
of this introduction: what are the possible explanations for the perceived
marginalisation of women in many spheres of economic and political importance?
The words 'perceived' and 'political' require qualification; 'perceived' because there
are indications that women and men are becoming more similar in attitudes and
aspirations than they appeared to be in the past, and I will enlarge on this later. The
issue remains that if women are seen or see themselves as marginalised, this itself
has negative implications on women's sense of participation in various
male-dominated spheres. 'Political' is used in the sense of power, of policy-making
opportunity, whether it be in an academic context, deciding who gets tenure or what
constitutes 'valid' research, or the more personal issue of which parent interrupts
their career to look after the children.

While this line of thought might seem to be straying away from the topic of my
thesis, it is important not to isolate the debate concerning women in mathematics
from the underlying context of women's relationships to society, knowledge and
science.

1.1.4. Women in society

We saw in the previous section how certain stereotypes regarding girls' behaviour
and aptitudes can affect the perception of their achievement, both actual and
potential. Here I will examine in more detail the extent and nature of such
stereotypes, as well as the strength of the evidence for them.

Amongst the received ideas regarding girls and women are notions that they have
less confidence in themselves, are more motivated by affiliative needs and are not
particularly interested in science and mathematics. All of these notions have their
place in the literature on gender effects in mathematics. Confidence is seen as
necessary for rule-challenging, and the need to relate to people is considered
detrimental to the development of the kind of independent learning patterns which
mathematics is thought to require (Hoffman 1975). While there have been a variety
of studies examining such differences between the sexes, the results have been
ambiguous at best, and contradictory at worst (Maccoby & Jacklin 1974). Of
concern to this study however is the observation that Maccoby and Jacklin
concluded that one of the four areas where there were 'well-established' (not to be
confused with 'large' or 'important') differences was mathematical achievement.
Boys also scored higher on measures of spatial visualisation (as seen above) and
 10

aggression. The other area was verbal ability, which I have already mentioned.
Amongst the unsupported beliefs were girls being more 'social', suggestible, having
lower self-esteem and achievement motivation.

Some researchers have suggested that where girls' or women's attitudes and
behaviour do conform to stereotype, the context is often male-appropriate and it is
therefore unfounded to generalise such traits (Stein & Bailey 1975). Again there is
some question of what is actually being measured in experiments purporting to show
gender differences. A case in point is that of the Assessment of Performance Unit
surveys which found evidence of differences in confidence levels between girls and
boys regarding mathematical ability (APU 1985: the results of these surveys are
discussed at greater length in Section 2.1.1). Here, confidence was assessed by how
difficult the subjects considered an exercise. One might argue that the observation
that boys rated exercises as easier more often than the girls reflects differences in the
expression of bravado, since the findings also showed that boys overestimated their
performance on written tests. Girls showed a stronger tendency to underestimate
their test results. In this case, conservatism seems to have been interpreted as relative
lack of confidence. One might argue that such observations could also imply
differences in interest levels or perceptions of the importance of the task.

Another point made by Maccoby and Jacklin is that many of the classic studies of
psychological gender differences were carried out with white middle-class American
college students, which again poses the problem of generalisation. There is some
indication that black women show different attitude and behavioural patterns to those
found amongst the previous samples (Safilios-Rothschild 1986). Walkerdine (1989)
also found indications of differences in parental attitudes and expectations between
working- and middle-class parents. It would therefore appear that the results of
psychological studies cannot be interpreted without considering sociological, and by
implication, historical, developments.

One must also consider the question of the publication and dissemination of research
results. Maccoby and Jacklin have pointed out that findings showing little
differences or unexpected ones can be deemed unworthy of publication. The picture
we obtain from published research is therefore biased by virtue of the academic
selection procedure. This includes self-selection by the researcher, who chooses
what s/he considers interesting or useful information, as well as the refereeing
process for publication in academic journals. Spender (1981) comments on the
 11

problems of the criteria used to determine the inclusion of publications in the British
Research Index, which result in references to 'marginal' subjects being rarely
included.

The statistical techniques used in the analysis of research findings influence to some
extent what is seen as 'valid' research and what is not: the need to be seen as
objective compels the researcher to test any differences observed for statistical
significance. If the differences are not significant, then the findings are often not
considered conclusive enough to warrant subjecting them to outside scrutiny. One
might speculate that if there were accepted significance tests for 'non-differences',
the picture of gender differences would be a very different one, giving much more
weight to data implying similarities between the sexes than is currently the case.

However, there are indications that research methods are shifting their focus. There
is now a debate in the field of gender studies in education concerning the relative
importance of statistical and educational significance. The criticism levelled is that
assessing the importance of differences in performance by statistical significance is
misleading, since for large samples very small differences can be statistically
significant (Walkerdine 1989). The meta-analysis of sex differences in measures of
quantitative ability for large-scale surveys indicates that such surveys tend to show
rather small differences (Hyde 1981).

The importance of the debate concerning statistical and educational significance

becomes more obvious when one considers that many people do not have access to
detailed research reports or are not inclined to make the effort. Therefore results of a
study on, say, sex differences in mathematical performance reported in the media as
'significant' can be interpreted as educationally significant. This type of unqualified
assumption reinforces the cultural myth that women are in some sense 'inferior' in
the domain considered. The problem lies in assuming that the samples are
homogeneous when in fact they are generally not, particularly for large samples.
Therefore differences attributed to one variable (sex, for instance) might be
influenced by other variables (such as race, class, 'ability' or interest levels) which
are not controlled for.

As we have seen, there appears to be a certain amount of controversy as to the extent

of any fundamental differences which might explain the observed imbalance of
power between the sexes. In the sphere of higher education, since this is one focus of
my study, the imbalance is shown by the underrepresentation of women amongst
 12

academic staff. In 1980, 40% of undergraduates were women, while only 14% of
full-time university staff were (Acker 1984). The historical aspects of women's
participation in higher education cannot be ignored. In the late 19th century, women
were advised against sustained mental activity on medical grounds: it was feared that
such activity could have detrimental effects, particularly on the reproductive organs
(Burstyn 1984). Again, there is the assumption that women and men move in
different spheres, that what is 'natural' for one is not so for the other.

The imbalance observed in higher education highlights Spender's argument that

education is a man's world: the definitions and ethos are male, and are subject to
validation by men (Spender 1981). This historical development is the 'outcome of
taking the power to determine the parameters of education' (p.161). The implication
is that men control the generation and dissemination of knowledge and that women
have therefore been excluded from the construction of knowledge. Although there
are many women in education, they are not at the locus of power.

I mentioned before that there seem to be indications that the imbalance is becoming
less extreme. Shifts in business practices mean that now there is a greater emphasis
on communication and negotiating skills, and the 'killer instinct' is no longer of
primary importance. The former skills have traditionally been associated with
women, and one might speculate that this would advantage modern women's career
prospects. Bumhill and McPherson's study of attitudes and aspirations amongst
academically well-qualified Scottish school-leaven showed that the occupational
intentions of women and men in 1981 were much more similar than they had been
10 years previously (Burnhill & McPherson 1984). In addition, there is some
evidence that observed gender differences in attainment in mathematics are
diminishing over time (Willms & Kerr 1987); and female participation in
mathematics at school has increased in England during the 80's (Cohen & Fraser
1992, in Appendix 6). There has also been a steady growth in the numbers of
women participating in higher education in Britain (Weinreich-Haste 1984;
Universities' Statistical Record 1987).

On the other hand, particularly at the higher levels, science education remains
male-dominated (Kelly 1976; Acker 1984), though perhaps to a lesser extent than
previously. The next section examines some of the aspects of women's rather
ambivalent relationship with science and science education, which might account for
such participation patterns.
 13

1.1.5. Women in science

We have seen in the preceding section how women historically have not been
included in educational policy-making and, by implication, have been excluded from
the construction of knowledge. I would like to examine the latter idea in this section
in order to see how the construction of scientific knowledge has led to a concept of
science which seems to appeal more to boys and men. I have mentioned that one of
my working assumptions is that knowledge is a social construct. I would also like to
clarify this idea in the particular context of scientific knowledge.

The poptilar image of science is one of rationality, objectivity, brilliant men in white
coats discovering fundamental truths about the universe and contributing to the
well-being of humankind. That of course is one facet of science. The other image is
the somewhat monstrous one of a monolithic impersonal, though still rational and
objective, structure within which socially irresponsible scientists conduct research
for the benefit of those who fund them, sometimes to the detriment of their fellow
human beings. The second view articulates the negative implications of the
perceived dislocation of science from society and the idea that science and scientific
'advancement' are in themselves positive things which need not be affected by social
considerations. Rationality and objectivity provide the constructors of science with
immunity from the negative consequences of their science. However, this also is
part of the monolithic interpretation of science, a view which I would like to dispute.

A study of the history of science leads to a somewhat different picture. Science was
very clearly developed by individuals who, like all individuals, had their own
assumptions and ideas about what the world was like.

Science was, and still is, the product of its time and of the individuals who lived in it.
There would therefore appear to be a fundamental dislocation between the platonic
ideal of science and the more immediate articulations of what might currently be
considered scientific truths.

The platonic ideal is that of the One Truth, the idea that this Truth exists
independently and we discover it. What actually occurs is that what we 'discover' is
affected to a greater or lesser extent by what we believe or want to believe about
how things operate. The illusion of the One Truth is maintained by a process of
consensus within the field of research. However the process is not perfect in the
 izI

sense that them is often dissent amongst the participants, leading to the formations of
various schools of thought. Where is the One Truth at this point? In a sense, the
answer is history. Many scientists today would not classify their theories as true or
untrue, but as how good they are as working models of the phenomena under study.
The idea of models as opposed to truths was made explicit in the field of
mathematical logic (Davis & Hersh 1981), in particular within the formalist school
which advocated that the ultimate meaning did not matter as long as the
mathematical symbols were manipulated correctly. 'Truth' was seen as provability
within the mathematical system (Davis & Hersh 1986). Godel's theorems can be
seen as influencial in effecting a change in the way mathematicians perceived the
concepts of proof and therefore truth. Since not all theorems could be proved or
disproved within a given axiomatic system (Godel's Incompleteness Theorem), they
assumed a somewhat ambiguous status: they were no longer strictly true or untrue in
the platonic sense, only within the context of a particular framework of axioms
(Davis & Hersh 1981).

There are many examples of such crises of belief in the history of mathematics and
their effects on the mathematicians of the period. Russell, who at first was convinced
that 'the principles of logic and the objects of mathematical knowledge exist
independently of any mind and are merely perceived by the mind' (Kline 1980,
p.218), spent many years trying to develop a thorough and consistent axiomatisation
of mathematical logic in order to provide a more solid foundation for mathematics.
In his Portraits from Memory (1958) he states
I wanted certainty in the kind of way in which people want religious faith. I
thought that certainty is more likely to be found in mathematics than elsewhere.
But I discovered that many mathematical demonstrations, which my teachers
expected me to accept, were lull of fallacies, and that, if certainty were indeed
discoverable in mathematics, it would be in a new field of mathematics, with
more solid foundations than those that had hitherto been thought secure. But as
the work proceeded, I was continually reminded of the fable about the elephant
and the tortoise. Having constructed an elephant upon which the mathematical
world could rest, I found the elephant tottering, and proceeded to construct a
tortoise to keep the elephant from falling. But the tortoise was no more secure
than the elephant, and after some twenty years of very arduous toil, I came to
the conclusion that there was nothing more I could do in the way of making
mathematical knowledge indubitable. (quoted in Kline 1980, pp.229-230)

By looking at the history of scientific development from a somewhat anthropological

point of view, one can validate the argument that scientific knowledge is a social
construct. For if it is not the science of the platonic ideal, waiting to be discovered,
then it is invented by people within a social framework. The historical narrative also
shows the human side of science, the dreams, ideals and despairs of the individuals
 15

involved. Why then is science seen as impersonal and therefore off-putting to girls
and women? (Kelly 1987a) One explanation is that
school science portrays a picture of both the positivist and reductionist tradition
of scientific methodology.[...] Positivist science portrays theories as logically
ordered sets of laws which explain reality. Arguably, such theories are
examples of masculine logic and explain the reality of masculine science.
(Bentley & Watts 1987, p.93)

The comment is specific to physical science, but the same might be said of the way
mathematics is traditionally taught. Kelly points out that while 'science and society'
courses have been developed, they have not had much impact in the early stages of
secondary education before pupils make their option choices.

I will return here to the quotation from Bentley and Watts cited above, since the idea
of science as a masculine subject has become one of the main tools in explaining the
dynamic processes which result in women's underrepresentation in the sciences. I
use the term 'dynamic' because social forces are involved in the shaping of the
situation and these are fluid and shifting as society changes.

The argument regarding the masculine nature of science is roughly the same as
Spender's contention that education is fundamentally masculine: that the
policy-makers and power-holders are men and it is in their interests to keep it that
way. This does not necessarily imply the existence of a male conspiracy to keep
women out, merely the presence of social inertia which contrives to maintain the
status quo. Easlea (1981) and Walkerdine (1989) take the argument further in the
specific context of science. They argue that men assured themselves the position of
power by defining science historically as something which did not possess feminine
characteristics: it was logical, rational and objective, all considered non-feminine
attributes. Easlea gives various examples from history which indicate that science
was seen as a powerful controlling force, a means of dominating nature (often
allegorised as female). This is the classical vision of science, one which does not
accept contradictions or questioning outside the established framework. In other
words, one can validly question the 'objective' truth of a scientific theory but not its
ideological value (whether it is 'good' or 'bad', or why it is of interest), since the
latter is a subjective consideration, one which is argued by appeals to emotion rather
than logic. However, history has shown us how definitions of what is logical and
rational can change: it is no longer necessary for Earth to be flat or be the centre of
the universe.
 16

The science and mathematics which is taught in schools usually adopts the classical
approach and this flavours children's early perceptions and impressions of science.
There have been curriculum developments in mathematics which have attempted to
emphasise the relevance and humanity of the subject, and make it generally more
interesting. However, this type of mathematics demands more teacher involvement
and interest to be effective, scarce resources considering the current disaffected state
of the profession.

It can be argued that humanising mathematics and science in schools would benefit
girls particularly since it presents an alternative to the rigid uncompromising
approach characteristic of the traditional 'masculine' view of mathematics (Eales
1986; Isaacson 1986; Head 1987). Emphasising the variety and richness of the
mathematical experience would encourage children's aesthetic appreciation Of the
subject and possibly result in an enrichment of mathematics, since it would attract
people who were genuinely interested in it, as opposed to those influenced by the
perceived prestige value of a difficult subject or by the promise of objectivity and
clear-cut answers to questions.

There has been some research carried out which implies that the proportion of
relatively immature pupils is higher for boys choosing science than it is for girls
doing so, and the former seem to have more rigid views on many issues compared to
boys who choose other subjects (Head 1987). Such findings suggest an alternative
interpretation of studies showing differences in confidence levels, since bravado is
often associated with immaturity and, as I have mentioned before, might be
interpreted as a display of confidence. Head's observations are supported by
teachers' and pupils' impressions of the classroom behaviour of girls and boys
(Bales 1986; Scott-Hodgetts 1986; Kelly 1987b). Head proposes a model of
psychological development which predicts that as long as science is taught as a rigid
subject which offers unambiguous answers to problems and is considered to be a
masculine subject, it will attract a large number of boys with rigid authoritarian
attitudes. The image of scientists as 'authoritarian, conservative and controlled in
their thinking' (Head 1987, p.19) is thus further perpetuated.

One of the solutions Head puts forward is interesting in the context of the differences
between the English and Scottish school systems and the observed variation between
these in girls' participation in mathematics. He suggests that delaying the timing of
subject choice would result in a reduction in the number of boys choosing science
 17

because of its 'masculine' image and approach. It might also increase the number of
girls since they would no longer be at the stage of development where they might be
unduly influenced by stereotypes concerning appropriate subject choice. Certainly in
Scotland where subject choice is not so severely restricted until the age of 17, there
are proportionally more girls taking Higher Mathematics and Physics than there are
taking A level Mathematics and Physics (Smail 1987), although Kelly (1978a) has
argued that there is still a large amount of channelling taking place early on which 'is
contrary to the whole philosophy of Scottish education' (Kelly 1978a, p.69). Head
also advocates that in order to attract more girls and students with imaginative
flexible minds, 'science must be relevant to the issues which concern them. The
probable implication is that science would need to be presented in the context of the
needs of society and individuals' (Head 1987, p.23). This is an attitude adopted by
many workers in the field, and the arguments involved apply equally to the specific
case of mathematics as it does to the physical sciences.

The idea of developmental differences between the sexes is discussed in more detail
in the work of Chodorow (1978, 1989) and Gilligan (1979, 1982). Their work
provides an interesting perspective of the issues involved in the differing attitudes
between the sexes in science and mathematics education. Using the framework of
object relations theory, Chodorow suggests that the process of identity development
is somewhat different for girls and boys due to the fact that it tends to be women
who parent. To vastly simplify the argument, while both sexes identify initially with
the mother, the boy child must eventually separate from her in order to develop a
masculine gender identity. This can lead to a 'negative' definition of the masculine
identity and thus to a rather elusive sense of masculinity: since the secondary
identification with the father (often a shadowy elusive figure himself, due to
time-consuming bread-winning pursuits) is not as strong as the initial relation with
the mother, under certain conditions the masculine gender identity may be defined in
opposition to the mother and thus the feminine gender identity. In such a case, being
male is defined as being not-female, and the boy child must therefore reject the
feminine within himself. An identity developed in this manner would be less secure
than the feminine identity developed as a result of the primary identification with the
mother.

The elusiveness of the masculine identity is closely associated with the notion of
'masculine' domains. Chodorow argues that
given that masculinity is so elusive, it becomes important for masculine identity
 Lii

that certain activities are defined as masculine and superior, and that women are
believed unable to do many of the things defined as socially important. It
becomes important to think that women's economic and social contributions
cannot equal men's. The secure possession of certain realms, and the insistence
that these realms are superior to the maternal world of youth, become crucial
both to the definition of masculinity and to a particular boy's own masculine
gender identity. (Chodorow 1978, p.181)

In the context of mathematics and superior realms, Spender (1986) comments that

last century, when it was classical languages which opened career doors, it was
widely established by reputable men that girls could not do languages. But now
that the focus has shifted and languages are not the testing ground for
hierarchies and success, girls have been found to be very good at the low-status
subject of languages. It is mathematics that they find difficult now. And
presumably if next century most power is still concentrated in the hands of
men, and child-rearing is decreed as the crucial determinant for career
advancement, it will soon be demonstrated that girls have no aptitude for
child-rearing practices. (Spender 1986, p.59)

The connection between this statement and Chodorow's argument is that it provides
an illustration of men's need and observed attempts to acquire secure domains of
influence and power.

What appears to be implicit in Chodorow's model is that there exist pressures on the
boy child to develop a masculine identity rather than a personal one, thus leading to
the rejection of the primary identification with the mother. These pressures might not
affect all boys to the same extent and one would presumably observe variations in
the development and expression of the masculine identity (and the feminine one, for
that matter). Chodorow does argue very strongly that the notion of gender difference
must be seen as relational rather than absolute in some sense (Chodorow 1989).
However, her arguments articulate themes from popular consciousness, such as that
of the fragile male ego. The importance of the type of parenting in the development
of gender-linked personality traits is further supported by a study which found that
well-fathered children exhibited higher levels of spontaneous sex-role blending
(Biller 1972). The implication here is that children who have strong relationships
with both parents feel secure enough about their personal identity to be able to adopt
both 'feminine' and 'masculine' styles of behaviour comfortably.

Chodorow's theory also has some important sociological implications regarding the
power relations within our society, relations which have historically favoured men. If
both sexes were equally involved in parenting, then the traditional division of labour,
which isolates women from the cultural, economic and political power bases, would
no longer prove an obstacle to women's equal participation in the power structure. In
 ifl
P 4

particular, the negative impact of conflicting domestic and academic responsibilities

described by the women postgraduates in Taylorson's study (Taylorson 1984) would
be greatly reduced. Thus the problem of who parents, and what effects this has on
the child, is intimately linked with wider concerns of power and social policy. The
connection is made explicit in Segal's comment
[..]'masculinity' gains its meanings, its force and appeal, not just from
intemalised psychological components or roles, but from all the wider social
relations in whichmen and women participate which simply take for granted
men's authority and privileges in relation to women. (Segal 1990, p.284)

I would like to point out that the concepts of 'masculinity' and 'femininity' are not
necessarily dichotomous and most individuals incorporate characteristics associated
with the opposite sex. Indeed, androgynous behaviour is more likely to allow the
individual to cope with a wide variety of situations, ensuring that the particular
demands of a situation and the personal needs of the individuals involved are dealt
with appropriately (Kaplan & Bean 1976). The terms 'feminine' and 'masculine' are
convenient descriptions of patterns of attitude, emphasis and behaviour which are
more readily associated with one sex than the other. This is not to say that men are
not influenced by affective ties, for instance, or indeed do not desire them. They may
merely be less likely to express such attitudes and behaviour as readily or in the
same way as women. 'Masculinity' itself is an ambiguous and convoluted concept
and therefore should really be thought of as a complex diversity of masculinities
involving differences of class, race, and political and sexual orientation (Segal
1990). There is a subtle and complex relationship between what is believed and
what is perceived which is beyond the scope of this study to examine, and I can only
ask for the reader's indulgence concerning my somewhat cavalier presentation and
refer them to the original works.

Gilligan (1979, 1982) presents another dimension of Chodorow's model. Studying

reactions to stories involving ethical dilemmas, she suggests that women's and
men's moral developments exhibit different areas of emphasis: responsibility and
relationships for women, and rights and rules for men. She claims that traditional
theories of psychological development make the implicit assumption that the
masculine mode of identity development through stages of separation (beginning
with the mother) is the norm. She argues that '[...]milestones of childhood and
adolescent development are described by markers of increasing separation. Then
women's failure to separate becomes by definition a failure to develop' (Gilligan
1979, p.434). The notions of separation and attachment are linked with the
 20

dimensions of independence and dependence, rule-challenging and rule-following,

confidence and lack of confidence. The idea of something being lacking in girls'
development is implicit in much of the traditional literature on gender differences
(Walkerdine 1989). Gilligan's work makes the point that this interpretation is only
valid if one considers the masculine mode of development to be the norm, and that
aspects of both feminine and masculine developments are necessary for true maturity
(Gilligan 1982). If the assumption that the masculine mode is the norm is
challenged, the 'problem' of women and mathematics would appear to be more a
problem of the definition of mathematics and how it is taught in schools. Brown
(1986) examines Gilligan's model in the context of mathematics learning and
suggests that the dominant mode of the existing mathematics curriculum emphasises
the 'masculine' concern with rules and a legalistic reasoning process (seeing things
in terms of precedents). On the other hand, very little in the curriculum caters to the
'feminine' concerns of context-boundedness (locating the phenomenon within a
broader context) and 'people-connectedness', which could be seen as relating to
considerations of responsibility. Again in the context of object relations theory,
Harding and Sutoris suggest that 'science presented as an interpretive, rather than a
controlling activity, is more likely to appeal to girls' (Harding & Sutoris 1987, p.33)

I have already mentioned the belief of several other researchers that altering the
image of mathematics and science and the way they are taught would benefit all
pupils, not just girls. My own research has similar implications and I would now like
to present the background to the project which is the subject of this thesis.
 'ii

1.2. The present study

1.2.1. How it started: a different picture

The initial impetus for this study came in the summer of 1986, when the Edinburgh
Centre for Mathematical Education undertook a project to assess the position of
women mathematics undergraduates at Edinburgh University. This involved a
survey of the educational careers of the entrants to the Mathematics Department
during the four-year period 1978-1982.

The most surprising finding in the survey was that the proportion of women taking
mathematics degrees was approximately 50% (Fraser & Cormack 1987, in Appendix
6). The national figures had indicated that the expected figure would have been
closer to 30% (Royal Society 1986). Two main questions emerged: whether the
results from Edinburgh University were typical of Scottish universities and if so,
what factors were involved in the greater participation of women in degree level
mathematics in Scotland. The first question entailed a more detailed examination of
the national data, and some of this is presented in Chapter 3 with further detail in
Cohen & Fraser (1992) (Appendix 6). In addition, a series of surveys were
undertaken / at Edinburgh University in order to achieve some insight concerning
students' attitudes toward mathematics focussing in particular on gender differences.
The analysis of these surveys and theft interpretation in the light of the research
described in Section 1.1 forms the main content of this thesis.

11.2. The Edinburgh University surveys

Three surveys were carried out between 1987 and 1989. Survey 1 involved
mathematics Honours undergraduates from all years. Survey 2 was considered as a
control and dealt with undergraduates from second and later years who were
qualified to study mathematics at university but had chosen to study other subjects.
Survey 3 explored in somewhat more detail second year mathematics
undergraduates' attitudes toward their learning experience, as well as the links
between achievement and attitude. The details of the sampling framework and
methodology of the surveys are given in Appendices 1-3, along with the
questionnaires used, full tabulations of responses to all questions and details of any
problems encountered with specific questions.
 22

The first survey of mathematics students was primarily exploratory due to the
difficulty of assessing to what extent previous research findings were applicable to
the particular context of the study. On these grounds, it seemed somewhat premature
to design a survey specifically to test whether these results would be replicated.
Unfortunately, resources were lacking to carry out in-depth case studies which
would have added a valuable dimension to the findings. It was therefore decided
that a second survey of mathematics students would be conducted (this was Survey
3) in order to examine in a more informed way some of the implications of the two
previous surveys. In order to reduce possible bias due to non-response or age
difference, the selected sample for Survey 3 was the entire second year Mathematics
Honours class.

The surveys thus all involved samples which were reasonably matched for
mathematical achievement at the end of secondary school. The students were also
matched in the sense that they had all chosen to study mathematics up to Higher/A
level. The general aim of Surveys 1 and 3 was to investigate the attitudes of
mathematics students to their courses and to examine, though in less depth, their
attitudes to mathematics at school and theft reasons for choosing to study
mathematics at university. The aim of Survey 2 was to investigate students who
could have studied for a mathematics degree but chose not to. The main points of
interest were theft experience of mathematics at school and their reasons for not
pursuing their study of the subject In all these surveys, gender differences were the
primary focus of interest.

For example, it was uncertain at the outset whether the results would replicate
findings of differences between the sexes such as estimation of ability and
expectation of success, attitudes toward mathematics (APU 1985) and importance of
affective relationships (Hoffman 1975). It was considered possible that women and
men who had been through the process of selection required to obtain good
mathematics qualifications at school might be similar in attitude and behaviour. The
hypothesis here was that women who choose to study mathematics at university
would tend to be the ones who had confidence in their mathematical ability and saw
the further pursuit of their mathematical education as valid and appropriate.
However, it was also considered possible that the somewhat competitive and
impersonal nature of university education might affect women negatively and there
would then be gender differences of the type found in the APU surveys.
 23

It was also uncertain whether there would be gender differences among the
non-mathematics students regarding motivation and attitude, such as women
expressing different reasons for not taking a mathematics degree or for theft choice
of degree subject.

Later chapters present the findings of the Edinburgh study and attempt to relate them
to previous research. Chapter 2 examines the school experience of students, while
Chapter 3 is concerned with the factors affecting choice of degree subject and
university. The results of the Edinburgh surveys regarding student attitudes toward
theft university experience are presented in Chapter 4.

1.2.3. Methodology

The samples obtained in these surveys were necessarily small since I had to do all
the work of questionnaire design, sample selection, interviewing, follow-up of
non-responders, coding and statistical analysis. At any rate, the relevant populations
are relatively small. It therefore has to be recognised at the outset that in any one
analysis, only a 'truly' large gender difference is likely to be detected by a,
significance test at conventional levels. Thus a gender difference in examination
performance of 15 percentage points, for example, might not be declared.
'significant', although it would be very big compared with differences observed in
larger national studies. A further major problem arises out of the large number of
tabulations and analyses carried out. It is obvious that if 1000 tables are analysed for
sex differences, 50 tables may be found to display a sex effect 'statistically
significant at the 5% level' even if there is no true sex effect. This is not a problem
peculiar to this study - it arises in any survey with multiple analyses. Even if we
ignore the screening out of 'uninteresting' or 'irrelevant' tables, there are 35 tables
of survey results in this text giving rise to approximately 142 possible tests of gender
effects in various aspects of mathematics education. Significance testing at the 5%
level would therefore be expected to produce about 7 'false positives' among the
rather few results which were found to be significant at this level.

At the same time, it is necessary to be aware of the limited power of the tests used
because of the small sample sizes, so that a 'true' difference of substantive
importance, 10 percentage points for example, might easily go undetected in
analysing a single table. In the analysis of 35 tables we may expect there to be a fair
number of 'false negatives'.
 24

In responding to these problems of conventional significance testing of survey data,

the first point is that such a small-scale study has to be regarded more as a case
study, hopefully one which indicates directions for future research, than as a study
capable of confirming or refuting specific hypotheses with a reasonable degree of
confidence. But the analysis is not done in a vacuum and it seems sensible to place
more confidence in results that are consistent with previous research findings than
results which are 'new' or in conflict with previous research. One must also be
extremely cautious in drawing firm conclusions from the latter in view of the small
sample sizes.

In view of the considerations discussed above, I decided to use the following

strategy. For each survey I first identified the comparisons that would provide
support either for hypotheses mentioned in the research literature or for hypotheses
generated from my own previous surveys. For any such comparison, I examined
whether the previous research hypothesis was supported by testing for significance at
the 5% level, If my survey gave a significant result, I regarded this as confirmation
(not proof) of the hypothesis that there is a 'real' difference between the groups
compared on the particular outcome variable, or a 'real' effect of the particular
variable used to distinguish the groups. Estimation of the size of the effect was
hardly worthwhile given that the small sample sizes would entail rather wide
confidence intervals. If my survey gave a non-significant result, I noted this but did
not attach much importance to it in view of the lack of power. Unless previous
research suggests a very large effect, it is not unlikely that a small-scale study such
as mine would fail to find an effect. In general, gender differences in this area are not
large, especially when other factors are taken into account (Hyde 1981).

Having dealt with comparisons related to previous research findings, I screened each
of my tables for possibly interesting or important results. Here the problem of
spurious results is clearly very serious because of the large number of possible
comparisons. One cautious approach would be to adopt a very stringent significance
level in individual tests so as to protect the overall Type I error probability of the
study, thus reducing the chance of 'false positive' results. However, attempting to
control the overall Type I error probability is not a particularly sensible solution for a
study such as this. No-one is going to make practical decisions or thaw firm
conclusions from such a small exploratory study, and the purpose of looking at
comparisons not related to previous research is to suggest ideas for further research.
In this context, it is appropriate to try and obtain reasonable power so as not to miss
 25

possibly interesting effects, while making it very clear to the reader that by doing so
any 'significant' results could easily be spurious. My approach was to use a 10%
significance level as an initial screening level. Any difference not significant at this
level was ignored. For differences significant at a level between 1% and 10%, I
made brief comments since the evidence that these effects are 'real' is rather weak in
view of the multiple comparisons problem. In this study, such differences will be
henceforth referred to as notable differences. Differences significant at a probability
level less than 1% were somewhat more worthy of attention and I devoted more
effort to trying'to relate these 'new' findings (of which there were only a couple) to
previous research. I would like to impress on the reader the fact that throughout this
study, the terms signiflcant(ly) and notable(ly) are used in the statistical sense only.

In this section, I have discussed differences significant at various levels without

giving any description of the tests applied. As we will see, the majority of the
tabulations consist of comparisons of percentages or mean scores of ratings on a
5-point scale. Tests of differences between two percentages were based on the
Pearson X 2 statistic with continuity correction applied to 2x2 tables of the numbers
of women and men in the sample making a particular response compared to the
numbers who did not Tests of homogeneity between three or more percentages were
also based on Pearson's X 2 . For comparisons of average ratings between two groups,
I used a two-tailed Mann-Whitney test. Although in these cases I included the
standard errors of mean differences in the tables for information, a t-test of mean
differences was deemed inappropriate due to the discrete rating-scale.
 26

Chapter 2.
Primary and secondary mathematics education

The previous chapter presented an overview of the research which provided the
general Theoretical background to this study. As the study was carried out in the
particular context of the British education system, it is desirable to examine the
initial conditions present in mathematics education in order to see to what extent
these conditions may affect the situation of women in university mathematics
education.

In this chapter, I will present the results of the major surveys in primary and
secondary mathematics education carried out in England, Wales and Scotland
(Sections 2.1 and 2.2). I will also describe in Section 2.2 the differences between
girls and boys in participation and performance at the public examinations in
mathematics and explore how the issues discussed in the first chapter might pertain
to the specific situation of British women in mathematics education. Finally, in
Section 2.3 I will discuss the results of the Edinburgh surveys concerning attitudes of
university students towards mathematics at secondary school.

2.1. Primary school

Major surveys of mathematics education in primary schools were carried out by the
Assessment of Performance Unit (APU) from 1978 to 1982 in England and Wales,
and by the Assessment of Achievement Programme (AAP) in Scotland in 1983 and
1988. The surveys also studied the situation at secondary level, but this will be
examined in Section 2.2. In addition to these large-scale surveys, there were some
smaller ones of interest. The Schools Council project on primary school mathematics
carried out by Murray Ward from 1972 to 1975 in England and Wales is relevant in
that it gives an indication of the then-current situation and so permits comparison
over a reasonably long time-period. Boroughmuir High School in Edinburgh did a
series of surveys of first-year intakes during 1977-1983, and this study provides
another dimension to the interpretation of the results of the 1983 AAP survey which
brings them in line with the findings of the other surveys
 27

2.1.1. England and Wales

The Schools Council project (Ward 1979) involved some 2300 10 year-old children
in the third year of junior school. In addition to the written test taken by the children,
there was a survey of the teachers asking them to rank the questions in the tests in
order of perceived importance. There were significant sex differences in 25 out of
the 91 questions. The questions on which the girls did significantly better were
mostly computational while those on which the boys did better were generally of an
applied and practical nature. The teachers ranked the 'girls' questions' as being more
important in terms of the children's eduction than the 'boys' questions'.

Each of the three APU surveys involved approximately 13000 11 and 15 year-olds
and consisted of written tests on concepts and skills, practical test interviews and
attitude questionnaires (APU 1985). The findings for the primary pupils were similar
to those of the Schools Council project: the topics in which the boys did better than
girls were applied and practical, and the ones the girls were better at were
computational and algebraic. Differences in the top attainment bands accounted for
most of the differences in performance. While there was little difference in boys'
and girls' enjoyment of mathematics and their perception of its usefulness, girls did
appear to see it as more difficult and expressed less confidence in their ability to
tackle questions involving measurement.

2.1.2. Scotland

In 1983, the Assessment of Achievement Programme surveyed roughly 1900

Primary 4 and 3000 Primary 7 pupils (ages 9 and 12 respectively) and found no
difference between the performance of girls and boys (AAP 1986). However, this
was possibly because the topics were not analysed separately, an oversight which
was remedied in the 1988 survey. The findings of the Boroughmuir High School
survey from 1977 to 1983, which involved approximately 1900 12 year-old pupils,
did give some indication that, as in England and Wales, the girls in the survey did
better than boys when the extent of computational skills was examined (private
communication by Peter Shannon, Boroughmuir High School). It seems therefore
likely that in the 1983 AAP survey boys were doing better than girls in other topics
and so the differences cancelled out to some extent. The 1988 AAP survey,
involving some 2900 Primary 4 and 3400 Primary 7 pupils, showed similar trends to
those found in the APU surveys (AAP 1989). Even as early as Primary 4, there were
differences in achievement for different topics.

2.1.3. Comments

While it was originally thought that no difference in mathematical performance

existed before the age of 11, the studies presented above, amongst others, show that
differences in performance, skills emphasis and attitude are present during primary,
even quite early on. I will not go into the possible reasons for these differences in
much detail since the theoretical aspect somewhat exceeds the scope of this study.

However, the results from the Schools Council project on the teachers' evaluation of
the importance of the different questions do raise the problem of how primary school
experience may affect boys and girls differently and so create the conditions leading
to the wider divergence observed at secondary level and beyond. The fact that girls
did better than boys in questions the teachers had rated as most important might
imply that the girls concentrated on the topics emphasised in class, perhaps to the
detriment of skills which would be more important at a later stage. Some researchers
have suggested that the teaching methods used at primary school combined with
differential expectations and interpretations regarding boys' and girls' achievements
and behaviour might have negative effects on pupils' mathematical development,
especially for girls (Scott-Hodgetts 1986; Wailcerdine 1989). If, as some of the
research presented in the previous chapter implies, girls are more concerned with
pleasing teachers than boys are (Hoffman 1975), then they may tend to develop work
patterns which bring them approval as well as success. These patterns, such as
diligence, care with presentation and a reluctance to 'muck around', may be seen to
disadvantage them later in a variety of ways. They may feel reluctant to approach the
teacher for help or tend not to use an exploratory approach to a problem they do not
know how to do, for fear of 'getting it wrong'. Of course, this interpretation is
somewhat simplistic since the actual dynamics involved are complex and depend on
a wide range of factors. It is also rather one-sided since one could equally
legitimately argue that lack of discipline is usually a hindrance to success at school.
 29

2.2. Secondary school

While variations in performance and attitude between boys and girls are already
present by the end of primary, it is during secondary education that the impact of
these differences becomes more obvious. National statistics show more girls opting
out of mathematics education at a relatively early stage, though less so in Scotland
than in England and Wales, as well as proportionally fewer girls in the top
achievement bands at the public examinations. These trends are of particular interest
for this study since they determine the numbers of women qualified to study
mathematics at university.

Since a pupil's attitude towards mathematics is very likely to affect the level to
which the subject is studied and how well s/he does in it, it seems appropriate to
begin the section by looking at the data from the APU surveys on attitude differences
between girls and boys at secondary level. I will then present the national data on
participation and performance in mathematics at the public examinations (Section
2.2.2) and discuss the implications of the differences observed between Scotland and
England and Wales (Section 2.2.3).

2.2.1. The APU surveys: findings on attitude

The differences in attitude between boys and girls were more marked at 15 than at 11
with girls finding mathematics more difficult, less interesting and less enjoyable.
They also expressed a relative lack of confidence in theft mathematical ability and
tended to attribute success to luck rather than ability more frequently than the boys.
In performance, girls were behind in all topics, but less so in those at which they had
done better relative to boys at age 11. However, though statistically significant,
many of the differences were rather small and, like those at primary, were mainly
due to more boys being in the top and bottom achievement bands. In addition, the
within-sex differences between topics were greater than the sex differences within
the same topic.

One result concerning work patterns was that boys attempted more problems but the
rate of correct responses for both sexes were very similar for many items. It would
thus appear that the girls omitted items rather than risk getting them wrong. Boys
also showed more confidence in tackling applied and practical problems (APU
 30

1985). It was suggested that the areas where the boys' mean scores were highest
relative to girls' were those which were important in subjects such as physics,
woodwork and technical drawing, options which are taken by more boys than girls
(APU 1982).

2.2.2. National patterns for participation and achievement at the public

examinations

It is during secondary education that personal preferences and aspirations influence

what subjects are studied and to what level they are taken. The patterns of subject
choice in secondary school reflect the idea that some subjects are more appealing to
one sex than to the other and it is a matter of concern to what extent social
expectations might channel able girls away from non-traditional subjects which,
given the opportunity, they might have enjoyed and been successful in. The Scottish
data on girls' participation in mathematics do suggest that the extent to which boys
and girls choose gender-appropriate subjects is at least partly determined by
restrictions imposed by the education system.

While the data presented below may be somewhat out of date, they describe the
prevailing conditions during the period the students surveyed at Edinburgh
University were nearing the end of secondary school. Table 2.1 shows the numbers
of male and female entrants and pass rates for CSE A level Pure and Applied
Mathematics in England and Wales. Tables 2.2a and 2.2b show corresponding
figures for SCE Higher and Certificate of Sixth Year Studies Mathematics in
Scotland. The decision to show pass rates at grades A-C for A level and A-B for
Higher was taken because these are the published entry requirements for
Mathematics degrees at English and Scottish universities respectively (see Section
3.2.2).

England and Wales

The proportions of boys and girls entering for 0 level Mathematics were roughly
similar since girls comprised approximately 47% of the total entry of 319108 for the
1984 examinations. However, proportionally fewer girls attained grade A and so
only 36% of those in the top achievement band were girls (Royal Society 1986).

At A level, female entrants were outnumbered approximately 2 to 1 and the situation

 31

Table 2.1
Female participation and performance in A level Pure and Applied Mathematics
Figures for 1984

women men % women

Number of entrants 21391 49196 30

Numbers passing at grades A-B 15231 33997 31

Pass rate % (71) (59)

Numbers passing at grades A-C 7915 19531 29

Pass rate % (37) (40)

Numbers passing at grade A 2374 7094 25

Pass rate % (11) C.4)

Source: Royal Society and Institute of Mathematics

Girls and Mathernatics,
and its Applications, 1986.

Table 2.2a
Female participation and performance in Higher Mathematics
Figures for 1984

women men % women

Number of entrants 7334 9041 45

Numbers passing at grades A-C 4646 5794 45

Pass rate % (63) (64)

Numbers passing at grades A-B 2411 3480 41

Pass rate % (33) (38)

Numbers passing at grade A 817 1430 36

Pass race .3 (11) (16)

Table 221,
Female participation and performance in Sixth Year Studies Mathematics
Figures for 1984

women men % women

Number of entrants for Paper I (Algebra) 293 589 33

ii (Calculus) 745 1621 31

III (Statistics) 111 178 38

rv (Computing) 201 703 22

V (Mechanics) 33 186 15

Numbers passing Paper II at grades A-C 440 753 37

Pass rate % (59) (46)

Numbers passing Paper II at grade A 83 166 33

Pass rate % (11) (10)

Source: special tabulations provided by the Scottish Examination Board.

 32

was more extreme in the top achievement band (Table 2.1). The pass rates for grades
A-E were similar for both sexes.

Additional data show that the proportion of female entrants varies with the particular
paper taken. It is lower for Applied Mathematics than for Pure and Applied
Mathematics, higher for Pure Mathematics, and highest for papers involving
statistics (Royal Society 1986; National Consortium for Examination Results 1987),

Scotland

There was very little difference in the proportions of each sex entering for 0 grade
Mathematics: out of a total entry of 33341 pupils, 16222 (approximately 49%) were
girls. The pass rate for grades A-C was somewhat higher for the boys: 10618 girls
(65% of the total entry of girls) and 11850 boys (69% of the total entry of boys)
obtained grades A-C. As for 0 level, proportionally more boys than girls obtained
grade A: 5900 boys (34%) compared with 4797 (30%) girls (figures provided by the
Scottish Examination Board).

Table 2.2a shows that though the proportion of female entrants to Higher
Mathematics was slightly smaller than for 0 grade entry, it was higher than that for
A level entrants. The pass rates for grades A-C were not notably different for boys
and girls, but relatively more boys attained the top grade.

The figures for entry to Sixth Year Studies Mathematics (Table 2.2b) show that at
this stage, the proportion of female entrants was more in line with that for A level.
As for A level, this proportion varied for the different papers offered. The Statistics
paper had the highest proportion of female entrants, and the Computing and
Mechanics papers the lowest. For the Calculus paper (Paper II), which is the paper
most frequently taken by pupils considering taking mathematics at university,
proportionally more girls passed at grades A-C than boys. The pass rates at grade A
were practically the same. However, SYS Mathematics is not a necessary
requirement for entry to a university mathematics course in Scotland and most pupils
doing SYS and considering university have unconditional offers from Scottish
universities based on their Higher results. It is therefore difficult to assess how
meaningful SYS grades are as indicators of ability, since they are quite likely to
reflect to some extent how motivated pupils are to work in the absence of external
pressure.
 33

2.2.3. Comments

Though the figures show that proportionally fewer girls than boys attain the top
achievement band for A level or Higher Mathematics, the differences are relatively
small and it appears to be mainly initial differences in participation that result in
boys outnumbering girls in the top performance bands. The fact that girls outperform
boys in the Sixth Year Studies Calculus paper suggests that it would be unjustified to
consider sex differences in performance at examinations as a simple indication of
lesser ability on the part of one sex or another, and that potential factors such as
differences in motivation, attitude and aspirations must also be taken into account.

The Scottish figures merit some discussion since the sex differences for Higher and
A level Mathematics entry are so dissimilar. The less specialised nature of the
Scottish education system probably encourages more pupils to continue with
mathematics after 0 grade, while their English counterparts have to decide what
three subjects will constitute their field of study for the next two years and determine
what courses they will be qualified to take at university. The Scottish system is thus
more likely than the English one to encourage girls not to opt out of mathematics at
an early stage since Scottish girls are not so restricted by the number of Higher
subjects they can take. The decision not to take mathematics is probably easier if
continuing with the subject entails giving up another one which is also of interest.
This is especially true if the pupil feels ambivalent towards mathematics in the first
place.

To what extent the prevailing view of mathematics as a somewhat 'masculine'

subject affects girls in this respect is difficult to assess since there is a certain
feed-back effect involved in the situation: if girls are negatively affected by the
image of the subject and opt out, then the image becomes self-fulfilling and is
perpetuated. Suffice to say that the current situation regarding girls' participation at
A level Mathematics does nothing to dispel this view and that a 'masculine' image
of mathematics is more likely to affect girls' propensity to study the subject
negatively rather than positively.

It would be interesting to ascertain whether mathematics has a less 'masculine'

image in Scotland than in England due to the relatively high proportion of girls
taking it at Highers. Data from the Edinburgh University surveys do seem to support
this idea since Scottish mathematics students tended to see their university course as
 34

less male-dominated than the English students did (Section 4.2).

 35

2.3. The Edinburgh surveys: school experience

The main purpose in asking the students questions about attitudes towards
mathematics at secondary school was to explore how those attitudes may have
affected their degree choice. Also, I wanted to examine whether the sex differences
in attitude observed in the APU surveys were present in the Edinburgh samples and
whether these differences varied between mathematics and non-mathematics
students. Some of the questions asked in the surveys concerning the secondary
education of the students are not considered in this chapter since the information
they provided either duplicated previous data or proved impractical to use. However,
the details are shown in the appendices.

2.3.1. Gender differences in attitude to mathematics at school

In all three surveys, the students were asked to rate the difficulty, interest and
usefulness of mathematics at school on a five-point scale (question 16 for Survey 1,
question 11 for Survey 2 and question 8 for Survey 3). The term 'usefulness' was
deliberately left undefined in order to allow for individual interpretations of how
mathematics might be useful. The results for the three surveys are shown in Tables
2.3a-c respectively as the means of the ratings for each group (all, men, women) as
well as the differences in the mean scores between men and women and the standard
error of the difference. The actual distribution of the scores can be found in the
relevant appendices.

The only significant difference was on the difficulty scale in Table 2.3b with the
women non-mathematics students rating school mathematics as more difficult than
their male counterparts did. This difference was in line with the APU findings
described in Section 2.2.1, although the APU surveys also reported that the girls in
their sample considered mathematics less interesting, a result the Edinburgh surveys
did not confirm. However, since the latter surveys were of mathematically-able
university students rather than a random sample of secondary school pupils,it is
perhaps not surprising that the differences reported by the APU were not consistently
replicated. One might expect mathematics students to display somewhat more
positive attitudes towards mathematics than those of the secondary school population
overall. Certainly the overall means among the mathematics students (Tables 2.3a
and 2.3c) are lower than those for the non-mathematics students (Table 2.3b),
 36

Table 2.3a
Question 16 Survey 1
Mean ratings by mathematics students of secondary school mathematics

scales used: Difficulty: 1 Very easy S very difficult

Interest: 1 very interesting S very boring
Usefulness: 1 very useful S a waste of time
standard
all men women ten-women error of
men-women

1.67 1.73 1.61 0.12 0.18

Difficulty
2.02 2.27 1.82 0.45 0.22
Interest
2.00 2.14 1.88 0.26 0.21
Usefulness*
N=81 t4=37 N=44

4 two women expressed no opinion and were not considered in the calculations.

Table 2.3),
Question 11, survey 2
Mean ratings by non-mathematics students of secondary school mathematics
standard
all men women ten-women error of
men-women

2.61 _3•45* 0.23

Difficulty 2.41 2.16

2.48 2.46 2.50 -0.04 0.22

Interest
2.43 2.61 -0.18 0.24
Usefulness 2.53

N=83 N=37 N=46

• pc0.03 on a two-tailed Mann-Whitney test for the difference between women

and men.

Table 2.3c
Question 8, Survey 3
Mean ratings by mathematics students of secondary school mathematics
standard
all men women men-women error of
men-women

2.10 2.20 -0.10 0.52

Difficulty 2.15

2.24 2.40 -0.16 0.26

Interest 2.32

2.29 2.30 -0.01 0.20

Usefulness 2.29

14=41 N21 14=20

 37

indicating somewhat more positive attitudes in general for the former.

2.3.2. Gender differences in the perception of encouragement

The students in the first and second surveys were also asked whether they felt they
had been particularly encouraged to do mathematics while at school (question 17 for
Survey 1 and question 12 for Survey 2). This question was motivated by various
studies on the connections between women's motivations and perceived affihiative
tendencies, which suggested that women might be more affected by others' opinions
and support (or lack of it) than men are (Hoffman 1975). The students in Survey 3
were not asked about encouragement in general terms since the focus of the survey
was somewhat different and the questions more specific.

The results from the first survey did support this hypothesis since the women
mathematics students were significantly more likely to say that they had been
encouraged to do mathematics while at school: 28 out of 43 women compared with
15 out of 37 men (65% and 41% respectively) with pc0.05 for a test on these
frequencies. Although more women non-mathematics students also said they had
been encouraged, the sex difference in Survey 2 was not significant nor particularly
large.

It is hard to say whether this finding implies that the women actually received more
encouragement or were merely more aware of it or more likely to admit to it, but it
does seem likely that positive encouragement of girls at school would offset the
potentially negative influence of seeing mathematics as an unusual subject for girls,
thus increasing their confidence in their mathematical abilities and perhaps
motivating them to continue studying the subject at university. There was some
indication of this in comments made by several of the women on how teachers'
attitudes had affected them, either negatively through 'chauvinistic' behaviour or
positively through active encouragement and support.

2.3.3. Possible question order effects

When comparing the ratings of school mathematics between the two surveys of
mathematics students (Tables 2.3a and 2.3c), there was a noticeable pattern of
difference. Almost all the mean ratings in Survey 3 were higher than the
corresponding ones in Survey 1, indicating that the students in Survey 3 had
expressed a more negative attitude than those in the first survey. The biggest
differences were amongst the women (two-tailed Mann-Whitney tests gave pcO.Ol
and pc0.02 on the difficulty and interest dimensions respectively for the differences
between the women in the two surveys).

It is possible that this result is an effect of a difference in question order between the
two surveys; the first sample was asked to rate mathematics at school (question 16)
after having been asked to rate their university mathematics course (questions 13 and
14), and school mathematics was generally rated as being rather less difficult and
more interesting and useful than mathematics at university (these results will be
presented in Chapter 4). It may be that the students tended to compare mathematics
at school with mathematics at university and thus rate the former more positively
than they might have otherwise done, particularly with respect to difficulty. On the
other hand, the second sample was asked to rate secondary mathematics (question 8
in Survey 3) before being asked to rate mathematics at university (questions 16 and
17).

There does not appear to be any obvious reason why women should have been more
susceptible to question order than men, though the data from Tables 2.3a and 2.3c
suggest this may be the case.

2.3.4. Reasons given for studying mathematics at school

The non-mathematics students were questioned on theft reasons for deciding to take
A level or Higher Mathematics (question 13 in Survey 2). Since the Scottish and
English education systems differ in the degree of specialisation at Higher or A level,
the differences in the responses between students with Highers and those with A
levels are shown as well as the differences between men and women (Table 2.4).

The question was an open one and the responses were classified into the following
categories:
Finding mathematics useful
Ability
Interest
Other
The students could give more than one response and the categories used were quite
 39

Table 2.4
çuestion 13, Survey 2
Reasons given by non-mathematics scudenrs for
having done A Level/Higher Mathematics
percent giving each reason

students with
Highers A levels women men all

Finding mathematics
useful 53 39 46 49 47

Ability 33 53 33 54 42

Interest 24 45 37 30 34

Otner 40 39 46 32 40

N=45 N=38 N=46 N=37 N=83

 HE

broad due to the variation in the wording of the responses. For instance, the category
'ability' included reasons such as mathematics being one of the student's best
subjects and confidence in being able to obtain a good grade. Mentions of A
level/Higher Mathematics being useful for entry to university, as opposed to being
specifically useful for the chosen university course, were classified as 'finding
mathematics useful', as were statements referring to mathematics as an important or
necessary subject. The category 'other' contained a fairly sizeable proportion of
responses to the effect that the choice of Higher/A level Mathematics had been
expected by the school or family or 'went well' with the other subjects taken.

The responses given by the students corresponded quite closely to the principal
reasons given by pupils for A level choice in the Office of Population Census and
Surveys study Young people's intentions to enter higher education (Redpath &
Harvey 1987).

The results are presented as percentages of each group (students with A levels,
students with Highers, women, men, all) mentioning each category of reason. As
mentioned above, the categories were not mutually exclusive and therefore the
percentages do not add up to 100. There were a few students who had done both A
levels and Highers; these were classified as having done Highers since their A level
grades tended to be relatively low.

The proportions of women and men mentioning each category of reason were
compared using x2
tests with one degree of freedom. This was also done for the
proportions of students with Highers and A levels giving each reason. None of the
differences were significant at the 5% level. However, there were a couple of
differences for which p<0.1 and these are presented below in accordance with the
methodology outlined in Chapter 1.

Differences between women and men non-mathematics students

The only difference of note was that men were more likely to mention ability as a
reason for having studied mathematics at school (p=O.OS). An obvious explanation
for this could of course be that the men had been in some sense more able, and the
data were examined to see if there was any evidence for this. One of the selection
criteria for the sample was obtaining grade A for Higher Mathematics or grades A-B
for A level Mathematics (see Appendix 2). So for practical purposes, all the students
with Higher Mathematics could be assumed to be of similar ability. I did find that a
 41

higher proportion of the men students with A level Mathematics had obtained grade
A; 14 out of 20 men (70%) compared with 8 out of 18 women (44%). However, the
exercise proved to be somewhat of a red herring since there was no notable
difference between the proportions of students obtaining A and B grades who said
they had studied A level Mathematics at school because of ability; out of the 22
respondents who had obtained A, 11 mentioned ability, compared with 9 out of the
16 students who had obtained B (50% and 56% respectively).

An alternative, though somewhat tentative, explanation could be that the difference

in mentions of ability between women and men reflects a difference between the
sexes in the expressed level of confidence in theft mathematical ability. The APU
surveys did find indications that girls appeared to show a relative lack of confidence
when compared with boys (Section 2.2.1).

If it is the case that the women did not have as much confidence in theft ability as the
men, then they may not have seen it as such an important factor in the choice of A
level or Higher Mathematics when compared with other factors such as interest or
usefulness. Of course, another aspect might be that men are more likely to feel the
need to emphasise their mathematical ability. I have discussed this idea at some
length in the previous chapter (Section 1. 1.5) and therefore I will not elaborate any
further, particularly since these hypotheses must remain speculative for the time
being due to the small sample sizes.

Differences between Higher and level non-mathematics students

Proportionally more A level students said that they had taken mathematics because
they had been interested in the subject (p=0.09). The differences between the
Scottish and English education systems could account for this variation in the
responses. Higher Mathematics is necessary for entry to many science courses at
Scottish universities (SUCE 1985), and is generally considered an advantage for
entrance to university even when it is not specifically required for the course. In
addition, any pupil considered at all capable of passing Higher Mathematics is
usually expected to take it. This may explain why considerations of interest were of
less concern to the Scottish students than the English ones, since the - latter are more
restricted in their choice and thus more likely to select A level subjects they are
particularly interested in. However, these ideas are also somewhat speculative at
present.
 42

2.3.5. Conclusions

The results presented above suggest that the women non-mathematics students had a
somewhat more negative attitude towards their mathematical abilities than their male
counterparts, a result which confirmed the APU findings (APU 1985). However, this
pattern was not found amongst the women mathematics students, which could imply
that women tend not to take a mathematics degree at university unless they have a
relatively high confidence in their ability. The apparent importance of any
encouragement they may have received at school to study mathematics would seem
to support this theory.

These data indicate that the results of large-scale surveys of school populations, like
those carried out by the APU, are perhaps not applicable to more self-selected
sub-populations, such as students choosing to do university mathematics degree
courses. This self-selection is both in terms of some sense of measurable ability as
perceived by the student and others, and interest in the subject. Although one must
note that here the term 'interest' is used in a general sense and may encompass
various aspects, such as aesthetic appreciation or maybe the idea of future usefulness
for a career.

The differences regarding women's and men's perception of encouragement at

school do imply to some extent that the self-selection process may be different for
the sexes, as discussed in Section 2.3.2. At this point, it is rather difficult to
ascertain whether the difference is due to internal differences in motivation patterns,
such as women's greater need for external support (Hoffman 1975), or whether it is
the perceived cultural definition of mathematics as 'masculine' which makes it more
difficult for women to pursue their study of the subject without some additional
motivation (Stein & Bailey 1975). As I have attempted to show in Section 1. 1, this is
an extremely sticky problem due to a pervasive circularity in social definitions and
expectations regarding gender-appropriate behaviour.

The students' reasons for their choice of degree are further examined in the next
chapter, which also studies the pattern of women's participation and performance in
mathematics at university in Scotland and England.
 43

Chapter 3.
Choice of degree and university

Data from Section 2.3 indicated that while women non-mathematics students in
Survey 2 found secondary school mathematics significantly more difficult than theft
male counterparts did, the women and men mathematics students showed no clear
evidence of attitude differences in this respect. On the other hand, the women
mathematics students in Survey 1 perceived themselves as having been more
encouraged to study mathematics at school, whereas the gender difference for the
perception of encouragement was relatively small among the non-mathematics
students.

Considering the above findings, it seemed possible that one would find gender
differences amongst mathematics students regarding the reasons for choosing to do a
mathematics degree. Some researchers have argued that women tend to consider the
wider implications of their choices and actions to a greater extent than men (Gilligan
1979), and therefore 'base life decisions on a wider range of criteria and less
systematically on academic and work criteria' (Maines 1985, p.317). It was
hypothesised that such a tendency would be particularly pronounced for women
mathematics students, since they might require additional motivation in order to
overcome ambivalent feelings concerning the appropriateness of such a degree
choice.

It was also considered interesting to compare the reasons given by mathematics and
non-mathematics students for their choice of degree, as well as examine the factors
which influenced the latter not to study mathematics as their main subject at
university. The discussion concerning the questions mentioned above is presented in
Section 3.1.

Taking into account the relatively high proportion of women amongst mathematics
undergraduates at Edinburgh University (Fraser & Cormack 1987), it seemed
worthwhile studying how such proportions vary between universities and what
aspects of a university might affect women's tendency to study mathematics there
(Section 3.2). Of particular interest is ascertaining how typical the situation at
Edinburgh University is of Scottish universities in general, and thus to what extent
the structure of the Scottish education system might affect the participation of
women in mathematics at university.

3.1. The Edinburgh surveys: choice of degree subject

The following section deals with the results from the Edinburgh University surveys
concerning reasons for degree choice. The survey questions asked varied somewhat
in form and content for the three surveys and are described in Sections 3.1.1 and
3.1.2. The results are presented and discussed in Sections 3.1.3-3.1.7. As usual,
details of the questions and results can be found in the relevant appendices.

3.1.1. Surveys 1 and 2: questions on choice of degree

In Survey 1, mathematics students were asked why they had chosen to do

mathematics at university in preference to other subjects (question 20). This was an
open question and the responses were classified into the following categories (results
shown in Table 3.1):
Ability
Enjoyment or interest
Career considerations
Influence of others
Other

In Survey 2, the question for the non-mathematics students was also open and
worded slightly differently, asking them how they had chosen their degree subject
(question 17). The answers were classified in a similar way to that of Survey I
(Table 3.3). In addition, the non-mathematics students were asked why they had
decided not to continue to study for a maths degree at university (question 14) and
the replies were grouped into the following categories:
Interest in other subjects
Finding mathematics lacking in usefulness or relevance
Not finding mathematics interesting or enjoyable
Finding mathematics difficult
Not seeing any career potential in mathematics or wanting a career in another
field
Other
 45

The results are shown in Table 3.4.

Each category of reason mentioned was counted as one mention. Thus if a student
gave two reasons which were classified together, this was still considered as one
mention rather than two. Students could however give more than one category of
reason.

Classification of the responses

The classification of the responses for the first two surveys did involve some
interpretation of the answers to the questions. The wording of the reasons mentioned
varied and it was impractical to consider precise classifications of the responses
since the samples were quite small. Therefore fairly broad categories were used and
this must be taken into account when considering the interpretation of the findings.

The reasons given by the mathematics students for degree choice were easily
classified since on the whole they concerned the favourable impressions students had
of mathematics and their own ability in it while at school. Due to the fairly small
size of the sample, it did not seem practical to distinguish between answers implying
self-perceived ability (such as finding mathematics easy) and those referring to more
'objective' manifestations of ability (such as obtaining good grades). So the two
types of reason were grouped in a single category. The categories for replies
mentioning the influence of career considerations and other people on the choice of a
mathematics degree were prompted by the literature on gender-linked differences in
motivation. Some investigators have postulated that men have a more instrumental
attitude than women and therefore would be motivated by practical considerations in
their choice of degree, such as its usefulness for obtaining a job (Maines 1985). The
APTJ surveys certainly found some empirical evidence of this (Joffe & Foxman
1986). On the other hand, some researchers see women as more likely than men to
be influenced by factors of a personal nature, such as encouragement and support
from others (Hoffman 1975). The category of influence of others included such
responses as knowing people who had done a mathematics degree as well as those
indicating direct or indirect encouragement or support.

While the classification of the non-mathematics students' reasons for degree choice
was along similar lines, the nature of the responses given varied somewhat from
those mentioned by the mathematics students. Some of the non-mathematics
students were studying subjects they had not done at school. So in these cases they
would have chosen their degree subject because they anticipated that they might be
good at it or interested in it, as opposed to knowing this from school experience. The
category 'career considerations' included responses which did not refer specifically
to career, but to the general practicality or relevance of the subject chosen. The
category 'other' was rather large due to the number of reasons mentioned. The
variety of degree subjects studied by the students in this survey made impractical the
precise classification of given reasons other than the ones mentioned most
frequently.

For the reasons given by the non-mathematics students for not having taken a
mathematics degree, it was deemed useful to classify remarks specifically
mentioning career considerations separately from those merely mentioning the
general lack of usefulness or relevance of mathematics. This was because career
considerations were relatively important in the choice of a non-mathematics degree
and therefore it seemed worthwhile to consider the effect of such considerations on
the decision not to take a mathematics degree. Of course, it is quite possible that
some students who just said that mathematics was not useful actually meant that it
was not useful for a career. But since prompting was avoided in the first two surveys
in order to reduce bias, it is not possible to ascertain to what extent this occurred.

3.1.2. Survey 3: questions on choice of degree

For Survey 3 (the second survey of mathematics students), the question on reasons
for degree choice (question 10) was fixed-response and the categories of response
were drawn up using information from the first two surveys. This format was used in
order to reduce both interviewer bias in recording and classifying responses and any
problems of analysis due to differences in the numbers of reasons given by
respondents. The students were given pre-printed cards with the question and the
response categories and asked to indicate on the card how important they felt each
consideration had been in influencing their decision to do a maths degree. They were
then asked if there had been any other reasons for their choice and these were
recorded. However, there were very few additional reasons given and so they were
not subjected to a formal analysis. As in all of Survey 3, any prompts used were
indicated on the questionnaire. The order of the responses on the pre-printed cards
was not randomised since it was thought that the time and effort involved in
producing and analysing many differently ordered cards outweighed possibly small
 47

gain in accuracy.

The first two reasons on the card 'being good at maths at school' and 'finding maths
easy' reflected different aspects of the 'ability' category identified in the responses
to the questions on degree choice in Surveys 1 and 2. The two aspects were seen as
sufficiently dissimilar to justify checking whether the students would respond
differently to the two reasons. Enjoyment of and interest in mathematics were seen
as fairly similar concepts and thus students were only presented with one category of
response for this type of reason. As for the perception of ability, there were two
categories concerning the influence of others. This was in order to distinguish
between the relatively indirect influence of having previously known people with
mathematics degrees, and the more direct effects of perceived encouragement and
support from others.

The students were also asked whether they had considered doing anything else
(question 11), but there was no notable difference in response patterns between
women and men and the results are not presented here (see Appendix 3 for details).

3.1.3. Patterns of response

The data for the first two surveys are presented in Tables 3.1, 3.3 and 3.4 as
percentages of each group of respondents (women, men, all) mentioning each
category of reason. It was noted that in Survey 1, women tended on average to give
more reasons than men: 26 out of 44 women and 13 out of 37 men (59% and 35%
respectively) mentioned two or more reasons. A x2
test on these frequencies with 1
degree of freedom gives p=0.05.

To allow for gender differences in the number of responses given, the data could also
have been presented as the proportions of mentions in each category to the total
number of mentions for each sex. The numbers of mentions are indicated in the
tables and the relevant percentages can be easily extracted from the given data if
required. However, as they do not indicate any different conclusions they have not
been presented separately.
 M

Table 3.1
Question 20, Survey 1
Reasons given by mathematics students for the their choice of degree

percent giving each reason

women men all

Finding mathematics easy or
being good at it at school 61 59 60

Being interested in mathematics 64 51 58

Seeing mathematics as useful

for a career 23 19 21

Influence of others 5 0 2

Other 11 14 12
N=44 N=37 N=81

Total number of categories mentioned 72 53

3.1.4. Survey 1: reasons for choice of degree

For this survey of mathematics students (Table 3. 1), the main reasons given for the
choice of degree were ability and interest. Career considerations were mentioned
relatively infrequently. There were no gender differences of note apart from the
tendency mentioned above for the women to give more than one reason for their
choice.

The fact that women tended to give more categories of reasons than men could be
interpreted in various ways. On the one hand, it could merely mean that women were
more communicative and so more likely to mention factors other than the principal
one. However, considering that the other open question analysed in Survey I
(question 22) did not elicit any difference in the numbers of reasons given, it might
be justifiable to interpret the observed difference as reflecting something other than
just a tendency for women to say more in response to an open question.

The other possibility is that the women tended to give more reasons because they did
not see a mathematics degree as such an obvious choice as the men did, and
therefore may have made a somewhat more considered choice in selecting the
subject. Such a hypothesis would be consistent with the interpretation given in
Section 2.3.2 of the women mathematics students' tendency to mention
encouragement to do mathematics at school more often than men. It must be pointed
out, however, that the evidence for such an explanation is somewhat weak in this
case.

3.1.5. Survey 3: reasons for choice of degree

There were no significant or notable differences in the response patterns between the
proportions of women and men rating each factor 'very important' (Tables 3.2a and
b). On the surface, this would appear to contradict the hypothesis that women might
be more motivated than men by support and encouragement from others, which was
supported by data from the first survey (Section 2.3.2). However, even if the women
in Survey 3 did not explicitly see encouragement from others as a particularly
important reason for degree choice, it is still possible that such support had some
influence on their decision, perhaps by making them more likely to contemplate
taking a mathematics degree than they might otherwise have been.
 50

Table 3.2a
Question to, Survey 3
women's ratings of the importance of factors
in their choice of a mathematics degree

of women finding each reason

very important fairly inoortant unimportant

Being good at mathematics

in school 90 10 o

Finding mathematics easy 30 53 5

Finding mathematics
rrresrinc 45 40 15
Thinking a mathematics
degree would be useful
for a career 35 33 10

Knowing people who had

taken a mathematics degree 0 25 75

ncouragement from
teachers or other people 15 40 45

N=20

Table 3.2b
Question 10, Survey 3
Men's ratings of the importance of factors
in their choice of a mathematics degree

of men finding eacn reason

very important fairly irnoortant unimportant

Being good at mathematics

in school 67 33

Finding mathematics easy 38 57 5

Finding mathematics
interesting 52 43 5

Thinking a mathematics
degree would be useful
for a career 43 38 19

Knowing people who had

taken a mathematics degree 0 5 95

Encouragement from
teachers or other people 14 43 43

N=21
 51

For both sexes combined, the ranking of the reasons according to the proportions of
respondents finding each factor 'very important' paralleled the ranking of the main
reasons given in the first survey of mathematics students (Tables 3.2a and 3.2b):
overall, 78% of the students in Survey 3 said that 'being good at mathematics at
school' had been very important, 49% rated 'finding mathematics interesting' as
very important, and 39% gave this rating for 'thinking mathematics would be useful
for a career'.

'Finding mathematics easy' was rated as very important by only 34% of the students,
and the two reasons concerning the influence of others were given this rating by a
very small minority of the students (15% for 'encouragement from teachers or other
people' and 0% for 'knowing people who had taken a mathematics degree').

The results on reasons for degree choice in the two surveys of mathematics students
clearly indicate that the principal motivation for doing a mathematics degree was
students' perceived ability in the subject. Interest was also a main reason in Survey
1, but somewhat less so than perceived ability in Survey 3. Career considerations
were not seen as particularly important in either survey. Although the difference in
question format makes comparison of the results of the two surveys rather difficult
regarding gender differences, the ranking within each survey of the relative
importance of each reason would seem to be similar for the two surveys.

3.1.6. Survey 2: non-mathematics students' reasons for degree choice

The non-mathematics students gave interest as the principal reason for their choice
of degree, with career considerations coming next (Table 3.3). Sizable minorities
mentioned other reasons which were not classifiable. Ability was a very minor
concern in this case. The main reason for not having chosen to study mathematics at
university was interest in other subjects (Table 3.4). The perceived lack of
usefulness or relevance of mathematics was also mentioned, but less frequently.
Other factors were not finding the subject interesting, finding it difficult, and career
considerations.

The only gender difference of note was that more women said they were not sure of
the reason for their choice of degree (a X 2 test gives p=0.08). On first inspection, the
findings (or lack of findings) in this case do not appear to support Gilligan's (1979)
and Maines' (1985) arguments that women base their 'life-decisions' on a greater
 52

Table 3.3
Question 17, survey 2
Reasons given by non-mathematics students for their choice of decree

• percent giving each reason

women men all
Ability 13 6 10
Enjoyment or interest 61 75 67
Career considerations 41 61 50
Influence of others 11 3 10
Not sure of reason 22 6 15

Other 20 22 21
N=46 N36 N=82
Total number of categories mentioned 77 64
* one response was missing.

Table 3.4
Question 14, survey 2
Reasons given by non-mathematics students
for not having done a mathematics degree

percent giving each reason

women men all
Interest in other subjects 46 32 40
Finding mathematics lacking
in usefulness or relevance 22 35 28

Not finding mathematics

interesting or enjoyable 20 16 18

Finding mathematics difficult 20 16 18

Not seeing any career potential
in mathematics or wanting a
career in another field 15 22 18

other 11 16 13
N=46 N=37 N=83
Total number of categories mentioned 61 51
 53

variety of considerations than men do. However, it is possible that the above
difference reflects this aspect in a less obvious way: not being sure of the reasons for
degree choice could imply that several equally appealing choices had been
considered, and thus the final decision would have been somewhat problematic. In
such a case, one might expect the reasons for the final choice to be confused and
hard to define. At any rate, this particular discussion is purely speculative, but
perhaps indicates areas needing further research. In this particular set of results,
gender differences are confounded with subject differences which might also
contribute to the observed gender differences in reasons for degree choice.

3.1.7. Comparison between mathematics and non-mathematics students

The reasons for degree choice given by the mathematics students in Survey 1
differed significantly from those given by the non-mathematics students. (Tables 3.1
and 3.3). The mathematics students mentioned ability more often and career
considerations less frequently than the non-mathematics students (for both
differences, pcO.Oi on X 2 tests with 1 degree of freedom).

These findings indicate that mathematics and non-mathematics students had

different motivations for their choice of degree. In the interviews, the former gave
the impression of having been unaware that, the mathematics taught at university
would vary somewhat from that taught at school, and appeared to have made the
choice to study mathematics assuming that their education would continue in the
same vein as it had at school. It would thus seem that the decision to study
mathematics at university often involved a 'drift' as opposed to a positive
decision.The relatively low frequency of mentions involving career considerations
tends to support the theory that mathematics students 'drifted' into the subject as
opposed to making a positive choice. A somewhat less negative interpretation is that
mathematics students had wanted to keep their options open at a stage when theft
career intentions were somewhat vague and therefore chose to do a relatively
non-vocational but well-regarded subject they thought they would enjoy and be good
at. Several mathematics students did specifically mention that theft choice had been
influenced by such considerations, and it must be pointed out that the two
interpretations are by no means exclusive.

However, it will be seen in Section 4.1.2 that the mathematics students' attitude
towards theft degree subject changed for the worse once at university, the main
 54

reason being increased difficulty attributed to the more theoretical nature of the
university course. The fact that the students' attitude became less positive once they
perceive the mathematics as getting more difficult, confirms the idea that for many
students the main appeal of mathematics in the first place was being able to do it at
school with relative ease, or perhaps having enough confidence in their ability to
believe that they could do it.

In contrast, non-mathematics students appear to have been motivated by an interest

which was less dependent on perceived ability and thus tended to persist throughout
their university career (Section 4.1.3).

Within-sex comparisons between mathematics and non-mathematics students

This particular comparison was originally intended to examine whether women

doing a non-traditional course, such as mathematics, and those studying more
conventional subjects would express different reasons for their choice of degree
subject. Only 7 out of the 46 women non-mathematics students were studying what
could be considered 'masculine' subjects (geophysics, civil engineering, physics,
chemistry and chemical physics). So in comparing the women mathematics and
non-mathematics students, we are implicitly comparing the motivations of women
who are doing a conventionally non-traditional subject with those of women doing
less 'questionable' subjects which are not seen as dubious choices on grounds of
traditional views regarding appropriateness or 'natural' aptitude. It was thought that
the former group might show higher levels of motivation concerning their choice of
degree, since the decision to study a non-traditional subject at university would
perhaps not be obvious and might therefore entail more consideration.

In fact, the hypothesis postulating differing levels of motivation for the two groups
of women was not confirmed since similar proportions of women mathematics
students from Survey 1 and women non-mathematics students mentioned interest as
a reason for choice of degree.

On the other hand, male non-mathematics students were somewhat more likely than
male mathematics students to give interest as a reason (a x2
test gives p=O.Oó). Such
a pattern could be seen as an indication that relatively unmotivated men may have a
tendency to take mathematics at university as an easy option. However, as the
difference was not statistically significant, this interpretation is merely speculative.
 55

3.1.8. Conclusions

The data suggest that while mathematics students on the whole are attracted to
mathematics because they see themselves as being good at it, the decision to study
the subject at university may be a somewhat more considered choice for women than
for men. The evidence is rather weak and indirect however, and further research is
needed to clarify the issues involved.

But the question raised is an interesting one which deserves some discussion, if only
of a purely theoretical nature. If there is some 'true' difference in the nature of the
decision to study mathematics between women and men, then it could be due to a
certain feeling of ambivalence among women when compared to men, possibly
because of the impression that mathematics is perhaps a subject more suitable for
men. If this is the case, then what is seen as an obvious choice for men might be a
somewhat less appealing one for women, and thus women might tend to require
more positive motivation than merely being good at mathematics in order to
continue studying the subject at university.
 56

3.2. Choice of university

I mentioned in Chapter 1 that one of the considerations which prompted this study
was the relatively large proportion of women mathematics undergraduates found at
Edinburgh University. My surveys did include some questions on reasons for the
students' choice of Edinburgh University, and the responses to these questions are
examined in Section 3.2.1.

In order to ascertain how typical the situation at Edinburgh was compared to other
universities, data on entrants to mathematical degree courses at other universities
were obtained. The summary of the findings from these data is presented in Section
3.2.2.

3.2.1. The Edinburgh surveys: choice of university

In Survey 1, the mathematics students were asked why they had chosen Edinburgh
University (question 22, Survey 1). The non-mathematics students in Survey 2 were
asked the same question (question 18, Survey 2).

The responses to the questions were classified in the following categories:

Location of university
Reputation of university or course
Structure of offered course
Influence of other people
Other
The results are shown in Tables 3.5a and 3.5b.

The category 'location of university' included reasons such as liking the city of
Edinburgh as well as considerations involving desirable distance from the student's
home. Responses concerning course structure typically referred to flexibility, both in
terms of choice of subjects and opportunity for change of degree subject. Despite
the relatively low numbers of mentions, the category 'influence of other people' was
considered separately since some previous research implied that women might be
more influenced by other people's advice and opinions (Hoffman 1975).

The main reasons mentioned in both surveys were location of university and
reputation of university or course.
 57

Table 35a
Question 22, Survey 1
Mathematics students' reasons for choosing Edinburgh University

percent giving each reason

women men all
Locality of university 80 81 80

Reputation of university
or course 59 62 60
Structure of offered course 27 8 19

:ofluence of other people 16 11 14

Other 18 27 22
N44 N=37 N=81

Table 3Jb
Question 19, Survey 2
Non-mathematics students' reasons for the choice of Edinburgh University

percent giving each reason

women men all

Locality of university 74 68 71

Reputation of university
or course 37 65 49

Structure of offered course 33 19 27

Influence of other people 11 11 11

Other 15 27 20

N=46 N=37 N=83

 W .
EAN

There were no gender differences for either survey which were significant at the 1%
level. The mathematics students (Table 3.5a) showed only one notable difference in
response patterns between men and women, with women being somewhat more
likely to mention course structure. Similarly, there was just one notable gender
difference for the non-mathematics students (Table 3.5b): the men in the sample
were more likely to give the reputation of the university or course as a reason. For
these differences, x2 tests with 1 degree of freedom gave p=O.OS and p=0 .02
respectively. There were no significant or notable differences between the
mathematics and non-mathematics students overall.

The data would appear to provide some indication, albeit tentative, that the women
mathematics students found the flexibility of the course offered at Edinburgh
University a more important factor in their choice of university compared to the
men. This may be a point worth examining in future studies in order to ascertain
whether women are more attracted by courses which keep their options open
regarding final degree choice and thus career opportunities, particularly for courses
which are not sanctioned by tradition and might therefore provoke feelings of
ambivalence or uncertainty about the wisdom of such choices. These considerations
also give rise to questions of how the structural differences between Scottish and
English university degree courses affect subject choice, since the Scottish degree
courses tend to allow more scope for change after entry to the course than the
English ones (Scottish Universities Council on Entrance 1985).

The men non-mathematics students' greater concern with the reputation of the
university or course could be interpreted in the light of Gilligan's (1979) finding that
men are more prone to use the notion of precedent to justify their decisions. In other
words, they tend to see past procedure as validating present course of action, rather
than consider the particular circumstances surrounding their present dilemma.
However, the mathematics students showed no gender difference for mentioning
reputation and so the evidence in support of this interpretation is somewhat
inconclusive in this case.

3.2.2. Entry to Mathematics degree courses in Scotland and England

In order to obtain a more general picture of the patterns of participation in

Mathematics degree courses at university, data were requested from the Universities'
Statistical Record (USR) on entrants to Mathematical Sciences degrees in Scotland
 59

and England for the years 1985-1987 inclusive. This period was selected because
prior to 1985, courses described by the Universities Central Council on Admissions
(UCCA) as 'Mathematics' included Computer Science and all joint courses with a
predominantly mathematical content. From 1985 onwards, the subject categories
Mathematics, Statistics, Computer Science and combined/other mathematical
subjects were grouped under the heading 'Mathematical Sciences' and the figures
compiled separately for each subject category. The data provided by the USR were
also broken down by year, university, domicile and sex.

For the purposes of this study, Mathematics, Statistics and combined/other degree
courses were treated as a single category. This was done for two reasons. Firstly,
some universities offered a relatively large number of joint degrees and had few
students taking the single subject Mathematics degree. Secondly, few universities
actually offered a single subject Statistics degree course, and as a result the recorded
number of entrants to Statistics degree courses was very small. The small numbers
of students made it difficult to make meaningful comparisons between the different
subjects and considering the variations found in the course definitions, it did not
seem worthwhile studying the figures for Statistics and combined/other degrees
separately. Entrants to Computer Science degrees were not considered in the
analysis.

It also did not appear particularly useful to analyse the data for individual years since
there did not seem to be much overall variation during the period studied.

In this section, I will only give a brief outline of the main features of the data
concerning women's participation in university mathematics. Further details can be
found in Appendix 4.

There was a fairly large difference between the proportions of women entering
mathematical degree courses at Scottish and English universities. At Scottish
universities, 40% of entrants were women, while the proportion for English
universities was 30%. Not surprisingly considering the numbers involved (Appendix
4), the significance level in this case for a X 2 test with 1 degree of freedom was very
high (pc0.0001). The result was similar when students' original domicile was
considered instead of the university they attended, with a greater proportion of
women recorded among Scottish-domiciled entrants than among English-domiciled
ones.
One explanation for the observed difference between Scottish and English intakes is
likely to be the difference in the proportions of Scotthh and English school-girls
qualified to study mathematics at university. In this case, 'qualified' means having
obtained grades A-C for A level Mathematics or A-B for Higher Mathematics. These
are the approximate entrance requirements for English and Scottish universities
respectively according to University entrance 1988: the official guide (Association
of Commonwealth Universities 1988) and the Scottish universities entrance guide
(SUCE 1985).

Using Tables 2.1 and 2.2a in the previous chapter, we see that the proportions of
girls among mathematically qualified English and Scottish school-leavers are similar
to the proportions of women among entrants to mathematical degrees at English and
Scottish universities respectively.

Since Tables 2.1 and 2.2a also show that relatively fewer girls obtain the highest
grade for A level/Higher Mathematics, it seemed useful to examine whether entrance
requirements affected the proportion of women among entrants to mathematical
degrees at individual universities. It proved possible to obtain some idea of the
requirements of English universities. However, the requirements of Scottish
universities turned out to be somewhat problematic.

University entrance 1988: the official guide provided data for 1986 on 'typical' A
level offers made and the range of grades accepted for entry to mathematics degree
courses for each university. For the period 1985-1987, the proportion of women
among entrants to English universities which required a grade A was 25%. The
proportion of women among entrants to universities asking for a grade B was 34%.
The highest grade required in a 'typical' offer was probably required in the
Mathematics paper, and certainly the former figure is the same as the proportion of
girls among pupils obtaining grade A in A level Mathematics in 1984 (Table 2.1).
The proportion of women among entrants to 'B' universities is not so different from
the proportion of girls among pupils obtaining grade B (this was 30% for the Pure
and Applied Mathematics paper in 1984 (Royal Society 1986)).

It can be noted that A level entrance requirements are generally lower for Scottish
universities than for English ones (ACU 1988). However, degree courses in Scotland
tend to last four years rather than three, and it would therefore be somewhat
misleading to conclude that Scottish universities have lower standards of entry than
English ones.
 61

The situation concerning Higher entrance requirements for Scottish universities is

difficult to assess due to several factors. Some universities consider entrance
requirements for a Faculty rather than a specific course, and some strongly
recommend CSYS qualifications. In addition, some Scottish universities have
relatively high proportions of English entrants with A levels, and these have to be
accounted for when considering the relationship between entrance requirements and
the proportions of women entering Scottish universities.

3.2.3. Conclusions

It would appear from the figures above that girls and boys who obtain the necessary
mathematics qualifications at Highers/A levels show similar participation rates for
entry to university Mathematics degree courses. That is to say that once they have
passed Higher/A level Mathematics, there is little difference in the proportions of
each sex continuing to study mathematics at university.It is therefore during the
secondary school period that women are most likely to terminate their mathematics
education.

Various researchers have pointed out that in most school systems, decisions have to
made during the period of adolescence regarding which subjects are studied at higher
levels. Choices may therefore be influenced by all the problems concerning gender
identity which this period often entails (Samuel 1983; Whyld 1983). At such a stage,
choosing mathematics may become problematic for girls because it does not fit in
with their concept of femininity, or it may simply be seen as irrelevant in terms of
their interests and aspirations (Eccles 1985). The higher proportion of women
studying mathematics in Scottish universities might therefore be a consequence of
the broad-based structure of the Scottish education system. By imposing less
restriction on the choice of subjects, Scottish schools seem to encourage girls to
continue studying mathematics up to the point they leave school to a greater extent
than the English system does (this idea has already been discussed in Section 2.2.3).

However, it remains to be seen what effects the introduction of the National

Curriculum in England and Wales might have on the patterns of girls' representation
at A level Mathematics.
 62

3.3. Summary

The differences discussed in this chapter have been selected ex post in accordance to
the methodology described in Section 1.2.3, and even then only attain rather weak
levels of statistical significance when tested independently (p=O.OS, p=0.08, p=0.05
and p=0.02 for the four gender differences mentioned in this chapter). It may
therefore be considered premature to try to draw substantive conclusions on such a
dubious basis. Nonetheless, it appears possible to me to suggest an underlying
theme linking these observations.

The main idea which emerges from this chapter is the question of whether the
decision to study mathematics at university is a more considered or problematic
choice for women than for men. There was some rather weak sign of this in the
women's observed tendency in Survey 1 t mention more reasons for their choice of
degree (Section 3.1.4). In the same survey, the women's expression of the
importance of the flexibility of the course in their decision to attend Edinburgh
University could tentatively be seen as also indicating a somewhat problematic
choice, since concern with keeping one's options open could imply an uncertainty
regarding the wisdom of one's choice. However, the problem remains to be resolved
of whether the decision is more considered because it is problematic, or because
women are to some extent more inclined to consider the wider implications of theft
decisions, as Gilligan (1979) suggests.

On the other hand, it may be the men who are making a 'negative' choice in deciding
to study mathematics at university, since they did tend to mention interest as a
reason for degree choice notably less often than men taking other subjects (Section
3.1.7). There is scope for further enquiry into this question, perhaps through more
in-depth interviews than I was able to carry out, which would appear to be of
particular interest in developing an understanding of the particular appeal of
mathematics for different individuals. Such an understanding is important in
pedagogical terms as it would hopefully enable teachers and curriculum developers
at all levels to have a better idea of how students might benefit more from their
mathematics education.

As regards the idea that the choice of mathematics at university might be more
problematic for women, there was no evidence from the non-mathematics students
that perceived difficulty was a more important factor for women not choosing to
 63

continue with their mathematics education (Section 3.1.6). The concept that
mathematics has a 'masculine' image and might thus be a less obvious choice for
women contemplating a university degree is explored, albeit in a somewhat
roundabout fashion, in Section 4.2.

In the next chapter, I examine the attitude and performance of the students at
university and attempt to identify how students might benefit or suffer from the
context of their university education. In keeping with the focus of my study, I was
particularly interested in ascertaining whether women and men were differentially
affected by the specific considerations and constraints involved in the university
learning experience and how such factors influenced their attitude and performance.
 Chapter 4.
Attitude and performance at university

The general aims of this chapter were rather varied, as befits the exploratory nature
of the present study. Firstly, in Section 4.1, I wanted to see whether there were
gender differences in the mathematics students' attitudes towards their degree
subject, and how their overall attitude changed once at university. The
non-mathematics students provided a control group for comparison purposes. It was
also considered interesting to examine how the students viewed the various
components of the mathematics course so as to obtain a more detailed understanding
of the effects the course had on the students.

In Section 4.2, we consider the perception of mathematics as a 'masculine' subject,

both for non-mathematics and mathematics students. It had been suggested that
seeing mathematics as a 'masculine' subject might make the choice to study the
subject at school a somewhat problematic one for girls who were concerned about
their feminine identity (Eccles 1985). I wished to ascertain whether students who had
demonstrated mathematical ability at the end of their school careers would still
perceive the subject as male-appropriate. An additional concern was how the
strength of such a belief was affected by school experience. It was hypothesised that
Scottish students might have less stereotyped views on the gender-appropriateness of
mathematics than English students since the Higher Mathematics classes attended by
the Scottish students tended to be less male-dominated than the English students'
A level Mathematics classes.

In order to further our understanding of the effects the university environment might
have on the attitudes of mathematics students towards their subject, the students in
Survey 3 were asked a series of detailed questions on various aspects of their
university learning experience. The questions and the students' responses to them are
discussed in Section 4.3. Section 4.4 explores gender differences in mathematics
students' achievement at university, as well as the relation between theft
achievement and opinions of their course.

A broader view of performance patterns for mathematical degrees is presented in

Section 4.5. This provides a background context for the findings of the present study
by showing the extent of the variation of degree results between women and men in
Scotland and England over a three-year period.

Section 4.6 rounds off the presentation of results from the Edinburgh surveys with
the findings regarding the career aspirations of the Edinburgh University students,
while Section 4.7 summarises the main points discussed in this chapter.

4.1. Attitude towards degree subject

Despite the findings from Surveys 1 and 3 showing no significant gender differences
in the mathematics students' attitudes towards mathematics at school (Section 2.3.1),
it was still thought that there might be differences in attitude towards mathematics at
university in line with the APU results (APU 1985). Since the women had expressed
a greater responsiveness to the presence of encouragement at school (Section 2.3.2),
the hypothesis was that the relatively impersonal style of university teaching might
have a more negative impact on their opinion of the mathematics course compared to
the men's. The findings on these questions are presented in Section 4.1.2.

The apparent differences in motivation between mathematics and non-mathematics

students concerning choice of degree subject (Section 3.1.7) prompted comparison
of the two groups' attitudes towards their respective degree courses in Section 4.1.3.
The mathematics students in Survey 1 emphasised the importance of ability as a
motivating factor for their choice of degree subject and then expressed rather
negative attitudes towards the university course during the interviews. It seemed
likely that the non-mathematics students would suffer less from the disenchantment
observed among the mathematics students, since their choice of subject had been
principally motivated by considerations other than having been good at it at school.

The APU surveys had shown that the differences between girls' and boys'
performance varied according to the topics involved (APU 1985). Algebra seemed to
be a topic girls were relatively good at, and it was thought that this tendency might
be apparent in the difference between women's and men's ratings of the algebra
component of the university mathematics course when compared with differences in,
the other topics. Since the APU had also found that boys did better than girls in the
practical and applied problems, it appeared probable that the physical component of
the Applied Mathematics 1 course would elicit more positive ratings from the men
than from the women. The findings concerning attitudes towards individual course
components are presented in Section 4.1.4.

It had seemed worth examining to what extent the non-mathematics students used
their school mathematical training in their degree course and whether their attitude
had changed since leaving school because of this (Section 4.1.5). I thought that such
questions would provide an indication of the students' perceptions of the relevance
or otherwise of the mathematics they had learnt at school and possible gender
differences in these perceptions (Eccles 1985).

4.1.1. Background information

Course structure at Edinburgh University

As I have previously pointed out, the Scottish education system differs somewhat
from that in England and Wales in that there is less restriction of subject choice
throughout the secondary education period. This is also true at university level since
during the first two years of the four-year Honours Bachelor of Science (BSc) or
Master of Arts (MA) course, students can take a variety of courses in subjects other
than their degree subject. It is therefore relatively easy to change degree subjects up
to the end of the second year (SIJCE 1985).

At Edinburgh University, Mathematics Honours students take a total of six subjects

in the first two years, including Mathematics IA and 2A. They have the option of
doing the Mathematics degree as an MA in the Arts Faculty, as opposed to a BSc in
the Science Faculty. In the former case, all the outside subjects may be taken from
the Arts and Social Science Faculties.

The first and second year Mathematics Honours courses, 1A and 2A, have three
components: algebra, calculus and analysis. At the time of the surveys, these
constituted roughly 30%, 50% and 20% respectively of the courses. Students with
good A levels may be exempted from the first year course and enter directly into
second year, but this is quite rare. In the third year of Honours Mathematics there is
more variety in the mathematics courses offered, but still no choice of options. At
this point, students taking the Mathematics/Statistics Joint Honours degree have a
slightly different curriculum, but follow many of the same courses as the single
subject mathematics students. At the end of third year, the students sit pan-final
examinations. Fourth year students select eight half-courses, or seven half-courses
 67

and a project, from a fairly wide range of options. The courses offered in the
academic year 1986/87 are listed in the questionnaire in Appendix 1.

In addition to the Mathematics Honours courses, students taking the BSc must take
the Applied Mathematics 1 course unless they can claim an exemption (on account
of A levels, for example). The course has a physical and non-physical component.
The non-physical component comprises numerical methods, computing, statistics
and probability. The physical component is mainly mechanics and vectors. Students
taking the MA do not have to take this course.

Survey questions

In Survey 1, the mathematics students were asked to rate the difficulty, interest and
usefulness of their current mathematics course as a whole, the individual
components of the course and the first year Applied Mathematics course (questions
13, 14 and 12 respectively). As the third and fourth year courses contained a variety
of topics which were not readily classifiable into the components algebra, analysis
and calculus, only the first and second year students were considered in the analysis
of attitude towards the different components of the course. The scales used for rating
were the same as those used for the questions on attitude towards school
mathematics in Chapter 2. The results for question 13 are shown in Table 4.1a and
those for question 14 in Tables 4.3a-c. In addition to the above questions, the
students were asked if they felt that they were having any particular difficulty in
their course compared to the rest of the class (question 23).

The non-mathematics students were asked to rate their degree subject (question 17,
Survey 2), whether they found a knowledge of mathematics useful for their course
and what level of mathematics they used (question 20). They were also asked
whether their opinion of mathematics had altered since school (question 21), and if
they thought mathematics could be useful to them in the future (question 22). The
results for question 17 are presented in Table 4.2.

The mathematics students in Survey 3 were asked to rate the Mathematics 1A and
2A courses (questions 16 and 17), the components of the second year course
(question 19) and Applied Mathematics 1 (question 21). In addition, they were
asked if their opinion of mathematics had changed since they came to university
(question 20). This last question was prompted by the results of the first survey
showing that the students expressed a more negative opinion of the subject once they
 T.

entered university. The findings for questions 16 and 17 are shown in Tables 4.1b
and 4. 1 respectively. Tables 4.4a-c present the results for question 19.

For Surveys 1 and 2, the answers to all the questions were recorded by the
interviewer. In Survey 3, the students were given pm-printed cards on which they
were asked to rate the components of the second year mathematics course and
Applied Mathematics 1. Again, this was done to minimise any bias in the responses
which might be induced by the ordering of the components of the question.

The data on the students' ratings of the courses are presented in Tables 4.1-4.4 as the
means of the rating scorns for each group (all, men, women) along with the
differences between the means for women and men. The standard errors of the mean
differences are also shown in order to place the magnitude of the differences in
context. However the rating scale was highly discrete and a comparison of the ratio
difference of means/standard error with the t-distribution would have been
unreliable. A non-parametric approach was judged more appropriate, and therefore I
used a two-tailed Mann-Whitney test to assess the significance of differences
between the distributions of women's and men's ratings. The previous research
discussed above had led me to expect differences, and so any differences found to be
significant at the 5% level were noted in the tables, in accordance with the
methodology presented in Section 1.2.3.

The complete frequency distributions of the rating scores can be found in the
relevant appendices.

4.1.2. Attitude of mathematics students towards their course

In Survey 1, there was a significant gender difference at the 5% level for the
'interest' ratings of the current course, with the women finding the course less
interesting (Table 4.1a). This was consistent with the APU findings (APU 1985)
relating to differential expressions of interest between girls and boys. However, the
ratings for the other two dimensions 'difficulty' and 'usefulness' showed no
significant gender differences, nor did the rating scores in Survey 3 of the
mathematics students' attitudes towards their Mathematics 1A and 2A courses
(Tables 4.1b and c). There was also no notable ilifference between the proportions
of women and men saying they felt they were experiencing particular difficulty in
the course.
Table 4.ia
Question 13, Survey 1
Mean ratings by mathematics students of their current mathematics course

Scales used: Difficulty: 1 very easy 5 very difficult

Interest: 1 very interesting S very boring
Usefulness: 1 very useful S a waste of time
standard
all men women men-women error of
men-women

Difficulty 3.30 3.08 3.48 -0.40 0.18

Interest 2.86 2.59 3.09 -0.50 0.22

Usefulness* 2.98 2.70 3.21 -0.51 0.23

N=81 N=37 N=44

pc0.04 on a two-tailed Mann-Whitney test for the difference between women

and men.
one woman expressed no opinion for this dimension and was not considered
in the calculations.

Table 4.lb
Question is. survey 3
Mean ratings by mathematics students of the Mathematics 1A course
standard
all men women men-women error of
men -women

Difficulty 3.05 2.81 3.33 -0.52 0.27

Interest 3.10 3.10 3.11 -0.01 0.20

Usefulness 2.72 2.86 2.56 0.30 0.32

N=39 N=21 N=18#

two women were direct entrants to 2A and had not done the lÀ course.

Table 4.1c
Question 17, survey 3
Mean ratings by mathematics students of the Mathematics 2A course
standard
all men women men-women error of
men-women

Difficulty 3.76 3.71 3.80 -0.09 0.22

Interest 3.05 2.90 3.20 -0.30 0.33

Usefulness 2.95 2.86 3.05 -0.19 0.33

N=41 N=21 N=20

 70

When the overall rating scores were examined, it was very obvious that the students'
attitude towards mathematics had declined during the course of their educational
career. The mathematics students in Survey 1 expressed consistently more negative
views of their current mathematics course compared with their attitude towards
mathematics at school, especially for the 'difficulty' dimension, (see the 'all'
columns in Tables 2.3a and 4.1a). The same trend was observed for the mathematics
students in Survey 3 (see the 'all' columns in Tables 2.3c, 4.1b and 4.1c).

Most of the students in Survey 3 (71%) said their attitude towards mathematics had
changed since coming to university. Amongst these, 48% had a worse opinion, 14%
a better one and 38% expressed the change in somewhat neutral terms (usually
saying they found mathematics at university more theoretical than at school).

The data on attitude change would appear to show that there are several factors
affecting mathematics students' attitudes. The more theoretical nature of the
university mathematics course was seen rather negatively, as can be seen in the
ratings of the analysis course (Section 4.1.4), which is the most abstract topic in the
first and second year courses. Comments by the students in Survey 3 in response to
question 20 on attitude change, indicated that they had not realised how much theory
was involved in the mathematics course.

The impression obtained from the interviews was that the students found the
theoretical aspect rather difficult and that it was the increased difficulty which put
them off the subject. This did seem to be the case for some students, judging by
comments made in Survey 3 such as 'I don't like it (mathematics) as much as at
school, it got hard'.

4.1.3. Comparison between mathematics and non-mathematics students

The non-mathematics students in Survey 2 found their degree courses significantly

more interesting and useful than the mathematics students in the first survey (a
Mann-Whitney test gives pc0.0001 in both cases). The ratings for difficulty were
very similar for the two groups (see Tables 4.1a and 4.2).

It may be noted that the gender differences for the non-mathematics students' ratings
of their degree subjects were very small for difficulty and negligible for interest and
usefulness (Table 4.2).
 71

Table 4.2
Question 17, Survey 2
Mean ratings by non-mathematics students of their current degree course

Scales used: Difficulty: 1 very easy 5 very difficult

Interest: 1 very interesting 5 very boring
Usefulness: 1 very useful 5 a waste of time

standard
all men women ten-women error of
men-women

:flficulty 3.29 3.38 3.22 3.15 0.20

:nterest 2.20 2.22 2.20 L32 0.20
Usefulness 2.20 2.19 2.22 0.27

N=83 N=37 N=46

 72

The negative change in the attitude of the mathematics students discussed in Section
4.1.2, when seen in conjunction with the non-mathematics students' attitudes
towards their degree subjects and the reasons for degree choice examined in the
previous chapter, further supports the hypothesis that there are differences in
motivation between mathematics and non-mathematics students regarding their
choice of degree subject. These differences would appear to affect their subsequent
attitude towards their chosen subject, since the non-mathematics students considered
their degree subjects significantly more interesting and useful on the whole despite
the two groups having rated their degree courses similarly for difficulty.

4.1.4. Attitude of mathematics students towards the components of their

mathematics courses

The Mathematics 1A and 2A courses

There were a few gender differences in Survey 1 which were significant at the 5%
level: the women found analysis and algebra significantly more difficult and calculus
less useful (Tables 4.3a-c). In Survey 3, as for the ratings of the course overall
(Section 4.1.2), none of the sex differences in the ratings were significant and most
were fairly small (Tables 4.4a-c). The overall conclusions from Tables 4.3 and 4.4 as
regards gender differences in attitudes to components of the mathematics course, are
similar to those regarding differences in attitudes to the current course as a whole
(Section 4.1.2).

The above results do not seem to particularly support the hypothesis that women
might show a more favourable attitude towards algebra. They also do not confirm
the more moderate proposal that any gender differences in attitude might be less
extreme in algebra than in other topics. However, algebra at university is rather
different from the kind of algebra taught at school level, which would have been
what the APU studies were referring to as 'algebra'.

The ratings of the different components of the mathematics course by the first and
second year students in Survey 1 indicated a fairly consistent ordering pattern (see
the 'all' columns in Tables 4.3a-c). Analysis was seen as the most difficult and least
interesting and useful topic, while algebra was considered less interesting and useful
than calculus. The results from Survey 3 also showed the same ordering pattern
(Tables 4.4a-c).
 73

Table 43
Question 14, survey 1
Mean ratings by 1st and 2nd year mathematics students
of the components of their current mathematics course

Scales used: Difficulty: 1 very easy 5 very difficult

Interest: 1 very interesting S very boring
Usefulness: 1 very useful 5 a waste of time

standard
all men women men-women error of
men-women
a Analysis
Difficulty 3.94 3.58 4.24 -0.66 0.27
:n:eresc 3.83 3.50 4.10 -.0.60 0.29
Usefulness 3.54 3.58 3.69 -0.11 0.31

hI Algebra
Difficulty 2.43 2.08 2.72 _0.64t 0.25

Interest 2.96 3.08 2.86 0.22 0.30

Usefulness 3.19 3.29 3.10 0.19 0.23

cI Calculus
Difficulty 2.64 2.46 2.79 -0.33 0.28

Interest 2.43 2.17 - 2.66 -0.49 0.23

Usefulness 2.41 2.00 2.76 _075* 0.23

N=53 14=24 14=29

p<0.05 on a two-tailed Mann-Whitney test for the difference between

women and men.
 74

Table 4.4
Question 19, Survey 3
Mean ratings by the 2nd year mathematics students
of the components of the Mathematics 2A course

Scales used: Difficulty: 1 very easy 5 very difficult

Interest: 1 very interesting 5 very boring
Usefulness: 1 very useful 5 a waste of time
standard
all men women men-women error of
men-women
a Analysis

Difficulty 3.66 3.38 3.95 -0.57 0.27

:ncerest 3.39 3.19 3.60 -0.41 0.36

usefulness 3.00 2.86 3.13 -3.29 0.31

b) Algebra
Difficulty 3.15 3.05 3.25 -0.20 0.26

interest 2.61 2.48 2.75 -0.27 0.30

Usefulness 2.73 2.62 2.85 -0.23 0.28

cI Calculus

Difficulty 3.00 3.14 2.85 0.29 0.21

Interest 2.46 2.62 2.30 0.32 0.30

Usefulness 2.56 2.67 2.45 0.22 0.27

N=41 N=21 N=20

 75

One of the more consistent findings which emerged from the two surveys c
mathematics students was the relatively negative attitude expressed towards th
analysis courses in the first and second years. As mentioned above 3 this course i
theory-oriented and emphasises the importance of formal and rigorous prool
Comments made by the students during the interviews indicated that many of ther
were put off by the formal and abstract aspect of the course and did not see the poir
of it. They also mentioned a lack of confidence when tackling questions in analysi
due to the difficulty in telling whether they had found the 'right' way to do th
exercise or were on the wrong track. They did not appear to be comfortable with th
idea that there might be several approaches to the problem. This attitude illustrate
some of the points I discussed in Section 1.1.5 concerning the way mathematics i
presented in schools.

This attituoe was not so obvious with the third and final year students, mayb
because of increased familiarity with the subject and a more varied course structure
The complex analysis course in third year seemed to be particularly popular, and it i
possible that at this stage students are more aware of and comfortable with th
techniques and methods taught in the analysis course during the first two years.
4 !

Since analysis is not really taught at school level, the course would be equall'
unfamiliar to both men and women initially. Yet the women in Survey 1 found i
significantly more difficult. It may be that the 'uncertain' aspect of the subjec
nientioned above advantages men in that they appear to be less put off by perceive
difficulty or unfamiliarity and more likely to anticipate success (APIJ 1985). On
possible interpretation for the difference in attitude towards analysis might be tha
men feel more comfortable about the possibility of 'failure' (not getting an exercis'
right first time) than women are. After having 'failed', men may tend to assume tha
they had just not gone about it the right way and try again, while women may fee
discouraged and not persist after the initial attempts (Hoffman 1975). This couli
induce a self-perpetuating cycle of feeling of failure and loss of confidence
Alternatively, the men may be more inclined to try and master the subject out o
competitiveness and a desire for control, whereas the women might be less likely t
feel motivated by such factors when faced with a relatively uninteresting subjec
(Seward & Seward 1980; Chodorow 1989).

However, these two explanations are not exclusive and both could well contribute t
the interpretation of the observed differences. Since there do not appear to be an:
 76

data at present on how men and women actually perform in different topics at
university, it is difficult to say how the differences in expressed attitude reflect
differences in achievement.

The Applied Mathematics I course

There were also a couple of significant differences at the 5% level in Survey 1

between women's and men's ratings of the non-physical component of Applied
Mathematics 1, with the men finding it more interesting and useful. The gender
differences for this component were non-significant in Survey 3, and were rather
small in both surveys for the physical component of the course (Appendices 1 and
3).

These results do not really support the hypothesis that the men would have a more
favourable attitude towards practical applications of mathematics. Since most of the
women in the sample had taken Higher/A level Physics, it is probable that they had
more inclination towards the physical sciences than the girls in the APU surveys. In
this case, the small difference in the women's and men's ratings of the physical
component of the Applied Mathematics 1 course was perhaps to be expected.

The gender differences in attitude towards the non-physical component were a bit
surprising since there had been several indications that statistics was a subject which
was relatively appealing to women (Section 2.2.2). However, the course also
included a substantial amount of computing and it was maybe this aspect which
affected the women's attitude, a notion which was briefly discussed in Section 1.1.5.

4.1.5. Non-mathematics students' attitude towards mathematics at university

Most of the non-mathematics students (76%) said that they found mathematics
useful for their course, and 90% thought that it might be useful for the future. There
were no notable sex differences in the responses to these questions. Since the courses
varied widely, the types of mathematics the students found useful were not easily
classifiable.

A majority of the students (60%) claimed their attitude towards mathematics had not
changed since leaving school, with 24% saying it had changed for the better and
16% for worse. Students in the Science Faculty were significantly more likely to
change their opinion, both for the better and worse, than those in other Faculties (a
 77

test with three degrees of freedom gives pcO.Ol) (question 21 in Appendix 2).
This might be explained by the fact that science students probably come into more
direct contact with mathematics and are often required to take a mathematics course
in conjunction with their degree subject

4.1.6. Conclusions

The results of the mathematics students' ratings of their mathematics courses appear
to indicate some difference in attitude between women and men, since all the
significant differences showed the men having a more positive attitude. However,
there were .not that many significant differences and quite a lot of rather small ones
in differing directions. It is therefore difficult to assess the true scale of any
measurable difference in attitude which may exist in the population under
consideration. Judging from the scale of the differences observed in my study, I
would guess that it is probably reasonably small.

As a comparison, the non-mathematics students' ratings of their degree courses were

consistently very similar for both sexes. As I pointed out before in Section 3.1.7, few
of the women in Survey 2 were studying what could be considered male-typed
subjects (ie. 'hard' sciences, with a predominance of men students), while even
fewer men in that particular sample were taking degree courses which might be
female-typed (in terms of having a predominance of women students). This suggests
that gender differences in attitude toward male-typed subjects, such as mathematics,
might tend to be exaggerated when compared to differences between women and
men taking subjects considered more gender-appropriate. Stein & Bailey (1975) take
a similar position when discussing observed gender differences in achievement
motivation.

There was stronger evidence of attitude differences between mathematics and

non-mathematics students toward their degree subjects, which supported the
tendencies noted in Section 3.1.7 regarding differential motivations for subject
choice between the two groups. Despite having rated their respective degree courses
very similarly for the 'difficulty' dimension, the non-mathematics students expressed
significantly more positive attitudes toward the 'interest' and 'usefulness'
dimensions than the mathematics students in Survey 1. Considering the mathematics
students' emphasis on ability as a reason for subject choice (Section 3.1.7) and their
negative reaction to the perceived difficulty of the university mathematics course
 M

(Section 4.1.2), it would appear that non-mathematics students chose their degree
subjects more out of a sense of inherent interest and were thus less likely to be put
off by subsequent difficulty.

The non-mathematics students also displayed an appreciation of the usefulness of

mathematics, both for their courses and their future careers. Since there were no
notable gender differences, it seems that most students of both sexes studying a wide
range of subjects are aware of mathematics' potential importance in the outside
world. One must however remember that the students in this sample had achieved
good results in mathematics at secondary school and might therefore be assumed to
feel relatively comfortable with the subject and its applications compared with
students who had experienced more difficulty at school.

4.2. Mathematics as a 'masculine' subject

The primary goal of the surveys was to explore whether and to what extent women
and men differed in basic attitude towards mathematics. But in order to obtain some
relatively objective measure of student attitude without appearing to be involved in a
somewhat contentious and often emotion-laden field of study there were very few
questions explicitly addressing the issue of women in mathematics. It was felt that
explicit questions about gender differences would prove detrimental to the
objectivity of the study by affecting student response. It was feared that by asking
personal questions on their perceptions of how gender affected attitude and
performance in mathematics, they might consider that the survey was specifically
concerned with gender issues and respond accordingly, rather than seeing the
questions in the wider context of mathematics education as a whole. Another
consideration was that some students might be less likely to participate in what could
be seen as a feminist survey and therefore bias the response pattern.

However, I was interested in finding out whether students did tend to characterise
mathematics as 'masculine' (more appropriate in some sense for men), since this
would be likely to affect women's inclination to select the subject for their degree
(Eccles 1985). While Survey 1 was primarily exploratory, it did provide some rather
tentative indications that even women studying mathematics at university differed
from the men in their attitude towards the subject (Section 4.1.2). It was
hypothesised that women's attitude might be adversely affected by a general
impression that mathematics was a more appropriate subject for men to study at
 79

university (Northam 1986; Isaacson 1986), but it was uncertain to what extent the
students would be inclined to express this specifically.

For the survey of non-mathematics students (Survey 2) and the second survey of
mathematics students (Survey 3), it was suggested that one might examine this
aspect in a fairly neutral manner by asking students what they thought the proportion
of women among university mathematics students was. I hoped that this would give
some indication of how the students perceived the subject in terms of
gender-appropriateness, and whether this view was affected by variables such as sex
or type of education system attended (ie. Scottish or non-Scottish). The latter
variable was considered of interest since Scottish Higher Mathematics classes tend to
have higher proportions of girls in them than A level Mathematics classes in
England and Wales (Section 2.2.2). It was thought that school experience might
influence the perception of mathematics to some extent due to the presence of
peer-models (Seward & Seward 1980). Therefore, students who had attended
secondary mathematics classes where girls constituted only a small minority of the
pupils might be more likely to perceive the subject as 'masculine' than students who
had been in classes where girls had been more in evidence.

4.2.1. Perception of the non-mathematics students

In Survey 2, the students were asked what they thought the proportion of females in
the university mathematics degree course was (question 16, Survey 2). Most of the
students were not attending the mathematics classes, and so the answers presumably
reflected how male-dominated the students perceived the subject to be, rather than an
informed statement. Since the responses to the question were fairly rough estimates,
they were dichotomised into the categories 'under 40% women' and '40% to 60%
women' for the purposes of analysis (no-one thought that there were more than 60%
women). The first category corresponded to a perception that the mathematics course
was male-dominated, while the second was seen as representing a view of the course
as roughly balanced regarding the proportions of women and men.

On the whole, 57% of the non-mathematics students thought the university

mathematics degree course was male-dominated. It is interesting to note that no-one
thought that the course might be female-dominated. There were no clear indications
that the students' perceptions were influenced by which school system they had
attended. There were also no notable differences between the women's and men's
 NIVA

responses (details in Appendix 2).

4.2.2. Perception of the mathematics students

Survey 3 included a question in which the students were asked whether the
proportion of women in their class at university was what they had expected it to be,
what they thought it was and what they had expected it to be (question 15, Survey 3).
If they mentioned a proportion under 40%, they were also asked why they thought
them would be (or were) fewer women.

Most (68%) of those who expressed an opinion thought that the class was roughly
balanced (the proportion of women in this particular class was 45%) and no-one
thought it was female-dominated. Again, there was no sex difference in this respect,
but there was a notable difference between Scots and non-Scots. Out of 28 Scottish
and 12 non-Scottish students, 6 and 7 (21% and 58%) respectively thought their
class was male-dominated (p=0.06 for a X2 test on these frequencies with one degree
of freedom). As before, the criterion for distinguishing between Scottish and
non-Scottish students was the type of public examinations taken (Highers or A
levels).

All but one of the 11 students who said it was not as expected had expected the
proportion of women to be lower (Appendix 3).

The reasons given by the students for women's expected or perceived

underparticipation did not seem to differ for women and men. The masons
mentioned mainly concerned women's apparent tendency not to study science due to
tradition. I have listed some of the responses below.

'I think it might be instilled that maths is for boys.' (woman)

'You don't expect girls to do maths, it is thought to be a difficult subject.' (woman)

'I felt that girls don't go for maths much. People seem to think that maths isn't arty.'
(man)

'Women don't seem to want to associate with the sciences.' (man)

'Probably males are given more push towards that sort of course.' (woman)

The notion of men being given more 'push' towards the sciences is an interesting
one in the light of Kelly's (1978b) finding that among Higher leaven relatively more
boys were not given the opportunity to drop science. Some of the reasons mentioned
were also general expectations due to the perceived situation in secondary school.

'At school more males were interested in maths and science.' (man)

'They don't seem to like science much. That's the impression from school, most
girls are in arts classes.' (woman)

4.2.3. Conclusions

There was no strong evidence from the non-mathematics, students that they saw
mathematics as a particularly 'masculine' subject since the proportion of students
thinking that the university mathematics course was male-dominated was not notably
different from the proportion of students thinking otherwise. However, one might
interpret that fact that no-one said they thought the university mathematics course
was female-dominated as an indirect indication that mathematics was perceived as
more 'masculine' than neutral.

One interesting finding was that the mathematics students who thought (erroneously
in this case) that their course was male-dominated were notably more likely to have
done A levels rather than Highers. An interesting hypothesis to investigate here is
that strong views of mathematics as a 'masculine' subject may be more apparent
among students who had not attended Scottish schools, since one would think that
students would have to have a fairly definite view of mathematics as a 'masculine'
subject for them to perceive their class as male-dominated when in fact it was not. It
therefore seems possible that some of the non-mathematics students did not actually
have particularly strong views on whether mathematics was male-dominated or not,
and so any difference between the responses of Scottish and non-Scottish students in
that sample may have been attenuated.

The mathematics students' explanations for their perception of the course as

male-dominated highlighted the impact of social expectations on secondary pupils'
subject choices, and thus their educational career. The observed difference between
 M
.

Scottish and non-Scottish students in their perception of the course, although not
significant, tentatively suggests that such expectations might have a stronger
influence in a more restrictive education system, such as the one in England and
Wales, than in the fairly broad-based Scottish system. These expectations not only
affect the pupils' personal preferences, but also school policy (what combinations of
papers the pupil can take, for instance). Therefore it may be more difficult to
combine a Mathematics A level with A levels in non-science subjects, a rather
untraditional choice which appears to be more popular among girls than among boys
(Cohen & Fraser 1992). Such restrictions would contribute to the attrition rate of
female Mathematics A level entrants. Since Kelly (1978b) suggests that timetabling
is an important factor even in Scottish girls' decisions to drop science before 0
grade, it appears reasonable to assume that timetabling restrictions would have rather
more impact on the choice of A levels.

The data presented above could be seen to illustrate how education policies at
secondary school level might foster a situation, both at secondary and university
levels, where expectations concerning the gender-appropriateness of certain subjects
and the tendencies of women and men to take such subjects are more (in the case of
England and Wales) or less (in the case of Scotland) self-fulfilling. Such a situation
indicates a need to reconsider education policies in the light of their potentially
negative effects in terms of unnecessary wastage and limiting of options. The
comment from one somewhat cynical (or perhaps realistic) mathematics student was

'The change for equality hasn't got there yet.'

4.3. The university learning experience

Since the second survey of mathematics students (Survey 3) was small-scale (under
50) and directed toward students who were relatively accessible, the length of the
questionnaire was not so restricted by time constraints. I was therefore able to ask a
variety of questions on specific aspects of the university learning experience. These
questions examined the students' opinions of the teaching staff and course
presentation, factors affecting students' approach to the course exercises and how
they explain and cope with perceived failure.

The motivation behind these questions was to see whether there were differences in
the students' perceptions of their learning experience which might affect their
 rM

attitude and performance, and which were possibly gender-linked. Some of the
questions were formulated using the findings of previous research on gender
differences in motivation for achievement (Hoffman 1975) suggesting that the
'human dimension' of a situation might be considered more important by women
than by men (Section 4.3.1). Several questions were inspired by studies involving
the causal attribution of success and failure (Frieze 1975) which indicated that the
types of explanations given for success or failure influenced (and were influenced
by) the expectation of failing or succeeding. There appeared to be differences in the
types of explanations women and men gave. The APU surveys found some
tendency amongst girls to attribute success in mathematics to luck rather than ability
(APU 1985), thus downgrading their capabilities and possibly eroding theft sense of
confidence (Section 4.3.2).

Other questions were motivated by interest in the way the students related to the
university learning environment (Section 4.3.3).

For reference, the questions dealt with in this section are nos. 23, 24, 27, 28 and
32-39 in the questionnaire for Survey 3 given in Appendix 3. The questionnaire
includes the lists for the pre-printed cards the students were asked to complete in the
course of the interviews for certain questions. As before (see Section 3.1.2),
pre-printed cards were used because of concern that if the lists were presented orally,
the ordering might influence the responses. The cards were to ensure that the
students were aware from the outset of the full range of responses and could select
the ones which were relevant to them.

In addition to the questions described below, the questionnaire included several other
questions which did not seem to yield particularly useful data. They are therefore not
presented in any detail in this chapter, but the results are shown in Appendix 3.

It must also be noted that there were not very many significant or notable differences
between the sexes for these questions. So in the absence of specific mention to the
contrary in the presentation of the results, the reader may assume that there were no
notable differences.
4.3.1. Affiliation and the personal dimensions of attitude

The students were asked to grade on a three-point scale (very important, fairly
important, unimportant) how important they found the mathematics lecturer in
determining how much they enjoyed a particular topic (question 23, Survey 3). They
were then asked to indicate on a pm-printed card, using the same scale, how
important they felt it was for the lecturer to have the qualities listed on the card
(question 24, results in Table 4.5). There was also a question on whether they would
like more feedback from the staff on how they were doing and on the staffs real
opinions about their strengths and weaknesses (question 27). The aim of these
particular questions was to examine to what extent students were affected by the
human element in the course, as opposed to the more impersonal and academic
aspects such as course structure and content.

Question 28 had somewhat less specific aims and asked the students to indicate on a
pre-printed card whether they agreed or disagreed with the listed comments on how
the course was taught. The list was compiled using the responses made by the
students in the first survey and the results are shown in Table 4.6.

In considering the results in this section, it is worth bearing in mind that relatively
few students actually saw their lecturers or tutors outside class-times, and those who
did did so very rarely (17% said they saw their lecturer and 22% their tutor out of
class, and this only once or twice a term. See question 26 in Appendix 3 for details).
Therefore, in assessing possible gender differences in attitudes towards teaching staff
and methods, one must take into account the fact that most of the students in this
survey had little, if any, contact with the staff outside of the formal lecture or tutorial
situation.

Overall 37% of the respondents said that the mathematics lecturer was very
important in determining how much they enjoyed a particular topic, 49% said s/he
was fairly important and the rest said s/he was unimportant.

In Table 4.5, the quality mentioned most often as very important for a lecturer to
have was clarity. The next most important quality was understanding the problems
students might have, followed by enthusiasm and a confident lecturing style, and an
interesting presentation of the subject. A helpful attitude was also considered very
important by a small majority of the students.
Table 43
Question 24, survey 3
students' ratings of the importance of a mathematics lecturer's qualities

percentage of students rating the following

qualities as very important
all women men

CLarity 95 95 95
:nderstanding students' problems 76 90 62
Enthusiasm 66 33 76

C:nfident lecturing style 66 30 81

:nteresting presentation of topic 63 75 52
Ee1pful attitude 59 43 57

Pleasant manner 17 20 14

Easy to gain access to 29 30 29

Being good at research 5 3 5

N=41 N20 N=21

Table 4.6
Question 28, survey 3
mathematics students' attitudes towards the way their course is taught

percentage agreeing with the following

comments on the teaching of the
university mathematics course
all women men

Challenging 88 85 90

Impersonal compared to school 76 90 62

Presented too quickly and not enough

time given to assimilate 63 80 48

Not enough examples or explanations* 61 85 38

Presented as too abstract 56 65 48

Encourages exploration of the subject 39 35 43

Uninteresting 34 35 33
N=41 N=20 N=21

* pc0.01 on a ZR test with 1 degree of freedom for the difference

between women and men.
There were two notable gender differences in the responses to question 24. Out of 20
women and 21 men, 10 and 17 respectively (50% and 81%) rated a confident
lecturing style very important, and 18 women compared to 13 men (90% and 62%)
considered as very important an understanding of students' problems (x2
tests on
these frequencies with 1 degree of freedom give p=0.08 in both cases).

A majority of the students (68%) said they would like more feedback from the staff
on how they were doing. Again them was a notable gender difference with 17
women and 11 men (85% and 52%) giving this response (p=0.06 for a X 2 test with 1
degree of freedom).

Table 4.6 shows that, for question 28, most of the students agreed that the course
was challenging but impersonal compared to school (88% and 76% respectively).

There was one significant gender difference for question 28, with women being
significantly more likely to agree that there were not enough examples or
explanations given. In addition, there were two notable gender differences in the
responses. The women showed a greater tendency to say that the university course
was impersonal compared to school and was presented too quickly (x2
tests on these
differences with 1 degree of freedom give p=0.006, p=0.08 and p=0.07 respectively).

While there was little evidence in the above results that women and men differed in
the importance they attached to certain personal qualities of the lecturer, there did
seem to be some support for the notion that the way mathematics is taught at
university affects women and men differently. The responses shown in Table 4.6 did
suggest, although rather tentatively, that the women found it more difficult to adapt
to university teaching (or were more likely to admit to it). However, it mustagain be
pointed out that the strength of such support is somewhat restricted by considerations
such as sample size and probability of Type II errors. Unfortunately the sample was
too small to allow a reasonable control for ability, and so it was not practical to
assess to what extent the sex effects noted in Table 4.6 could have been attributed to
a slightly lower mathematical ability (as measured by examination performance)
among the women in this particular sample as compared to the men. If this were the
case, there would be no support in either table for a real gender difference in
responses to these questions.

On the other hand, the lack of any significant gender difference in rating a lecturer's
qualities was not necessarily inconsistent with previous research findings that
women regard the 'human dimension' as more important and my own finding that
women showed a greater responsiveness to personal encouragement at school. It
could simply be that in responding to question 24, all the lecturers or lecture
occasions were viewed as an undifferentiated whole. Although both sexes rated
certain personal qualities highly, it was not clear to what degree they felt that any of
their lecturers actually possessed these qualities, or indeed whether women and men
differed in their attribution of these qualities to the lecturers concerned. If, for all or
some of the respondents, the responses to question 24 represented a portrayal of
desirable qualities in an 'ideal' lecturer, then the lack of any sex effect appears less
inconsistent with previous research.

Nonetheless, the rather low ratings given by both sexes to such qualities such as
'pleasant manner' and 'helpful attitude' suggest that whether or not there is any
substance to the hypothesis that women are more affected by the 'human dimension'
of a situation, among the mathematics students in the sample this dimension was less
important than qualities relating to good communication skills. Both sexes appeared
to have a fairly instrumental attitude to the qualities required of a mathematics
lecturer, clarity being the most highly rated. Insofar as the gender differences in
Table 4.5 were at all meaningful, it could be argued that the quality with the highest
difference in favour of women ('understanding students' problems') is the one
closest to the notion of being encouraged at school (Section 2.3.2). The quality with
the highest difference in favour of men ('confident lecturing style') reflects a more
instrumental attitude among men - a confident lecturer being one who will succeed
in reducing the uncertainty surrounding mathematics and its use. Of course these
interpretations are wholly speculative, and it would require a much deeper
psychological investigation to ascertain whether they are well-founded.

4.3.2. Attribution and motivation

Question 35 was concerned with students' attributions of the causes for

underachievement in examinations. Firstly, question 35(a) asked whether the
student had ever felt that s/he had not done as well in an exam as s/he would have
liked. If the student responded positively, s/he was asked in question 35(c) to
indicate on a pit-printed card the reasons s/he had not done well. The list of reasons
was drawn up using the results of previous research on causal attribution (Frieze
1975). There has been some research suggesting that women might be more likely
 M

than men to attribute underachievement to internal factors, such as lack of ability,

which would lower their expectation of future success (Frieze 1975; Seward &
Seward 1980). However, other studies have not shown any positive evidence of this,
and it was therefore not a particularly strong hypothesis. The results can be seen in
Table 4.7.

As shown in Table 4.7, the most often mentioned reason for lack of success in
examinations was not having done enough work. The next most common reason was
examination difficulty.

There was a significant difference between women and men in attributing lack of
success to lack of ability, with the women being more likely to do so. This result
supports the hypothesis that women tend to attribute failure to internal stable causes
which are not likely to change and over which they have little control. One would
therefore expect such an attribution pattern to have a negative effect on the student's
expectations of future success. One should note that the effect can be seen as cyclic
and self-fulfilling, since it seems reasonable to infer that a low expectation of
success due to lack of confidence would probably result in attributing failure to
internal factors. It is therefore not particularly constructive to try to specify cause
and effect in this case. As discussed in the previous section, controlling for
achievement as well as gender was considered impractical and therefore it was
uncertain to what extent the above gender difference could have been attributed to
differences in ability in this case. However, the achievement patterns of the students
and the links between achievement and attitude are examined in Section 4.4.

In question 39, the students were asked to indicate on a pre-printed card the factors
they felt encouraged them in theft studies. The list of factors was compiled using
findings from previous research on motivation for achievement which appeared to
imply that women might not be motivated by the same considerations as men, these
having to do with aspects of mastery and competitiveness (Hoffman 1975; Seward &
Seward 1980).

The factors most often mentioned as being encouraging were getting the gist of the
subject as a whole (mentioned by 95%), and succeeding at something seen as
difficult (93%). These were followed by past success (88%), the hope of future
success (85%) and comments by staff (78%). Sex differences in these rates were on
the whole very small (Appendix 3). Therefore, these results did not confirm the
hypothesis that women might be more motivated by social considerations (such as
Table 4.7
Question 35(c), survey 3
Mathematics students' causal attribution of perceived lack of success

percentage attributing lack

of success to each reason
all women men

Not enough work 74 67 60

Examination difficulty 63 67 60
Not having studied a
particular topic 53 56 50
Lack of ability* 29 50 10

Sad luck 26 17 35

Lack of interest 21 33 10

Not feeling well 5 11 0

N=38 N=18 N=20

* p<0.02 on a x2 test with 1 degree of freedom for the difference

between women and men.

(I
 me

comments by staff) and less so by appeals to their sense of mastery (succeeding at

something seen as difficult).

4.3.3. Learning styles

This section presents the results of the questions which did not strictly pertain to the
concerns examined in Sections 4.3.1 and 4.3.2, although they did examine some of
the same considerations indirectly. These remaining questions are therefore loosely
grouped under the term 'learning styles'.

Some of the previous research had suggested that women were more affected by
interaction with others than men were (Hoffman 1975; Seward & Seward 1980), and
many of the following questions attempted to examine potential ramifications of this
hypothesis in the context of the university course structure.

In question 32, the students were asked to indicate on a pre-printed card which of the
specified considerations would influence them to attempt an exercise on a
mathematics worksheet. The aim was to assess how influenced students were by
external pressures, and what sorts of internal factors might motivate the students'
choice of exercises.

93% of the students said that they would be influenced to attempt an exercise if it
was a hand-in question. The next most frequently mentioned reasons were thinking
they could do it (90%), thinking it seemed important for the course (85%) and
thinking it looked interesting (76%). While 68% overall agreed that thinking a
question looked easy would influence them to try it, the women were notably more
likely to say this. Out of 20 women and 21 men; 17 and 11 respectively (85% and
52%) made this response (p=0.06 for a X 2 test on these frequencies with 1 degree of
freedom), and it was the only notable difference found for this particular set of
questions. Again, such a finding might be interpreted as indicating a lack of
confidence on the part of the women, that is, assuming the men were being accurate
about their motivation in this case. An alternative view might possibly be that the
men were not admitting to being influenced by the question appearing easy because
they did not wish to downgrade their ability. This interpretation will be discussed in
more detail in Section 4.3.4.

Questions 33 and 34 examined the students' reasons for and reactions to failing to
 91

complete an exercise. The results for question 33(a) are not shown here since it was
felt that they did not provide any particularly useful information. Question 33(b)
asked whether they sometimes felt they could have completed the exercises if they
had had more time, or less other work to do. This response was seen as a measure of
how confident the student was of her/his ability to eventually solve the problem.
Results from the APU surveys implied that the men might express more confidence
in this respect (APU 1985); however, there was no notable gender difference. About•
half the students (44%) agreed that they might have completed the questions they
had given up on if they had had more time.

Question 34 asked what they tended to do when they decided they could not do an
exercise: forget about it, look up the solution sheets, ask their classmates, ask their
tutor or lecturer. Again, the responses were not exclusive. They were also asked
what they used to do at school in the same situation. The aim of these questions was
to examine to what extent the students invoked the help of others when they were
having difficulty, as opposed to the more impersonal method of consulting the
solution sheets.

For both sexes, the most common reaction to failure was to look up the solutions
(90% of students made this response). Consulting classmates was mentioned by 71%
and 59% said they asked their lecturers or tutors.

In addition to the above questions, the students were asked whether they ever
guessed at an answer in a university mathematics examination without being sure
(question 36), if they tended to go over the questions after an examination to see
what they had done wrong (question 37), and whether they compared their
performance in mathematics to others in their class (question 38). The responses to
these questions were seen as possible indirect measures of exploratory and
risk-taking behaviour, motivation to succeed in a 'masculine' domain (seen as
implicit in the importance accorded to examinations), and competitiveness. All these
attributes have been traditionally considered to be more in evidence among men than
among women (Stein & Bailey 1975; Seward & Seward 1980; APU 1985; New
Scientist 1988), and it was therefore hypothesised that the men would be more likely
to respond positively to the questions.

However, the majority of the students responded positively to all three questions (the
percentages being 78%, 73% and 80% respectively), with no notable gender
differences.
 92

4.3.4. Conclusions

The data did not clearly support the theory that women may show a relative lack of
confidence in their ability in fields considered 'masculine', such as mathematics
(Stein & Bailey 1975). While they did mention lack of ability as a reason for
underachievement significantly more often than the men, there was no notable
difference between the proportions of women and men admitting that they guessed at
answers in examinations, and both sexes expressed similar confidence in being able
to complete questions given enough time. It would therefore seem that the APU's
(1985) findings regarding males' greater propensity for risk-taking behaviour and
confidence when tackling questions are not supported in this case. However, as
mentioned previously in Section 2.3.5, all the students in the sample were
self-selected to some extent regarding their confidence to do a mathematics degree,
and it is therefore perhaps not surprising that they do not display gender differences
in attitude such as those found by the APU.

Considering the responses to the question on factors affecting the students'

motivation to attempt an exercise on a worksheet (question 32), there was no
evidence that women were more influenced by the opinions of the staff (ie. whether
the exercise was a hand-in or recommended question). They did seem more affected
by how they perceived their ability to do the question, but the responses in this case
varied with the wording of the question: while there was very little difference in the
proportions of men and women saying they were motivated by the thought that they
could do the exercise, notably more women said that they were motivated by the
question looking easy.

One possible, though speculative, interpretation of this response pattern might be

that the men were less willing to admit that they were motivated by the exercise
appearing easy because this could give a negative impression of their ability. It
might imply that success in the question could be attributed to lack of difficulty
rather than the mathematical ability of the student, thus minimising the value of the
success. This interpretation would be in line with the reasoning of researchers such
as Chodorow (1978) and Easlea (1981) who argue that males may not be willing to
admit to weakness in certain domains which are perceived as 'crucial both to the
definition of masculinity and to a particular boy's own masculine gender identity'
(Chodorow 1978, p.181). Traditionally, mathematical reasoning is considered to be
one of these domains (Easlea 1981; Walkerdine 1989). In a rather perverse way,
 93

admitting to difficulty can be a form of confidence: the confidence that recognising

one's shortcomings will not prove excessively detrimental to one's sense of self.
Again, further work is needed in order to disentangle what is being said from why it
is being said.

Although the women students showed no obvious inclinations to be more motivated

by social factors with regard to their work-patterns, they did tend to attribute greater
importance to lecturers having good interrelational skills and expressed a notably
greater desire for more feedback from the staff. It is possible that despite the
impression that students did not see the teaching staff as particularly involved with
their course-work outside the lecture or tutorial situation, during the relatively
restricted periods when there were possibilities of student/staff interaction, this
interaction assumed a somewhat greater importance for the women. However, the
evidence was very weak in this case.

Some of the findings discussed in this section, such as attributing failure in

examinations to lack of ability (question 35) and thinking that an understanding of
students' problems is a very important quality for a lecturer to have (question 24),
could be considered typical of relatively low achievers (Frieze 1975). These might
be likely to have little confidence in their abilities and therefore feel the need for the
support others could provide. In order to ascertain whether this theory is justified,
the university achievement of the Edinburgh mathematics students and the links
between achievement and attitude for the students in Survey 3 are examined in the
next section.
 94

4.4. Achievement and attitude at Edinburgh University

In Section 2.2.21 described the gender differences in performance observed in the

secondary public •examinations. The figures showed that while the overall pass-rates
were similar for girls and boys, proportionally more boys obtained grade A for
Higher and A level Mathematics. The situation was similar at university level, with
14% of male and 10% of female graduates obtaining Mathematics First Class
Honours degrees in 1983 (Royal Society 1986). This trend was also apparent in the
1986 survey of Edinburgh University mathematics graduates described in Section
1.2.1 (Fraser & Cormack 1987). The academic achievement of the mathematics
undergraduates was examined in the subsequent surveys, and the results of those
studies are presented in section 4.4.1. Section 4.4.2 deals with the links between
attitude and achievement observed for the mathematics students in Survey 3.

Assessment of the Edinburgh University Mathematics Honours students during the

first two years consists of two class examinations, one at Christmas and the other at
Easter, and a degree examination at the end of the Summer Term in June. The results
of the class examinations do not affect a student's advancement to the next year, but
are considered in conjunction with the results of the degree exam when awarding
Merit certificates for the year.

In order to pass the first two years, the students nominally need to obtain a score of
at least 50% in the degree examination. A Second Class Merit certificate is awarded
when the average score of the two class examinations and the degree examination is
between 65% and 75% approximately. An average score of at least 75% is needed
for a First Class Merit. The cut-off points for the award of Merit certificates may
vary somewhat from year to year according to the students' performance.

At the end of the third year, the students sit part-final examinations, the results of
which count for 317ths of the marks for the final Honours degree results. These
results are banded into First Class, Upper Second Class, Lower Second Class and
Third Class.

4.4.1. Achievement of the mathematics students

For Survey 1, there was no evidence of consistent sex differences in achievement for
the selected sample. Due to inaccuracies in the students' reports of theft
 95

performance, I was not able to examine the relation between performance and
attitude for the achieved sample, and it therefore did not seem useful to present a
detailed analysis of the students' performance. The relevant figures can be found in
Appendix 1 (question 15).

The results on achievement in Survey 3 were not affected by inaccurate reporting by

the students as the design and scale of the survey permitted detailed verification of
the examination results. It was also easier to compare performance since the students
were in the same year and had thus sat the same examinations.

The students in Survey 3 were asked a series of questions on past and anticipated
performance in the class and degree examinations. In question 40, they were asked
what grades they had got for the Mathematics 1A class and degree examinations and
whether they had been awarded a Merit certificate for the 1A course. They were also
asked whether they had expected to get a Merit and what grade they had obtained in
the Mathematics 2A December class examination.

Question 41 examined the students' expectations and asked what they thought they
might get in the Mathematics 2A class examination in March and the degree
examination in June. In addition, they were asked whether they expected a Merit
certificate for the 2A course. This question was inspired by previous research which
indicated that males tended to have higher expectations of success in certain areas,
such as mathematics, despite there being little difference in subsequent performance
(Seward & Seward 1980; APU 1985). The actual grades and Merits awarded were
obtained after the 2A June examinations in order to verify the students' predictions.

I decided to use the award of Merit certificates as the measure of achievement since
this took into account overall performance during the year, as opposed to a one-off
measure such as the result of the degree examination. However, the expectations of
Merit awards were found to be difficult to interpret since some students seemed to be
influenced by modesty and did not make a clear reply when asked whether they
thought they might get a Merit for Mathematics 2A ('I hope so', for instance). It
therefore seemed simpler to study expectations by looking at the predicted letter
grades for the 2A degree examination. The method used in this case is described in
Appendix 3.

Tables 4.8a and 4.8b show the percentages of each group of students (women, men,
all) awarded Merit certificates for the Mathematics 1A and 2A courses. The
Table 4.3
Question 40, survey 3
Performance of the second year mathematics students

a) Mathematics 14
percentage obtaining:
1st class Merit 2nd class Merit no Merit
women 11 3] 56
ten 52 14 33 N=2l

all 33 23 44 N=39

* ;c0.03 on a x2 test with 2 degrees of freedom for the difference

between women and men.
two women were direct entrants to Mathematics 2A and had not taken
Mathematics 1A.

Mathematics 24
percentage obtaining:
1st class Merit 2nd class Merit no Merit
women 10 15 75 N=20

men 33 19 48 N21

all 22 17 61 N=41

Table 4.9
Question 41, Survey 3
Mathematics students expectations of their 2A degree examination results

percentage of students
underestimating overestimating obtaining grade
obtained grade obtained grade predicted
women 22 56 22 M=18

men 15 50 35 N=20

all 18 53 29 N=38*

* two women and one man said they did not know what grade they might obtain.
 97

students' predictions for their 2A degree examination results are presented in Table
4.9 as the percentages of each group underestimating their letter grade,
overestimating it and obtaining the predicted grade. Again, it must be pointed out
that the small sample size makes the use of percentages somewhat misleading.

Table 4.8a shows that the men in the sample did significantly better than the women
in obtaining First Class Merits for the 1A course. However, there was no notable
difference in performance for the 2A course (Table 4.8b). In addition, there were no
notable differences between the proportions of women and men obtaining Merits of
any class for either IA and 2A. These results were similar to those from Survey 1 in
failing to provide evidence of a consistent difference in performance between
women and men.

There were no notable differences in the proportions of women and men

overestimating or underestimating their 2A degree examination results (Table 4.9).
The findings from the APU surveys that the boys tended to have exaggerated
expectations and confidence regarding their future performance were therefore not
confirmed in this case.

4.4.2. Achievement and attitude: Survey 3

Since the data on achievement from the mathematics students in Survey 3 were
checked for accuracy, it seemed a worthwhile exercise to analyse some of the
questions on attitude using achievement as the independent variable rather than
gender. As the sample was fairly small, achievement was assessed in terms of
obtaining or not obtaining a Merit certificate, irrespective of class.

I decided to consider the award of a Merit certificate for Mathematics 1A as the

performance indicator rather than the award of a Merit for 2A, since it was thought
that the students' explicit awareness of their past performance might be a strong
factor affecting their confidence and attitude. In fact, using the award of a Merit
certificate for 2A as the performance indicator did not affect the results to any great
extent. The initial theory was that high achievers would have a more positive
attitude towards the mathematics course (Bell et at. 1983), and it seemed reasonable
to hypothesise that such a difference in attitude would be found in the students'
perceptions of the way the course was taught although I was not sure how this would
be expressed. It was also thought that high achievers would show a lesser tendency
to attribute lack of success in examinations to lack of ability (Frieze 1975), thus
expressing a higher level of self-esteem regarding their mathematical ability.

The survey questions have already been described in previous sections and therefore
I will not describe them in great detail. I will merely list the ones selected for this
analysis.

Question 8: rating of attitude towards school mathematics (Table 4.10a).

Question 16: rating of attitude towards the 1A course (Table 4. 10b).
Question 17: rating of attitude towards the 2A course (Table 4. lOc).
Question 28: attitude toward the way the mathematics course is taught (Table-4.1 1).
Question 35(c): causal attribution of underachievement in examinations (Table 4.12).

Questions 27, 36 and 38 were also analysed, but no significant or notable differences
between the two achievement groups were found.

The results of the ratings of school mathematics for Merit and non-Merit students are
shown in Table 4.10a as the means of each group's rating scores for the three
dimensions (difficulty, interest, usefulness). The difference of the means between the
two groups is given for each dimension, as well as the standard error of the
difference of the means. The rating scales are shown in the table. The ratings for the
Mathematics 1A and 2A courses are presented in the same format in Tables 4.10b
and 4. lOc respectively and any significant results are indicated in the tables.

While Table 4.10a shows relatively little difference in attitude towards mathematics
at secondary school between students who obtained Merits for 1A and those who did
not, Tables 4. lOb and 4. lOc indicate significant and consistent differences in attitude
towards the 1A and 2A mathematics courses (apart from the difference for the 2A
rating for difficulty, which is not notable and rather small). So non-Merit students
showed strong tendencies to rate the 1A and 2A courses more negatively than the
Merit students, a result which confirmed the preliminary hypothesis.

It might be noticed that the differences in attitude towards the Mathematics 1A and
2A courses were larger and more consistent for the different achievement groups
than they were for women and men (see Tables 4.1b and 4.1c in Section 4.1.2 for the
gender differences in attitude). One might tentatively suggest that in this case,
achievement appeared to be a better predictor of attitude at university than gender,
especially since there was no notable difference between the proportions of women
Table 4.10
;.escions 8, 16 and 17, Survey 3
Ratings of school and university mathematics courses by achievement

S:aies used: Difficulty: 1 very easy 5 very difficult

Interest: 1 very interesting very boring
Usefu!ness: I very useful S a waste of time
Difficulty Interest Usefulness
Secondary school mathematics

;erit students 2.14 2.46 2.18 N=22

:on-Merit students 2.12 2.06 2.53 N=17

:ifference of neat scores 0.02 0.40 -0.35

Standard error of difference 0.29 . 0.28 0.27

University: Mathemadcs lÀ

erit students 2.73 2.78 2.41

Ncn-Merit students 3.47 3.53 3.12

flfference of mean scores _0.74* _0.75* _0.71*

Standard error of difference 0.26 0.27 0.31

C) University: Mathematics ZA

Merit students 3.64 2.64 2.55

Non-Merit students 3.88 3.53 3.41

tifference of mean scores -0.24 -0.89 _0.86*

Standard error of difference 0.21 0.31 0.33

cc0.05 on a Mann-Whitney test for the difference between Merit and

non-Merit students.
 Oro

Table 4.11
Question 28, survey 3
Differences between Merit and non-Merit mathematics students'
attitudes towards the way the mathematics course is caught

percentage agreeing with the following

comments on the teaching of the
university mathematics course
Students with Non-Merit
Merits for 1A students
Challenging 91 82
impersonal compared to school 59 94
Presented too quickly and not enough
time given to assimilate 50 76
Not enough examples or explanations 50 71
Presented as too abstract 36 82
Encourages exploration of the subject 36 47•

uninteresting 14 53
N=22 N=17

Table 4.12
Question 35(c), survey 3
Differences between Merit and non-Merit mathematics students
causal attribution of perceived lack of success

percentage attributing lack

or success to eacn reason
Students with Non-Merit
Merits for IA students
Not enough work 80 65
Examination difficulty 50 76

Not having studied a

particular topic 45 59
Lack of ability 15 47

Bad luck 15 35
Lack of interest 20 24
Not feeling well 5 6
N20 N=l7
 101

and men among students obtaining Merits for 1A. Of course, the evidence is very
weak and it would have been more appropriate to analyse the questions in this
section controlling for both variables. However, the small sample size meant that
such detail of analysis was somewhat unrealistic considering the techniques I was
using.

The data from questions 28 (Table 4.11) and 35(c) (Table 4.12) are presented as the
percentages of each group agreeing with the specified comments on how the
mathematics course is taught and the reasons given for not having done as well as
hoped for in mathematics examinations.

There were several notable differences between Merit and non-Merit students in
their attitude toward the way the course was taught (Table 4.11). Non-Merit students
found it more impersonal compared to school; presented as too abstract and
uninteresting. The p-values for x2 tests with 1 degree of freedom on these
differences were respectively p=0.03, p=O.Ol and p=0.02. Since previous research
had not really permitted me to predict these particular results, I adopted a more
stringent level of significance in this case in accordance with the methodology
described in Section 1.2.3, and therefore these results are merely notable rather than
significant.

There was also one notable difference for the causal attribution of perceived lack of
success (Table 4.12), with non-Merit students showing a greater tendency to blame
lack of success in examinations on lack of ability (a X 2 test with 1 degree of freedom
gives p=0.08). There was therefore some rather weak indication that the Merit
students were more likely to attribute failure to causes which did not reflect
negatively on theft ability. In this respect, the differences between Merit and
non-Merit students were similar to those observed between women and men (see
Table 4.7 in Section 4.3.2).

4.4.3. Conclusions

The above findings did indicate that there was a rather strong relation between
achievement and attitude, as measured by numerical ratings of various aspects of the
students' mathematics course. However, the evidence was very weak for the more
qualitative measures of attitude in questions 28 and 35(c). There was also no strong
evidence of consistent differences between the performance of the women and men
 IL,!1

in Surveys I and 3, and no evidence that the men in Survey 3 had exaggerated
expectations of success compared to the women.

The size of the sample in Survey 3 made it difficult to analyse the responses to the
questions while controlling for both gender and achievement-linked differences, and
it is therefore not clear how these two variables interact to influence the response
patterns. In retrospect, I should perhaps have used a somewhat more sophisticated
technique, such as multiple regression, for the analyses in this section. However,
time constraints did not permit me to undertake the additional analyses. I am
therefore reduced to merely pointing out that the findings in Section 4.4.2 taken in
conjunction with those in Section 4.3 do suggest the necessity of controlling for
achievement when interpreting findings from studies of gender differences in
attitude.
 103

4.5. Performance at Scottish and English universities

In order to obtain a wider view of achievement patterns in mathematics at university,

data were requested from the Universities' Statistical Record (IJSR) on the final
degree results of university graduates in Mathematical Sciences in Scotland and
England for the years 1985-1987 inclusive. The nature of the classifications used for
the data has already been described in Section 3.2.2. The data used in this section
were originally broken down by year, university, class of degree and sex.

The data from Cambridge University were not used in the following analyses due to
the fact that Cambridge does not classify its degrees in the normal way. The USR
returns from Oxford University for 1985 had been misclassified and so the Oxford
results were for 1986 and .1987 only. The actual figures are shown in Appendix 5 for
reference.

Considering the differences in girls' and women's participation in mathematics

between Scotland and England, both at secondary school and university (see
Sections 2.2.2 and 3.2.2), a comparison of degree performance between the Scottish
and English universities seemed interesting. In the interests of brevity, the terms
Scottish and English in this section refer to graduates from Scottish and English
universities respectively, regardless of original domicile.

The overall degree results are shown in Table 4.13 as the percentages of each group
(all, women, men) obtaining each class of Mathematical Science degree (excluding
Computer Science). It can be seen that men's performance was slightly more
extreme than the women's, with proportionally more men obtaining First Class,
Third Class and Pass/Ordinary degrees. This is a pattern which has been found in sex
differences in educational achievement at various levels (Willms & Kerr 1987;
Clarke 1988). The percentages of women and men awarded Upper Seconds were the
same, and proportionally more women were awarded Lower Seconds. However, it
must be pointed out that the sex differences in this case were rather small on the
whole, although the different distribution between classes for women and men was
highly significant due to the large numbers involved (pc0.0001 for a X 2 test with 4
degrees of freedom).

Table 4.14a shows the percentages of each group obtaining each class of degree for
Scottish and English universities. Since the Scottish graduates were rather more
 104

Table 4.13
Degree results for Mathematical Sciences graduates from
Scottish and English universities for 1985-:?7 inclusive
(excluding Computer Science graduates and Cambndge graduates)

percentage obtaining each class of degree

1st upper Lower 3rd Pass/ordinary
2nd 2nd
all 16 27 32 13 8 N=7169

women 15 27 35 17 6 N=2480
men 16 27 30 12 8 N=4689

Source: Universities' Statistical Record.

 105

likely to obtain Pass/Ordinary degrees than the English graduates, it was considered
more appropriate to conduct a separate analysis in order to compare Honours degrees
(Table 4.14b). The differing natures of the Scottish Ordinary degree and the English
Pass/Ordinary degree merit some comment in view of the differing proportions of
Scottish and English graduates awarded Pass/Ordinary degrees, and I will discuss
this after the presentation of the main results regarding sex differences in degree
performance.

As far as obtaining Pass/Ordinary degrees went (Table 4.14a), the English graduates
displayed the same sex difference in performance patterns observed in Table 4.13.
The sex difference for the Scottish graduates was smaller and not significant.

When Honours degrees were considered separately (Table 4.14b), the gender
differences for English graduates were similar to those already commented on in
Table 4.13, small despite being highly significant (pc0.0001 for a X 2 .tst with 3
degrees of freedom). Scottish graduates showed a more marked sex difference in the
attainment of First Class Honours, but the overall sex difference was not significant
due to the smaller numbers. It may be noticed that the main difference between
graduates from Scottish and English universities was that a larger proportion of
Honours graduates from Scottish universities were awarded First Class degrees
(pcz0.005 for a X2 test with 3 degrees of freedom). This was due to a fairly high
proportion of male Scottish graduates in the top achievement band compared to male
English graduates. The percentages of Scottish and English women graduates
obtaining First Class degrees were reasonably similar.

The differences between Scottish and English graduates were striking enough to
warrant some discussion. One possible, though rather simplistic, explanation for the
difference found between the proportions of graduates from Scottish and English
universities obtaining Pass/Ordinary degrees might be lower standards of entry in
Scottish universities (Section 3.2.2). However, the links between performance in the
secondary school public examinations and at university are somewhat tenuous
(Clarke 1988) and probably not strong enough to account for the difference in
performance patterns. In addition, such an explanation does not take into account the
fundamentally differing structure of the Scottish and English degree courses, such as
the longer duration and later specialisation of the former.

Therefore, an alternative, and to my mind more plausible, explanation is that the

USR data show that Scottish students tend to attend Scottish universities, and these
 106

Table 4.14a
Degree results for Mathematical Sciences graduates
in England and Scotland for 1985-1987 inclusive
(excluding Computer Science graduates and Cambridge graduates)

percentage obtaining each class of degree

1st upper Lower 3rd ?ass/Ordinary
2nd 2nd
England all 15 27 33 18 6 N=6364

women 15 27 37 17 4 N=2148
men 16 27 31 19 7 N=4216

Scotland all 18 22 25 13 22 N805

women 14 23 27 15 21 N=332
men 21 22 23 11 22 N=473

Table 4.14b
Degree results for Mathematical Sciences Honours graduates
in England and Scotland for 1985-1987 inclusive
(excluding Computer Science graduates and Cambridge graduates)

percentage obtaining each class of degree

- 1st upper Lower 3rd
2nd 2nd
England all 16 29 35 19 N=5990

women 15 28 38 18 N=2067
men 17 29 34 20 W=3923

Scotland all 24 29 31 16 N=629

women 18 29 34 19 N=261
men 27 29 30 14 N=368

Source: universities Statistical Record.

 107

students might not regard a Pass/Ordinary degree as a sign of failure to the same
extent that the English students do. Traditionally, it would appear that within the
Scottish system taking an Ordinary degree can be seen as a positive choice
(McPherson 1972), whereas in the English system, a Pass/Ordinary degree is a failed
Honours degree. Therefore students in Scottish universities might show a greater
tendency to opt for Pass/Ordinary degrees if their circumstances are not entirely
favourable to the pursuit of an Honours degree. One might speculate that the length
of degree courses in Scotland could be an influencing factor in this case, since some
students might think that three years is enough time time to spend in an academic
institution and want to begin a career as soon as possible. This did appear to have
previously been the case for students entering teacher training, although more
recently there has been a drop in the numbers of students taking this option (Burnhill
& McPherson 1983).

The difference between Scottish and English universities regarding the award of
First Class degrees would not appear to have an obvious explanation. Again, it may
be that standards are lower in Scotland, or the marking more generous. It could also
be that the four-year course allows more time to assimilate the course material and
so improves degree results, although why this should apparently affect the men more
is unclear. However, these theories are purely speculative at present due to lack of
more detailed data It is also difficult to assess how the English results were affected
by omitting the Cambridge data, since the Cambridge intake is relatively large and
highly selected (see Appendix 4).

Conclusions

The overall picture of performance in terms of degree results is that there was a
small difference between the performance of women and men at either end of the
achievement band, with men slightly overrepresented. The complementary
difference in Second Class degrees appeared entirely in the award of Lower Second
Class degrees where women were overrepresented. Of course, the data reflect a
predominantly English pattern since about 89% of the graduates were from English
universities. The observed pattern is one which appears in a wide range of degree
subjects and is commented on at length by Clarke who attributes the differences to
the influence of social and institutional pressures (Clarke 1988). In this specific case,
however, it might be argued that the observed differences are too small to warrant
much in the way of elaborate explanation.
 P.
The discrepancies in performance between the men Honours graduates from Scottish
universities and those from English universities regarding the award of First Class
Honours degrees are somewhat puzzling in the context of this study. This
discrepancy makes it difficult to assess the potential impact of women's higher
participation rate in mathematics education at Scottish universities (Section 3.2.2) on
gender differences in degree results. It is nominally clear that Scottish men are
overachieving relative to English men, but it is less clear what the position of the
Scottish women is. However, the differences between Scottish women and men were
not significant and it is possible that the somewhat unusual performance pattern
displayed by Scottish men was due to random variation. Further study would be
necessary to determine whether this is actually the case. Until we have built up a
reasonably accurate picture of degree performance in Scotland, any attempts to study
possible links between women's participation and their performance relative to men
remain unproductive.
 109

4.6. Career aspirations among the Edinburgh mathematics students

I thought it interesting to examine the students' career aspirations in order to obtain

some idea of how theft experience of mathematics at university may have affected
theft career intentions. Anecdotal evidence and departmental records had indicated
that most mathematics students took up careers in finance, with few graduates
becoming involved in more industrial applications. It was thought that perhaps one
might find more women thinking about careers in teaching, as a result of the
observed trend for women to be employed in the 'caring' professions (USR 1987),
and for a higher proportion of women graduates to enter education training
compared to men graduates (Weinreich-Haste 1984).

The mathematics students in Survey 1 were simply asked if they had any ideas for
the future (question 27). Originally, this question was intended for the third and final
year students only since it was thought that they would be more inclined to have
considered the question of a career. However, in the end the question was asked of
all the students, apart from the first ten first year students interviewed. The second
survey of mathematics students included a question examining whether the students
wanted a job using mathematics, what type of job they wanted, and whether they
thought they would get one (question 22, Survey 3). In addition, they were asked if
they had any idea what they wanted to do after they graduated (question 42).

The responses to question 27 in Survey 1 were classified into the categories shown
in Table 4.15. For this question, the students could make more than one response.
S

The sample in Survey 3 was rather small and the results similar to those of Survey 1,
therefore the details of the findings for question 42 in Survey 3 are not discussed to
any great extent in this section (see Appendix 3).

The results for question 27 showed only one significant gender difference, which
was for mentions of careers in academic research. Most of the students who
mentioned a career were considering employment in the financial sphere, and there
was no notable difference between men and women in this respect. In retrospect,
considering Burnhill and McPherson's (1983) findings that gender differences in
Scottish university students' aspirations had diminished substantially over.a decade,
it was perhaps not such a surprising result. Burnhill and McPherson also comment
on the shrinking employment prospects for trainee teachers, and it seems a
reasonable assumption that such considerations had some effect on the students'
 110

Table 4.15
Quescion 27, survey 1
Career aspirations of mathematics studert:s

percentage mentioning
careers in each field
all Women

Finance 41 44
Did not mention a career 27 26

Research 13 0

Other 13 15

Teaching S 8 3

Management 6 10

Computing 4 3 €

N=71 N=39

o<O.Ol on a x2 test with 1 degree of freedom for t:e difference

between women and net.
 111

career intentions since rather few mentioned teaching. There was no difference
between the women and men in this respect.

Regarding the men's greater propensity to consider careen in research, it may be

that the women were somewhat more pragmatic and saw research as being a
relatively insecure area of employment, which is a fairly accurate assessment of the
current situation. Another explanation might be that they were more likely to feel
that they were not capable of a research career in an area they perhaps perceived as
male-dominated and difficult. One hypothesis to consider would be that the rigid
hierarchical structure of academic life, among other things, proves offputting to
those who are not sufficiently committed at a relatively early stage. That men should
show more of this commitment, or single-mindedness, would not be too surprising in
the context of Gilligan's findings relating to women's greater concern with overall
context, and thus perhaps more ambivalent feelings about the advantages of an
academic career. The disadvantages, such as the relative insecurity and inflexibility
of the occupation, might be seen to outweigh any initial interest they may have had.
Various researchers have commented on women's marginalisation in academic
occupations, both economically and experientially (Simon et cii. 1969; Astin &
Bayer 1975; Acker 1984; Taylorson 1984). However, it is difficult at present to
ascertain how such considerations might affect women's initial decision not to enter
the higher levels of academia, and the question would require further detailed study
of the factors influencing career choices.

It is also possible that women might not have seen mathematics as interesting or
important enough to warrant pursuing it further. Certainly the women in Survey I
rated their mathematics course as significantly less interesting than the men did, a
result which would support this hypothesis. It seems plausible that all of these
explanations are valid to various extents, and that the factors involved affected
individuals differently depending on their personality. In view of men's possible
tendency to 'drift' into mathematics (see Section 3. 1), one might speculate that some
of those mentioning a career in research were merely continuing the 'drift' and were
reluctant to leave an environment they felt comfortable with. However, there is no
doubt that women continue to be underrepresented at the higher academic levels in
mathematics (Acker 1984; Chetwynd 1992), and the observed reluctance of women
to consider further research in the field would appear to be a potential factor.

In Survey 3, 83% of the mathematics students said they would like a job using some
 112

form of mathematics, again mainly as actuaries or accountants. 71% thought they

would get such a job and 15% did not know whether they would. When the students
were asked if they had any idea of what they wanted to do after they graduated
(question 42), 61% said they did and the pattern of response regarding types of
careers considered was similar to that in Survey I (Appendix 3), including the fact
that no women mentioned a career in research. However, there were no notable
gender differences for this sample.

It was apparent during the course of the surveys that few of the mathematics students
had explored the full range of opportunities available to mathematics graduates.
None of the students mentioned fields such as operations research, and relatively few
had considered careers in industry or teaching. It is difficult to say whether this was
due to lack of interest or ignorance of the possible options, but it would appear that
the majority of mathematics students see the financial sphere as the most obvious
and lucrative field in which to begin their careers.

4.7. Summary

The few significant gender differences found for Survey I in the mathematics
students' ratings of their mathematics course (Sections 4.1.2 and 4.1.4) showed men
expressing a somewhat more positive attitude. But there were no strong indications
from Surveys 1 and 3 that women and men mathematics students differed much in
their attitudes towards the aspects of their university learning experience discussed
in Section 4.3. When women and men did express notable differences in attitude,
there was a tendency for these differences to provide some support for Gilligan's
(1979) observations that women appear to attribute more importance to the personal
dynamics involved in a situation than men do. However, only two of the gender
differences in Section 4.3 were significant and neither of those was directly relevant
to Gilligan's argument, therefore any interpretation along such lines must remain
tentative.

The data on the mathematics students' achievement showed no consistent gender

differences (Section 4.4), and so one would be hard-pressed to attribute what
differences in attitude there were to differences in measurable ability between
women and men. The data on degree results for Scotland and England over a
three-year period also provided no evidence of noticeably superior performance on
the part of men mathematics graduates, although the large numbers involved did
 113

mean that almost all of the differences discussed in Section 4.5 were highly
significant. On the other hand, there were significant differences in the students'
ratings of their mathematics course between different achievement groups in Survey
3, implying a fairly strong link between achievement and attitude in this particular
case. Again, the relationship was less clear for the more subjective measures, but the
results did nonetheless highlight the importance of taking such interactions into
account in future studies.

There were some indications that, as an academic subject, mathematics still carries
somewhat 'masculine' connotations (Section 4.2). Although the subject was not seen
as male-dominated by an overwhelming majority of the non-mathematics students,
the data indirectly suggested that these students considered it to be more 'masculine'
than strictly neutral. While most of the mathematics students in Survey 3 accurately
described the proportion of women and men in their class to be equally balanced,
there was still 30% who thought that the class was male-dominated. One could see
this as further evidence that despite efforts to change the image of the subject, there
still exists an enclave of opinion which continues to express the traditional views of
mathematics as appealing more to men and thus being more male-appropriate in
some sense. The fact that some of the students in this particular study were
apparently disregarding the evidence of personal experience is a testimony to the
strength and pervasiveness of such views. However, the explanations given by the
mathematics students for thinking or having thought that their class was
male-dominated emphasised the importance of social expectations in influencing
theft views, rather than the belief in innate and immutable differences in
mathematical ability between women and men.

There was very little difference in aspirations between women and men mathematics
students, though significantly more men in Survey 1 did mention research as a
possible option. It would be interesting to examine this tendency further in order to
ascertain why women who have reached a high educational level in mathematics
show a reluctance to continue in the field. It must be pointed out that this pattern is
also found in a variety of academic disciplines. Of course, the majority of
mathematics students of both sexes opted out of considering careers in research for
various reasons, presumably including the perceived lack of sufficient ability.
However, there did seem to be a minority, consisting wholly of men, who managed
to consider the subject important enough to disregard aspects which had already
convinced theft classmates that theft future lay elsewhere.
 114

Chapter 5.
Conclusion

My original intention in undertaking this study was to see whether some of the
supposedly more widespread findings concerning gender differences in mathematics
were replicated in chosen samples which were matched for pre-university
mathematical achievement, and subject choice in the case of the mathematics
students. As one can see, the exercise was a mixed success in this respect since it
mainly showed that the situation is rather more complex than I may have originally
been led to believe. In the course of my research, I have been obliged to question
many of the assumptions inherent in some of the previous studies. Concepts such as
'lack of confidence' and 'affiliation motives' made me somewhat uncomfortable due
to the sense of powerlessness they implied. It felt ideologically unsound to use such
terms without lengthy and complex qualifications, since the 'passive wimp' theory
of femininity appears unjustifiable (Walkerdine 1989). A change of perspective was
needed, a perspective which I 'found initially hinted at in Walkerdine's Counting
girls out. Various aspects of a more workable interpretive framework were further
elaborated in Easlea's Science and sexual oppression, Gilligan's In a different voice,
Spender's Men's studies modified, Segal's Slow motion and Chodorow's Feminism
and psychoanalytic theory. Although these works may seem to represent a range of
views which perhaps do not appear isomorphic to a cursory glance, they have in
common a commitment to the deconstruction of the gendered and value-loaded
meanings our society imposes on such seemingly neutral concepts as 'confidence',
'mathematical knowledge', 'independence', and the list goes on. Despite these
misgivings, I will continue to use some of these terms in the rest of this final chapter,
where I will sum up the main findings of my research and attempt to give a
reasonably coherent synthesis. I hope the more ideologically pure reader will forgive
my linguistic disgressions.

5.1. Confidence

Confidence is a recurrent theme in much of the available work on gender

differences, and is possibly the most often used excuse for the perception of women
as being inferior in mathematics relative to men (Seward & Seward 1980; APU
 115

1985; Walkerdine 1989). However, issues concerning the lack of confidence of one
sex or another are problematic, as I have pointed out earlier in this work (Section
1.1.4). Noted differences in inclination to admit to difficulty between the sexes (APU
1985; Walkerd.ine 1989) constitute a vastly confounding factor in the determination
of what exactly one should interpret as an intrinsic lack of confidence, or indeed if
there is such a beast.

There are several aspects of confidence. One is reflected in expectations of success

at a task. Risk-taking behaviour, such as guessing at answers to a question when one
is not sure of being right, is also usually considered to indicate confidence. My study
did not find any significant or notable gender differences among the mathematics
students in Survey 3 in either expectation of success or risk-taking behaviour
(Section 4.3.3). Expectation of success in this case was measured by whether the
students thought they could have completed questions on exercise sheets they had
given up on if they had more time available. There was also no indication that the
women in this sample were more likely to underestimate their performance at degree
examinations, which would imply that they did not tend to underestimate their
ability, contrary to the APU findings (APU 1985).

The only direct suggestion of a lack of confidence among the women mathematics
students was in response to question 35(c) in Survey 3 where significantly more
women attributed lack of success at examinations to lack of ability (Section 4.3.2).
However, this is possibly a somewhat questionable interpretation in view of
Chodorow's theory that men appropriate domains of power in order to prop up an
elusive sense of masculine identity (Section 1.1.5). Men would therefore be less
likely to admit to weakness, such as lack of ability, which could be a threat to their
sense of self. In this case, mathematical ability might be more important to the men's
self-definition than to the women's, a concept further discussed in the next section.

5.2. Motivation

The finding that the women mathematics students in Survey 1 were significantly
more likely to admit to having been encouraged to study mathematics at school
(Section 2.3.2) raises some interesting questions regarding women's motivations to
study the subject at university. When this result is considered in conjunction with the
observation that these women also showed a notable tendency to give more than one
reason for their choice of degree more often than the men did (Sections 3.1.3 and
 116

3.1.4), there would appear to be some indication that the decision to take a
mathematics degree may have been a more considered choice on the part of the
women. The evidence in this case is rather weak, but does perhaps suggest that
further studies along such lines might prove enlightening.

There are two main themes involved here. The first concerns the nature of
mathematics as it is taught in schools and whether such a representation is less
appealing to women, therefore rendering the decision to study the subject at
university more problematic. The second theme involves the idea that women tend to
base their life-decisions on a wider range of criteria than men do (Gilligan 1979).
The idea of mathematics as a masculine domain has been discussed at length in
Sections 1.1.3-1.1.5 and there seems no need to repeat myself at length. However,
another aspect of the 'masculinity' of mathematics requires some further thought as
it indicates a link between the two themes.

The appeal of mathematics for boys and men can be seen in the context of
Chodorow's theory that men's sense of identity tends to be more fragile than
women's (Chodorow 1989), as discussed in Section 1.1.5. Mathematics offers a
secure domain, one which appears to combine power and certainty because of the
prestige of the subject within our society and the reductionist and positivist way in
which it is presented and perceived. I use the term appears in this case for reasons
discussed in Section 1.1.5 concerning the fundamental uncertainty inherent in
mathematics and science. This uncertainty is not apparent in the type of
mathematics traditionally taught in schools, or even at undergraduate level in most
universities. Mathematics was historically seen as being the key to knowledge of the
universe, a powerful concept indeed. The implication of Chodorow's arguments is
that such perceived power and security are more important for men, albeit perhaps
on a subconscious level, since admitting to a need for secure domains, even to
oneself, could indicate weakness. This would fit in with Gilligan's argument that
women make more considered decisions and my tentative observation of a somewhat
more conscious decision-making process among the women in Survey 1.

In such a context, the women's greater awareness of, or inclination to admit to,
having been encouraged to study mathematics at school could also be seen as
according a greater in of validity and importance to the opinions of others,
something that men may find more difficult and ego-threatening.
 117

5.3. Mathematics students' attitudes towards their course

As presented in Section 4.1, there were some indications that the women
mathematics students in Survey 1 held more negative opinions of the university
mathematics course than the men did, confirming to some extent the APU results
(APU 1985). This would appear to suggest, although only tentatively at present, a
differential sense of involvement with mathematics at university, a view which is
supported by the tendency for the men in the sample to consider mathematics
research as a career significantly more often than the women (Section 4.6). Again,
one can interpret these observations in terms of Chodorow's theories concerning the
greater importance for men of secure domains.

However, it was difficult to establish how representative the observed differences

were due to the size of the samples and the variation of the results between the two
surveys of mathematics students (Section 4.1.2). There was also no notable
difference in Survey 1 between the sexes in admitting to experiencing particular
difficulties with the course, implying that the more negative attitudes expressed by
the women did not appear to affect theft confidence to do the course.

One must also bear in mind the possible effects of teaching styles on the students'
attitudes, since there was some rather weak suggestion from Survey 3 that the
women attributed more importance to student/staff interactions (Section 4.3.1).

5.4. Achievement and participation in university mathematics

My surveys showed no evidence of consistent significant differences in

mathematical performance at university between the sexes (Section 4.4.1). The
analysis of degree results in Scotland and England over a three-year period showed
only very small gender differences overall, with some rather wild and at present
unexplained variations in performance from the men Scottish graduates (Section
4.5). The differences found in the national data were consistent with the APU
findings of boys' overrepresentation in both the top and bottom achievement bands
(Joffe & Foxman 1986).

My hunch that the relatively large proportion of women mathematics students found
at Edinburgh University during the preliminary survey (Section 1.2. 1) was typical of
Scottish universities was confirmed in the analysis of the national data (Section
 118

3.2.2). I have argued in Section 2.2.3 that such differences in women's participation
in university mathematics can be seen as being influenced by, if not the direct
outcome of, the relative openness of the Scottish education system ensuring that
more girls are mathematically qualified at the level required for university entry, a
view supported by Small (1987) in the context of physics education.

5.5. Synthesis and outlook

The general picture which emerges from this study is one which shows little overall
gender difference in mathematical performance at university, and somewhat more
obvious differences in participation which appear to be influenced by the extent of
specialisation present in the secondary education system. There was only rather
weak evidence of attitudinal differences between the women and men mathematics
students in my surveys: while there were a few significant differences observed, as
discussed in the previous sections in this chapter, they must be evaluated in the
context of the rather large number of tabulations involved in the study. In addition,
there were indications that attitude varied between achievement groups (Section
4.4.2), which would imply a gender-achievement interaction not accounted for in
this study.

What gender differences I did observe did appear to be consistent with the
interpretive framework developed by Chodomw and elaborated by Gilligan, as
discussed in Sections 5.1-5.3. However, there is a problem with using such a
framework with surveys of this nature insofar as I am ascribing complex and often
subconscious motivations to my subjects just because their apparent behaviour, as
reflected in their responses to my questions, seems to fit the patterns described by
Chodorow and Gilligan as the consequence of certain types of object-relationships.
Perhaps investigations based on case studies are the answer to this dislocation, a
solution which was not available to me at the time of the study. Despite my
impression that gender differences were rather small in my study, I do feel that
research in the field is of continuing importance. As long as gender differences are
felt (or used) to restrict women's opportunities in society, it is crucial to examine the
source of such differences and, most importantly, the belief in them. I have referred
to some beliefs regarding women's innate incapacity for intellectual activities in
Sections 1.1.4 and 1.1.5, beliefs which to us now appear laughable. Perhaps they
were not so at the time, and hopefully, in view of the apparent changes towards
 /

119

gender equality mentioned in Section 1.1.4, many of our currently popular beliefs
regarding the inclinations and limitations of one sex or another will prove equally
laughable to our descendants.

One particular concept I would wish to see examined in future studies is that of
'gender saliency', a notion discussed by Chodorow 11989) which implies that 'it is
likely that sex role will only mediate intellectual functioning of those persons who
evaluate it as central to their self definition' (Nash 1979, p.291). Gender saliency is
an important concept since, as Chodorow puts it
• Difference is psychologically more salient for men in a way that it is not for
women, because of gender differences in early formative developmental
processes and the particular unconscious conflicts and defenses these produce.
This salience, in mm, has been transmuted into a conscious cultural
preoccupation with gender difference. It has also become intertwined with and
has helped to produce more general cultural notions, particularly that
individualism, separateness, and distance from others are desirable and requisite
to autonomy and human fulfillment. Throughout these processes, it is women,
as mothers, who become the objects apart from which separateness, difference,
and autonomy are defined. (Chodorow 1989, p.112)

In other words, belief in the existence of gender differences, against which identity
can be defined, is more important for men than for women. She goes on to say
It is crucial for us [ ... ] to recognize that the ideologies of difference which
define us as women and as men, as well as inequality itself, are produced,
socially, psychologically, and culturally, by people living in and creating their
social, psychological, and cultural worlds. [ ... ] To speak of difference as a final,
irreducible concept and to focus on gender differences as central is to reify
them and to deny the reality of those processes which create the meaning and
significance of gender. To see men and women as qualitatively different kinds
of people, rather than seeing gender as processual, reflexive, and constructed, is
to reify and deny relations of gender, to see gender differences as permanent
rather than as created and situated.

We certainly need to understand how difference comes to be important, how it

is produced as salient, and how it reproduces sexual inequality. But we should
not appropriate differentiation and separation, or difference, for ourselves and
take it as given. (Chodorow 1989. pp. 112-113)

This is an approach which I believe is a more positive and constructive way of

dealing with the multi-dimensional phenomenon that is gender and its ramifications.
 120

References

Acker, S. (1984) Women in higher education: what is the problem? in Acker &
Piper.
Acker, S. & Piper, D.W. (eds) (1984) 15 higher education fair to women? Guilford:
SRHE & NFER-Nelson.
Assessment of Achievement Programme (1986) Mathematics 1983. Edinburgh:
Scottish Education Department.
Assessment of Achievement Programme (1989) Mathematics: second survey 1988.
Edinburgh: HMSO.
Assessment of Performance Unit (1982) Secondary survey report, no 3. London:
HMSO.
Assessment of Performance Unit (1985) A review of monitoring in mathematics:
1978 to 1982. London: HMSO.
Association of Commonwealth Universities (1988) University entrance 1988: the
official guide. London: ACU.
Astin, H.S., & Bayer, A.E. (1975) Sex discrimination in academe, in Mednick et al.
Bell, A.W., Costello, J. & KUchemann, D. (1983) A review of research in
mathematical education, part A: research on learning and teaching. Windsor.
NFER-Nelson.
Ben-Chaim, D., Cappan, G. & Houang, R.T. (1984) The effect of instruction on
spatial visualisation skills of middle school boys and girls. American educational
research journal, 25(1), pp.5 1-71.
Bentley, D. & Watts, M. (1987) Courting the positive virtues: a case for feminist
science, in Kelly.
Biller, H.B. (1972) Sex-role learning: some comments and complexities from a
multidimensional perspective. Washington DC: Symposium on sex role learning in
childhood and adolescence.
Brown, S.I. (1986) The logic of problem generation: from morality and solving to
deposing and rebellion, in Burton.
Burnhill, P.M., & McPherson, A.J. (1983) The Scottish university and undergraduate
expectations 1971-1981. Universities quarterly, 37(3), PR257-270•
Burnhill, P. & McPherson, A. (1984) Career and gender: the expectations of able
Scottish school leavers in 1971 and 1981, in Acker & Piper.
Burstyn, J. (1984) Educators' response to the scientific and medical studies of
women in England 1860-1900, in Acker & Piper.
Burton, L. (ed) (1986) Girls into maths can go. Eastbourne: Holt; Rinehart and
 121

Winston Ltd.
Burton, L. & Townsend, R. (1986) Girl-friendly mathematics, in Burton.

Chetwynd, A.(1992) Sex ratios in UK university mathematics departments. London -

Mathematical Society newsletter, January, 190.
Chodorow, N. (1978) The reproduction of mothering: psychoanalysis and sociology
of gender. Berkeley: University of California Press.
Chodorow, N.J. (1989) Feminism and psychoanalytic theory. Cambridge: Polity
Press and Basil Blackwell.
Clarke, S. (1988) Another look at the degree results of men and women. Studies in
higher education, 13(3), pp. 315-331 .
Cohen, G. & Fraser, E.J.P. (1992) Female participation in mathematical degrees at
English and Scottish universities. Journal of the Royal Statistical Society A, 155(2),
pp.241-258.

Crossman, M. (1987) Teachers' interactions with girls and boys in science lessons,
in Kelly.

Dale, R. R. (1974) Mixed or single-sex schools? vol 3. London: Routledge and Kegan
Paul.
Davis, P.J. & Hersh, R. (1981) The mathematical experience. Boston: Birkhauser.

Davis, P.J. & Hersh, R. (1986) Descartes' dream. London: Penguin.

Durdell, A., Siann, G. & Glissov, P. (1990) Gender differences and computing in
course choice at entry into higher education. British educational research journal,
16(2), pp. 149-162.

Eales, A. (1986) Girls and mathematics at Oadby Beauchamps College, in Burton.

Easlea, B. (198 1) Science and sexual oppression. London: Weidenfeld and Nicolson.

Eccles, J. (1985) Model of students' mathematics enrollment decisions, in Fennema.

Fennema, E. (1985) (ed) Explaining sex-related differences in mathematics:

theoretical models. Educational studies in mathematics, 16, pp.303-320.
Fraser, E.J.P. & Cormack, S. (1987) Women and mathematics: a different picture.
Teaching mathematics and its applications, 6(4), pp. 157-159.
Frieze, I.H. (1975) Women's expectations for and causal attributions of success and
failure, in Mednick et al.

Gilligan, C. (1979) Woman's place in man's life cycle. Harvard educational review,
49, pp.431-446.

Gilligan, C. (1982) In a different voice: psychological theory and and the

development of women. Cambridge, Mass.: Harvard University Press.
 122

Harding, J. & Sutoris, M. (1987) An object-relations account of the differential

involvement of boys and girls in science and technology, in Kelly.

Head, J. (1987) A model to link personality characteristics to a preference for

science, in Kelly.
Hoffman, L.W. (1975) Early childhood experiences and women's achievement
motives, in Mednick et al.

H'usen, T. (ed) (1967) International study of achievement in mathematics, vol 2. New

York: Wiley.

Hyde, J.S. (1981) How large are cognitive gender differences? American
psychologist, 36, pp. 892-901 .
Isaacson, Z. (1986) Freedom and girls' education: a philosophical discussion with
particular reference to mathematics, in Burton.
Joffe, L. & Foxman, D. (1986) Attitudes and sex differences-some APU findings, in
Burton.
Kaplan, A.G. & Bean, J.P. (eds) (1976) Beyond the sex-role stereotypes; readings
towards a psychology of androgyny. Boston: Little, Brown.
Kelly, A. (1976) Swings and roundabouts; trends in SCE science presentations since
1962. Scottish educational studies, 8(1), pp. 4-18 .
Kelly, A. (1978a) Sex differences in science enrolements: a provisional analysis of
Scottish patterns between S3 and tertiary education. CES collaborative research
newsletter, 3, pp. 61-69 .
Kelly, A. (1978b) Sex differences in science enrolements: reasons and remedies.
CES collaborative research newsletter, 4, pp.61-68 .
Kelly, A. (ed) (1987) Science for girls? Milton-Keynes: Open University Press.

Kelly, A. (1987a) Why girls don't do science, in Kelly.

Kelly, A. (1987b) The construction of masculine science, in Kelly.

Klem.ke, E.D., Hollinger, R. & Kline, A.D. (eds) (1980) Introductory readings in the
philosophy of science. Buffalo: Prometheus Books.
Kline, M. (1980) Mathematics, the loss of certainty. New York: Oxford University
Press.
Kolodny, A. (1981) Dancing through the minefield: some observations on the
theory, practice and politics of a feminist literary criticism, in Spender.

McPherson, A.F. (1972) The generally educated Scot: an old ideal in a changing
university structure, in Bernstein, 0., McPherson, A.F. & Swift, D.F. Eighteen
plus the final selection, E282, Unit 15. Bletchley:the Open University.

Maccoby, E.E. & JackliiI, C.N. (1974) The psychology of sex differences. Stanford:
 123

Stanford University Press.

Maines, D.R. (1985) Preliminary notes on a theory of informal bathers for women in
mathematics, in Fennema.
Mead, M. (1962) Male and female. London: Penguin.

Mednick, M.T., Tangri, S.S. & Hoffman, L.W. (1975) Women and achievement.
Washington: Hemisphere.

Nash, S.C. (1979) Sex role as a mediator of intellectual functioning, in Wittig,

M.A. & Petersen, A.C. (eds) Sex related differences in cognitive functioning:
developmental issues. New York: Academic Press.
National Consortium for Examination Results (1987) Examination results 1984.
Trowbridge: Chief Education Officer for Wiltshire County Council.
New Scientist (1988) Girls 'less willing' to guess answers. New Scientist, 26 May,
1614, p. 34.

Northam, J. (1986) Girls and boys in primary maths books, in Burton.

Redpath, B. & Harvey, B./Office of Population Census and Surveys for the
Department of Education and Science (1987) Young people's intentions to enter
higher education. London: HMSO.
Rothschild, J. (ed) (1983) Machina ex dea: feminist perspective on technology. New
York: Pergarnon Press.

The Royal Society and the Institute of Mathematics and its Applications (1986) Girls
and mathematics. London: the Royal Society.
Safilios-Rothschild, C. (1986) Sex differences in early socialisation and upbringing
and their consequences for educational choices and outcomes, in Organisation for
Economic Co-operation and Development Girls and women in education. Paris:
OECD.

Samuel, J. (1983) Mathematics and science introduction, in Whyld.

Scott-Hodgetts, R..(1986) Girls and mathematics: the negative implications of

success, in Burton.

Scottish Universities Council on Entrance (1985) Scottish universities entrance

guide. St Andrews: SUCE.
Segal, L. (1990) Slow motion: changing masculinities, changing men. London:
Virago.

Seward, J.P. & Seward, G.H. (1980) Sex differences: mental and temperamental.
Lexington, Mass.: D.C. Heath & Co.

Sherman, J.A. (1975) A summary of psychological sex differences, in Mednick et al.

Siann, G., Durndell, A., Macleod, H. & Glissov, P. (1988) Stereotyping in relation to
 124

the gender gap in participation in computing. Educational research, 30, pp.98-103 .

Simon, R.J., Clark, S.M. & Galway, K. (1975) The women PhD: a recent profile, in
Mednick et al.

Skemp, R.R. (1986) The psychology of learning mathematics. London: Penguin.

Smail, B. (1987) Encouraging girls to give physics a second chance, in Kelly.
Spear, M. (1987) The biasing influence of pupil sex in a science marking exercise, in
Kelly.
Spender, D. (ed) (198 1) Men's studies modified. Oxford: Pergamon Press.

Spender, D. (1984) Sexism in teacher education, in Acker & Piper.

Spender, D. (1986) Some thoughts on the power of mathematics, in Burton.

Stein, A.H. & Bailey, M.M. (1975) The socialization of achievement in females, in
Mednick et al.
Taylorson, D. (1984) The professional socialization, integration and identity of
women PhD candidates, in Acker & Piper.

Trickett, L. & Suilde, F. (1988) Low attainers can do mathematics, in Pimm, D. (ed)
Mathematics, teachers and children. London: Hodder and Stoughton/Open
University.

Universities' Statistical Record (1987) University statistics 1986-87 ,vol 2: students

and staff. Cheltenham: USR.
Walden, R. & Walkerdine, V. (1986) Characteristics, views and relationships in the
classroom, in Burton.

Walkerdine, V. & the Girls and Mathematics Unit (1989) Counting girls out.
London: Virago.

Ward, M. (1972) Mathematics and the ten-year-old. London: Evans/Methuen

Educational.

Weinmich-Haste, H. (1984) The values and aspirations of English women

undergraduates, in Acker & Piper.

Willms, J.D. & Kerr, P. (1987) Changes in sex differences in Scottish examination
results since 1975. Journal of early adolescence, 7, pp.85-105.
Whyld, J. (ed) (1983) Sexism in the secondary curriculum. London: Harper & Row.

Whyld, J. (1983) More than one way forward, in Whyld.

 125

I. Survey 1

1.1. Methodology

The target population consisted of all first and second year students who had a
Director of Studies in the Mathematics or Statistics Departments and the third and
fourth year Mathematics Honours students. In total there were 211 students, 81
women and 130 men.
The target population was stratified by year and sex and a sample of 100 was chosen
by systematic selection with random start. The sample was equally split by sex and
comprised 30 students from each of the first two years and 20 from each of the final
two.
The survey was piloted on a small sample of 8 third year students during March
1987 and the main survey took place in the third term. The questionnaires were
completed by the interviewer in the course of one-to-one interviews, with the
students. Due to constraints of time and resources, I did all the interviewing. There
was therefore no opportunity to control for interviewer effect on the responses.
Letters were sent out via lecturers and tutors asking students to participate in a
survey on attitudes towards mathematics and to sign up for an interview at a
convenient time. The letters for the final year students were put in their work-room
since they were all attending different lectures and tutorials. There was a choice of
two interview locations, one in the main George Square campus and the other in the
science campus at the King's Buildings. In the cases of non-response, letters were
sent to the term-time address two weeks later. If there was no response after this, the
telephone number was obtained and the subject contacted by phone, either to arrange
an interview time or to complete the questionnaire over the phone. The Halls of
Residence proved difficult to reach by telephone and a few interviews were carried
out in-situ, but this was very time-consuming.
In the end, 54 students responded to the first letter and 17 to the second one. A
further three came in for an interview after being contacted by phone, five were
interviewed over the phone and two at home. However, some of the second year
students said they had not received the first letter. Since these had been given to
various tutors to distribute, it was difficult to ascertain how many had actually been
delivered before a second letter was sent. The students' responses did not appear to
be greatly affected by these variations in the contact and interview procedure, but it
is not possible to be certain of this as the numbers were quite small.
The final year students were the main source of non-response, possibly due to the
approaching final examinations. A detailed breakdown of the selected and achieved
samples is shown below.
 126

population number sampling achieved total

selected fraction sample response
selected rate
year

1 women 19 15 0.78 15 1.00

men 40 15 0.37 12 0.80

2 women 26 15 0.55 14 0.93

men 29 15 0.51 12 0.80
3 women 22 10 0.45 10 1.00
men 32 10 0.31 7 0.70
4 women 14 10 0.71 5 0.50
men 29 10 0.34 6 0.60

all women 81 50 0.61 44 0.88

years men 130 50 0.41 37 0.74

all 211 100 0.47 81 0.81

 127

1.11. Results

All results are given as frequency counts. Unless otherwise stated, N women-44 and
N men=37.
Qi Sex of respondents
The breakdown is shown in the table describing the sampling frame.
Q2 Age of students
It was considered more useful to look at students in different years than students of
different ages.
Q3 Age at entry
These data were not considered worth analysingsince very few of the students had
entered university straight after Highers and so the majority had been 18 on entry.
Q4 Type of degree
Most of the degrees classified as 'other' were joint Mathematics/Computer Science
degrees.
women men
BSc Mathematics 25 26
MA Mathematics 4 2
BSc Mathematics/Statistics 9 6
Other 6 3

Q5 Type of school attended

Since the numbers of students from independent and single-sex schools were fairly
small, it was not practical to compare the responses from students having attended
different types of school. The responses for this question are therefore only shown
for interest. None of the students had attended independent'comprehensive schools.
Though an unlikely category, these had been included in the list for completeness.
women- men
Single-sex state supported selective 3 2
Mixed state supported seleètive 7 1
Single-sex independent selective 5 3
Mixed independent selective 1 2
Single-sex state supported comprehensive 1 0
Mixed state supported comprehensive 26 28
Other 1 1
Q6 Location of school attended - -
This question was superfluous since it was decided to classify the students by the
type of public examinations they took (ie. Highers or A levels) rather than consider
whether the school itself was Scottish or English. Therefore students who attended
Scottish schools but did A levels instead of Highers were in a sense seen as having
had an English education. It was important to make this distinction because I had
initially wished to compare the effects of the Scottish and English education systems
on the students' responses. However, most of the analyses using this variable proved
 128

somewhat inconclusive.
Q7 Type of public examination taken
One women had taken the International Baccalaureat.
women men
Students with Highers 25 25
Students with A levels 18 12
Q8 A level/Higher results
This information was not used in the analyses since it was felt that performance in
subjects other than mathematics did not have a direct bearing on the field of study.
Mathematical achievement in the public examinations showed very little variation,
probably due to the nature of the sample, and was therefore not controlled for.
Q9 Certificate of Sixth Year Studies Mathematics options offered and taken
Algebra Calculus Statistics Computing Mechanics
offered
women 19 24 4 19 11
men 15 21 7 17 7
taken
women 15 •23 0 8 7
men 10 20 2 13 5

Q10 University courses taken

The responses to this question were not analysed since they did not appear to be of
particular interest to this study. It had been thought that women might show a greater
tendency to have done non-science subjects, but the numbers doing so were too
small to be conclusive. There was also some difficulty in comparing students from
different years who had done different numbers of courses.
Qil Applied Mathematics 1 results
Analysing this question proved somewhat complicated since students could pass one
half of the course and fail the other half. In addition, since the course was
compulsory for students taking the BSc, differences in levels of interest could affect
performance patterns and make straightforward comparison difficult. Controlling for
different levels of expressed interest (see Q12) did not seem worthwhile due to the
small numbers involved.
Q12 Attitude towards Applied Mathematics 1
Applied Mathematics 1 consists of two half courses:
Ah Non-Physical Applied Mathematics
Bh Physical Applied Mathematics
N women=40, N men=33
Ten students were exempted from the course and one man did not take Applied
Mathematics 1 Bh.
 129

difficulty (very easy) 1 2 3 4 5 (very difficult)

Ah
women 3 10 23 3 1
men 1 18 10 4 0
Bh
women 0 11 7 13 9
men 3 6 7 11 5
interest (very interesting) 1 2 3 4 5 (very boring)
Ah
women 0 11 19 9 1
men 4 14 8 7 0
Oh
women 0 12 14 8 6
men 2 13 5 6 6
usefulness (very useful) 1 2 3 4 5 (a waste of time)
Ah
women 3 13 13 10 1
men 7 14 9 3 0
Bh
women 4 14 11 6 5
men 1 10 13 5 3
Q13 Attitude towards current mathematics course
One woman did not rate her course for usefulness.
difficulty (very easy) 1 2 3 4 5 (very difficult)
women 0 5 16 20 3
men 0 10 14 13 0
interest (very interesting) 1 2 3 4 5 (very boring)
women 2 8 19 14 1
men 6 12 11 7 1.
usefulness (very useful) 1 2 3 4 5 (a waste of time)
women. 1 10 17 9 6
men 5 1013 9 0

Q14 Attitude towards the components of the mathematics course

All students were initially asked this question. However, the third and final year
students were taking a variety of courses (shown in the questionnaire) which proved
impractical to classify in a comparable way to the first and second year courses.
Students doing the Mathematics/ Statistics degree are exempt from some third year
courses and were also difficult to compare. For these reasons, only the results for the
first and second years are given.
N women=29, N men=24
 130

difficulty (very easy) 1 2 3 4 5 (very difficult)

algebra
women 1 12 10 6 0
men 7 11 3 3 0
analysis
women 0 1 4 11 13
men 1 3 6 9 5
calculus
women 5 5 12 5 2
men 3 9 10 2 0
interest (very interesting) 1 2 3 4 5 (very boring)
algebra
women 2 7 14 5 1
men 3 5 7 5 4
analysis
women 0 2 4 12 11
men 2 3 5 9 5
calculus
women 1 12 12 4 0
men 6 10 6 2 0
usefulness (very useful) 1 2 3 4 5 (a waste of time)
algebra
women 0 5 16 8 0
men 1 3 10 8 2
analysis
women 0 4 8 10 7
men 2 3 2 13 4
calculus
women 1 11 12 4 1
men 6 13 4 1 0
Q15 Achievement in the mathematics course
Since it was not possible to verify that the students had accurately reported their
examination results, the results shown below were obtained from Departmental
records and are for the entire selected sample. It was felt that this would give a more
representative picture of the achievement patterns than only considering the
performance of the achieved sample.
2A class: 7 women and 8 men obtained Merit passes for 1A (Nw=l5,Nm=15)
3A class: 7 women and 4 men 'I (Nw=10,Nm=10)
4 women and 2 men 2A
4 class: 4 women and 8 men 1A (Nw=10,Nm=10)
6 women and 6 men 2A
3A results: 1 woman and 4 men obtained A (1st)
1 woman and 5 men B (2.1)
6 women and 1 man C (2.2)
2 women obtained 0 (3rd)
Q16 Attitude towards mathematics at secondary school
Two women did not rate school mathematics for usefulness
 131

difficulty (very easy) 1 2 3 4 5 (very difficult)

women 22 18 3 1 0
men 18 13 4 2 0
interest (very interesting) 1 2 3 4 5 (very boring)
women 18 18 6 2 0
men 11 12 9 3 2
usefulness (very useful) 1 2 3 4 5 (a waste of time)
women 17 15 8 2 0
men 10 16 7 4 0
Q17 Perception of encouragement to do mathematics at school
One woman's response was missing.
encouraged discouraged neither
women 28 3 12
men 15 3 19
Q18 Other people's influence on students' choice of mathematics at school and
university
This was presented as an open question.
women men
at school
no-one 27 24
teachers 13 10
parents 5 7
career advisers 3 0
other 1 1
at university
no-one 33 29
teachers 9 4
parents 2 4
career advisers 3 2
other 2 1
Q19 Occupation of parents
Due to the difficulty of accurately assessing the social class of the parents from the
job descriptions given by the students, the only classification considered was manual
and non-manual employment. 66 students had fathers in non-manual employment,
with 14 students having fathers in manual employment (one father was dead and no
information was obtained on him). Since the number of students from working-class
families was so small, it was not considered worthwhile controlling for class
differences in this study. However, there were more women from working-class
backgrounds: 9 women and 5 men had fathers in manual occupations.
Q20 Reasons for having chosen to do a mathematics degree at university
This question was presented as an open question and the responses classified into the
categories below.
 132

women men
Found mathematics easy/was good at it 27 22
Was interested in the subject 28 19
Thought mathematics might be useful
for a career 10 7
Other reasons 5 5
Precedents amongst family or friends 2 0
Q21 The proportions of girls in secondary school mathematics classes
The results for this question showed the same patterns seen in the national figures
(Tables 2.1, 2.2a and 2.2b, Chapter 2), and it was therefore considered unnecessary
to present them here.
Q22 Reasons for choosing Edinburgh University
This question was also presented as an open question.
women men
Locality of university 35 30
Reputation of university or course 26 23
Structure of offered course 12 3
Influence of other people 7 4
Other 8 10
Q23 Particular difficulty experienced by the students
women men
Experiencing no particular difficulty 31 28
Experiencing some difficulty 12 9
Did not know 1 0
Q24 Parts of the course students liked best
The responses to this question were difficult to analyse since they were not really
comparable. Some students mentioned course components, while some mentioned
more detailed sub-topics. Since the lecturer also seemed to influence attitude towards
the course, students from different years varied in their opinions of the same course
components. Since it was impossible to control for these factors, detailed analysis
was not considered worthwhile or appropriate.
Q25 Students' opinions of the style of presentation of the course material
Most students seemed to think that it would be difficult to teach the course any
differently from the way it was taught. Some students said that more hand-outs of
lecture notes would be helpful. But this was not really what the question was
designed to examine, which was whether students thought that other forms of
learning, such as projects or group-work, might improve their learning experience.
The responses to this question seemed to indicate that the students had not thought
very much about the way the course material was presented and it was therefore
decided that further analysis would not be worthwhile.
Q26 Number of questions attempted on worksheets
Questions on the worksheets are categorised as A. B or C depending on the degree of
difficulty. Questions marked 'C' are peripheral to the course material and are
 133

provided for interest. The responses to Q26 were difficult to study because they were
often not precise. In retrospect, this question should not have been presented as an
open question and the responses were not analysed.
Q27 Ideas for the future
At first, the first and second years were asked if they were going to continue to study
for a mathematics degree and not asked about their ideas for the future. However, it
later seemed worthwhile to ask the latter question to all the students (excepting the
10 first year students interviewed before the change of policy). Only 3 students said
they were going to change their degree.
N women=29, N men=23
women men
Finance 17 12
Did not mention a career 10 9
Research 0 9
Other 6 3
Teaching 3 3
Management 4 0
Computing 1 2
 134

Lifi. Questionnaire I
Course year running number

10. Wli•. LIz%v,r ity .ot' racs h.,c you taker?

I physics (10) U
I. sex I male — 2 au theuria ii os I phys 1>4
2 female (I) 3 cusinpuuter science
4 statistics 4
2. age 5 applied maths -
(2) 111112 6 other science subjects -
3. year of entry 7 other arts subjects_ -----
11. 11 applied maths was taken what results did you get? --
4. degree I
2
BSc Maths
IdA Maths
112 1 merit 2 pass 1st attenupt 3 subsequent pass 4 fail (ll)
3 BSo Waths/Stats
4 other - 12. How did you find the applied maths course? Grade it on a
1 to 5 scale with 0 no opinion
Ab lit
5. What type of school did you attend before university? (5) LII
1 single-sex state supported selective I very easy 5 very difficult (12) 1,1111 A [1
2 mixed-sex state supported selective
3 single-sex independent selective 1 very interesting 5 very boring [1 6 0
4 mixed-sex independent selective
5 single-sex state supported comprehensive 1 very useful 5 a waste of time LI C [4
8 mixed-sax state supported oomprehnsjve
7 single-sex independent comprehensive 13. How do you find your current maths course on the whole?
-4
6 mixed-sex independent comprehensive Grade on p 1 to 5 scale with U)
9 other___________________________________ U'
6. Was it 1 Scottish I very easy 5 very difficult (13) [Uk
(6) LI
2 English
3 H. Ireland 1 very interesting S very boring [J B
other 1 very useful 5 a waste of time i_i C
7. Did you do Highers or A-levels? I Highers (7) LII
2 A-levels
3 other_______ 14. How do you find specific parts of the course?
8. What were your five best Highers or three best
A-levels and what results did you get? (6) LI ss course j$efulness
1 __________ A B C D E
2
3
A B C D E
ABCDE
LIIIJAS
L11sco
algebra
T
4 A B C D E
unnslyusso I
5 __________ A B C D E deiculus
9. (a) What maths SIB options were offered at your school?

(b) Which ones did you take (if applicable)?

 IS. How did you find ealhs at school. Grade on a scale from
1 to 5 for (a) difficulty
(h) interest A
I very easy 5 very difficult 161 B
difficulty Cc) usefulness -

interdot. I very interesting 5 very boring L__ C

,,sefuliieos I very useful 5 a waste of time
V/. Did you feel that you were particularly encouraged
-
to do maths while still at school?
(ii) jIJj
14. How do you find specific parts of the course ? I encouraged
2 discouraged
0 no opinion
18. Do you feel that your decision to do maths at
(a) school
(14 university (18) ._r. A
was influenced by a particular person or persons ?
0 no One
I male teacher
2 female teacher
3 male careers advisor
4 female career advisor
for 4th veers S father
6 mother -
14. (a) which courses are you taking? (14) 7 other_.._____ -____
Lebesque integration and Fourier analysis 4 - ._Ja —
18 functional analysis - -
complex analysis - -
3A Galois theory What are the oocupations of your parents?
38 algebraic coding theory - -
number theory -- mother_,_
48 mathematical logic father_____ ._.. -

basic topology
68 topology and geometry of surfaces
I-
GA numerical analysis --
68 numerical methods for EDE's - Why did you choose to do maths at university in preference
7?. mathematical programming - to other subjects?
7B applied graph theory -
BA calculus of variations and PDE's -
88 nonlinear methods 21. What was the proportion of .ales to females in your maths class
10?. ,oathematics;education and history - at school? (a) 0 grade/0-,
higher/A-I]
(b) Grade each on a 1 to S scale for S'iS

A difficulty 22. Why did you choose Edinburgh University?

B interest
C usefulnass
23. Do you feel that you have anyparticular difficulty in this
15. What results have you had on this course? course compared to the rest of the class?
1 yes
2 no (23) IIIJ
0 don't know
If so, what difficulties are you experiencing?
 Which perts of the course do you like best and why?

What do you think at the style at presentation of the material?

About bow many questions do you attempt on each tutorial sheet?

for let and 2nd years

will you continue to do a maths degree?
U)

for 3rd and 4th years

27. Do you have any ideas for the future?

 138

H. Survey 2

11.1. Methodology

The target population for this survey consisted of students in second year and above
in the Faculties of Science, Social Science, Art and Medicine who had obtained
grade A fof Higher Mathematics or grades A or B for A level Mathematics and were
not registered for a mathematics degree. The list of students was provided by the
Registry Office and classified the population by Faculty, sex and degree. Within
degrees, the names were listed alphabetically. Since the students were selected on
the basis of theft year of matriculation, there were a few cases where the student was
actually taking a first year course, particularly in the case of medical students having
done a pre-medicine course.
Due to the difficulty of contacting non-mathematics students and assuring their
cooperation, as well as the fact that the questionnaire was fairly short and
straightforward, this survey was not piloted. The format was similar to that of
Survey 1 (questionnaire filled in by interviewer during a one-to-one interview).
Again the sample was chosen by systematic selection with random start and the
detailed composition is shown in the table below. Address labels were then provided
by Registry, but not telephone numbers.
In the beginning of the Autumn term 1987 letters were sent to the students'
term-time address asking them to participate in the survey. As a result of responses
to these letters, some subjects had to be replaced in the sample (medical students on
electives abroad, foreign language students on theft year abroad and students with no
known address). The initial response was poor and in December 1987 another letter
was sent and efforts made to increase the response rate (forms and self-addressed
envelopes enclosed in order to set up interview times, telephone numbers and
convenient times requested for telephone interviews). When the second wave of
responses petered out, a sample of the remaining students was selected to be
interviewed at home in the beginning of the second term.
Only 22 students responded to the first letter. The second letter attracted 25 more.
Six students responded by giving phone numbers and were interviewed by phone
and 30 were interviewed at home. The breakdown of the response-rate by Faculty
and sex is shown below.
 139

population number sampling achieved total

selected fraction sample response
selected rate

Faculty

Arts women 78 15 0.19 8 0.50

men 79 15 0.19 9 0.60

Medicine women 229 14 0.07 5 0.36

men 261 15 0.06 8 0.50

Science women 239 30 0.13 20 0.67

men 642 30 0.05 15 0.50

Social women 135 15 0.11 13 0.87

Science men 184 15 0.08 5 0.33

All women 681 74 0.11 46 0.62

Faculties men 1166 75 0.06 37 0.49

All 1847 149 0.08 83 0.56

 140

11.11. Results

Again, all the results are presented as frequency counts. Unless otherwise stated,
N women=46 and N men=37.
Qi Sex of respondents
The breakdown is shown in the table describing the sampling frame.
Q2 Age at entry to course
For the same reasons as in Survey 1, it was decided not to analyse the responses to
this question.
Q3 Year of entry
As above.
Q4 Faculty and degree
This information is shown in the table on the previous page.
Q5 Type of school attended
women men
Single-sex state supported selective 3 3
Mixed state supported selective 3 1
Single-sex independent selective 4 4
Mixed independent selective 6 8
Single-sex state supported comprehensive 0 0
Mixed state supported comprehensive 27 20
Single-sex independent comprehensive 2 1
Mixed independent comprehensive 1 0
Q6 Locality of school
For the sane reasons as those mentioned for Survey 1, the responses to this question
were not analysed.
Q7 Type of public examination taken
Students who had done both Highers and A levels were regarded as having done
Highers for the purposes of this study.
women men
Students with Highers 28 17
Students with A levels 18 20
Q8 Public examination results
For the same reasons as in Survey 1, this information was not used in the final
analysis.
Q9 Scottish students having taken CSYS Mathematics
One women said her school had not offered SYS Mathematics and one man had
done A levels instead of SYS.
N women=27, N men=16
 141

women men
Students having taken SYS 15 9
Students not having taken SYS 12 7

Q10 University courses taken and attitude towards mathematics course (if
taken)
There was too much variation in the courses taken to permit an analysis of this
question. Only a few students were taking non-compulsory mathematics courses and
it therefore did not seem worthwhile considering this particular part of the question.
Q 1 Attitude towards secondary school mathematics
Science students: N women=20, N men=15
difficulty (very easy) 1 2 3 4 5 (very difficult)
all students
women 3 20 16 6 1
men 13 12 7 3 2
science students
women 0 12 5 3 0
men 7 7 0 0 1

interest (very interesting) 1 2 3 4 5 (very boring)

all students
women 8 16 14 7 1
men 7 13 11 5 1
science students
women 5 7 6 2 0
men 1 6 6 1 1

usefulness (very useful) 1 2 3 4 5 (a waste of time)

all students
women 6 17 13 9 1
men 7 15 9 4 2
science students
women 4 10 4 2 0
men 3 7 3 1 1

Q12 Perception of encouragement to study mathematics at school

The wording of this question was the same as for Q17 in Survey 1.
encouraged discouraged neither
women 29 1 16
men 20 0 17

Q13 Reasons for having taken Higher/A level Mathematics

This was an open question, and the responses were categorised as shown in the table.
'Finding mathematics useful or necessary' included responses describing
mathematics as important or basic, or necessary for university entry.
The category 'other' includes a fairly sizeable proportion of responses to the effect
that the student had been expected to take the subject or that it 'went well' with the
other subjects s/he was taking. Although the latter reason may be taken to mean that
the student saw mathematics as useful or necessary for other subjects, the response
was not categorised as such unless usefulness or necessity were specifically
mentioned.
 142

students with
women men Highers A levels
Finding mathematics useful
or necessary 21 18 24 15
Ability 15 20 15 20
Interest 17 11 11 17
Other 21 12 18 15
Q14 Reasons for not having chosen to do a mathematics degree
One woman had changed to medicine after having started a mathematics degree and
one man had changed from a joint Mathematics/Philosophy degree to Philosophy.
Their reasons for changing to non-mathematics degrees are included in the following
results.
students with
women men Highers A levels
Interest in other subjects 21 12 21 12
Finding mathematics lacking in
usefulness or relevance 10 13 12 11
Not finding mathematics interesting
or enjoyable 9 6 9 6
Finding mathematics difficult 9 6 9 6
Not seeing any career potential in
mathematics or wanting a career in
another field . 7 8 7 8
Other 5 6 5 6
QiS Proportions of girls and boys in school mathematics classes
Since the responses to this question were similar to those for Q21 in survey
1, it does not seem worthwhile to show them here.
Q16 Perceived proportion of women in the university mathematics course
I considered three independent variables for the analysis of this question: sex,
educational background (ie. whether the student had done Highers or A levels), and
the proportion of girls in the student's Higher/A level mathematics class. A class was
considered male-dominated if the proportion of women was less. than 40%, and
roughly balanced if there were 40%-60% women.
Perception of the university
mathematics course as:
male-dominated roughly balanced
sex
women 25 21
men 22 15
educational background
Highers 23 22
A levels 24 14
proportion of girls in mathematics class
all male 4 6
up to 40% 18 5
40%-60% 17 21
over 60% 0 3
all female 8 1
Q17 Reasons for choice of degree subject and attitude towards degree subject
The first part of this question was open and the categories of response are shown in
the table.
 143

The response of one man was missing for the first part of the question.
Reasons for degree choice
students with
women men Highers A levels
Ability 6 2 6 2
Enjoyment or interest 28 27 30 25
Career considerations 19 22 21 20
Influence of others 5 3 6 2
Not sure of reason, type of
Highers/A levels taken 10 2 6 6
Other 9 8 8 9

Attitude towards degree subject

difficulty (very easy) 1 2 3 4 5 (very difficult)
women 1 10 15 18 2
men 0 7 13 13 4
interest (very interesting) 1 2 3 4 5 (very boring)
women 8 26 7 5 0
men 11 11 11 4 0
usefulness (very useful) 1 2 3 4 S (a waste of time)
women 13 17 10 S 1
men 14 10 8 2 3
QiS Reasons for the choice of Edinburgh University
women men
Locality of university 34 25
Reputation of university or course 17 24
Structure of offered course 15 7
Influence of others 5 4
Other 7 10
Q19 Occupation of parents
As for Survey 1, it was decided to consider the father's occupation in assigning
social class. Since the number of students who could be considered as originating
from the working-class was small, it did not seem worthwhile controlling for this
factor.
One woman's response was missing. Mother one's father was dead and his
occupation was not ascertained.
Only 8 students (6 women and 2 men) came from a working-class background.
Q20 Finding a knowledge of mathematics useful for chosen course
useful not useful
women 34 12
men 29 8
Q21 Altered opinion of mathematics since school
 Om

changed for no change

better worse
sex
women 11 5 30
men 9 8 20
Faculty
Arts 5 1 11
Medicine 2 1 10
Science 10 11 14
Social Science 3 0 15
Q22 Usefulness of mathematics in the future
The type of mathematics the students thought might be useful varied greatly
according to degree course and envisaged career. It was therefore not considered
practical to analyse the responses to the second half of this question.
useful not useful did not know
women 40 5 1
men 35 1 1
Q23 Ideas for the future
Again the responses varied with degree subject and for reasons of practicality it was
decided to only distinguish students with career plans from those without.
In the sample, 37 women and 30 men had career plans.
 145

ILifi. Questionnaire 2
course year -—- running number

10. university courses t.ikc..

I. sex I male
2 female (I) LIII
2. age at entry to course (2)
(if maths course taken) What did you think of the naths course?

3. year of entry (3) LI

4. faculty and degree
8 Science
Arts 4 Medicine (4)
9 Social Science
9
5. type of school attended before university
I single-sex state supported selective
(5)
-
LI grade from I to 5
difficulty lvery easy S very difficult (10) (a)
2 mixed-sex state supported selective
interest I very interesting 5 very boring (b)
3 single-sex independent selective
usefulness I very useful S a uaste of tine (C)
4 mixed-sex independent selective
single-sex state supported comprehensive 11. maths attitude at school
6 mixed-sex state supported comprehensive difficulty (II) (a)
7 single-sex independent comprehensive interest (b)
8 mixed-sex independent comprehensive usefulness - (C)
9 other coments; how did you feel about it in general?

6. was it I Scottish (6) 11

English
3 N. Irish
4 other

7. type of exams I Higher. - (7) LI

12. level of maths encouragement at school I encouraged
2 discouraged
(I?) 1111
2 A-levels 3 no opinion
3-other coeseenta: whet kind of dis/encouragement?

8. five best Higher. or three best A-levels

1
() 1111
2 ABCDE
3 ASCOE 13. Why did you decide to do Higher/A-level Maths?
4 ABCDE
5 ABCDE

9. SYS Maths papers (a) offered

(b) taken

14. I-shy did you decide not to continue to study for aflaths degree at university?
(if appiscable) Will did you choose to do a maths course at university?
 IS. proportion of female, to malt, in school macit. class
0-grade/C-level
II islier IA-level
(C) ITS

What do you think the proportion of females to males is in the university

maths degree course? -

how did you choose your degree subject?

grade from I to S for

difficulty (17) - (a)
interest (b)
usefulness (c)

IS. Thy did you choose Edinburgh Uni ver s it y ?

occupations of parent,
mother
father

Do you find a knowledge of mathematics useful for your course? (20)

I yes
2 no
0 no opinion
comaentsl At what level?

Has your opinion of mathematics altered since school? How?

Do you think mathematics could be useful to you in the future? What type?

23. ideas for the future

III. Survey 3

111.1. Methodology

The aim of Survey 3 was to interview all the second year students who were
intending to take Mathematics or Mathematics/Statistics Honours degrees. It was
decided not to include students taking other joint degrees since previous experience
had indicated that such students were fairly likely to change their degree subject after
second year. The selected sample comprised 55 students, 24 women and 31 men.
The questionnaire was piloted with three final year students in the beginning of the
Spring term. Some modifications were made as a result, and letters sent out to the
selected students' term-time addresses asking them to come and be interviewed for a
survey on attitudes towards mathematics and mathematics education. A second letter
was sent at the beginning of the Summer term to those students who had not yet
responded and then a third one mentioning the possibility of a home-visit. In
addition to the letters, an announcement was made during one of the lectures and the
students contacted by phone in order to arrange interview times.
The format of the questionnaire varied somewhat from that of the two previous
surveys in that the students were asked to indicate their responses to some questions
(those asterisked in the questionnaire) on pre-printed cards. This was done in order
to assure that the students' responses were not unduly influenced by the order of the
components in the questions. It was therefore not practical to conduct the interviews
by phone. Prompted and unprompted comments were recorded separately so as to
permit an assessment of whether the students showed differences in their willingness
to express their opinions. However, there were very few comments for most
questions and therefore they were not analysed in much detail.
Four of the students who responded had to be eliminated from the sample: two were
doing joint degrees, one was going to change his degree subject, and one was an
exchange student The achieved sample then consisted of 41 students, 20 women and
21 men. Twenty-one students responded to the first letter, 14 came to be interviewed
after the phone-call. 4 after the second letter and 2 after the third. The only
home-visit was to a student who was subsequently not included in the sample. The
responses did not appear to be much affected by the number of letters required to
obtain the interview, but it is difficult to be certain of this due to the small numbers
involved.
 149

IH.II. Results

As for Surveys 1 and 2, all results are shown as frequency counts. For some of the,
tables, the wording of the precoded answers was changed slightly from that in the
questionnaire for presentation purposes. Unless otherwise stated, N women=20 and
N men=21.
Qi Sex of respondents
Shown above.
Q2 Type of degree
Two women were doing Mathematics MAs, one was planning on changing to the
BSc.
Q3 Location of school attended
As for the first two surveys, it was decided to distinguish between English and
Scottish students by considering whether they had taken A levels or Highers (Q4).
Q4 Type of public examination taken
Students who had taken Highers but had done A levels instead of CSYS were
considered as having done Highers for the purposes of this study.
women men
Students with Highers 13 16
Students with A levels 7 5
QS A level/Higher results
For the same reasons as in Survey 1, this information was not used in the final
analysis.
Q6 CSYS Mathematics options offered and taken
One man and one woman did not take SYS Mathematics and did not know what
options were offered at their school. Some students prepared SYS papers outside of
the offered options, and therefore there are some discrepancies in the table between
the numbers of students being offered certain papers and the numbers having taken
them.
Algebra Calculus Statistics Computing Mechanics
offered
women 8 12 7 8 3
men 12 15 4 12 5
taken
women 7 12 3 3 1
men 13 14 2 9 4
Q7 Other university courses taken
This question was not analysed because the numbers were too small for the results to
be useful.
 150

Q8 Attitude towards mathematics at secondary school

difficulty (very easy) 1 2 3 4 5 (very difficult)
women 3 11 5 1 0
men 6 7 8 0 0
interest (very interesting) 1 2 3 4 5 (very boring)
women 2 9 8 1 0
men 5 8 6 2 0
usefulness (very useful) 1 2 3 4 5 (a waste of time)
women 3 8 9 0 0
men 5 7 7 2 0

Q9 Students' feelings about mathematics at secondary school

Preliminary impressions of the responses to this question did not indicate that the
students were giving other than the basic reasons for which they chose to take a
mathematics degree at university. Probing responses such as liking mathematics did
not tend to clarify the 'deeper' reasons why the students felt the way they did about
mathematics. Quite a few students mentioned
liking the problem-solving aspect of mathematics and being able to do it, but it was
very difficult to obtain more detailed responses without very time- consuming
in-depth interviews. Perhaps a less structured survey might have proved less
inhibiting for questions of this nature.
Q10 Reasons for having chosen to do a mathematics degree at university
The students were asked to indicate on a pre-printed card how important each
consideration had been in influencing their choice of a mathematics degree.
very fairly unimportant
important important
Being good at maths at school: women 18 2 0
men 14 7 0
Finding maths easy: women 6 13 1
men 8 12 1
Finding maths interesting: women 9 8 3
men 11 9 1
Thinking a maths degree would
be useful for your career: women 7 11 2
men 9 8 4
Knowing people who had taken
a maths degree: women 0 5 15
men 0 1 20
Encouragement from teachers
or other people: women 3 8 9
men 3 9 9
Qil Whether the students had considered doing something other than a
mathematics degree
12 women and 17 men had considered other degree subjects. For Higher and A level
students, the numbers were 20 and 9 respectively. The other subjects varied
 151

somewhat and the numbers were too small to justify a detailed analysis, but there did
not appear to be any notable difference in the subjects considered by women and
men.
Q12 Why the students had chosen to do the BSc degree rather than the MA
Most students had not considered doing an MA and therefore there were no positive
reasons for the decision to take the BSc degree. The majority of students said that
they had not known about the MA and that mathematics was a science. Since there
were only two students studying for an MA and one planned to change to a BSc,
there did not seem much point in a detailed analysis of this question.
Q13 Preparation given by the school to do a mathematics degree
Most students thought that the preparation given by the schools was adequate, with
some Scottish students pointing out that having done SYS Mathematics was quite
important. However. 6 women and 2 men said their preparation had not been
adequate. The principal reason mentioned was that schools emphasised the
importance of acquiring techniques to pass examinations, while the university course
required a different, more theoretical, approach.
Q14 Proportion of girls in school mathematics classes
There did not appear to be any difference between the responses in this survey and
those in the previous ones.
Q15 Expectations concerning the proportion of women in the university
mathematics course This was the only question which touched directly on the
particular aspect of the perceived 'masculinity' of mathematics and the phrasing
presented some difficulty. The wording of the question was carefully chosen in
order to minimise bias and thus obtain a realistic impression of how the student
perceived mathematics. As a result, the analysis of the responses was not as
straightforward as it might have been and the presentation of the findings reflect this
to a certain extent. The reasons given for the perceived/expected proportion of
women being under 40% are shown Section 4.1.2. As before, a roughly balanced
class is defined as one with 40%-60% women.
Three female Higher students had not had any expectations concerning the
proportion of women in the mathematics course and one of these said she did not
know what the proportion actually was.
students with
women men Highers A levels
Seeing the class as roughly balanced 13 14 22
Seeing the class as male-dominated 6 7 6
of women was as expected 10 16 20
was not as expected 7 5 6
 152

Q16 Attitude towards Mathematics 1A

Two women were direct entrants to Mathematics 2A and are therefore not included
in this analysis.
difficulty (very easy) 1 2 3 4 5 (very difficult)
sex
women 0 1 11 5 1
men 2 6 7 6 0
lÀ Merit
Merit students 2 6 10 4 0
non-Merit students 0 1 8 7 1
interest (very interesting) 1 2 3 4 5 (very boring)
sex
women 1 3 8 5 1
men 0 6 8 6 1
lÀ Merit
Merit students 0 8 1 3 0
non-Merit students 1 1 5 8 2
usefulness (very useful) 1 2 3 4 5 (a waste of time)
sex
women 1 9 5 3 0
men 2 7 5 6 1
lÀ Merit
Merit students 2 12 5 3 0
non-Merit students 1 4 5 6 1
Q17 Attitude towards Mathematics 2A
difficulty (very easy) 1 2 3 4 5 (very difficult)
sex
women 0 0 5 14 1
men 0 1 7 10 3
lÀ Merit
Merit students 0 1 8 11 2
non-Merit students 0 0 4 11 2
interest (very interesting) 1 2 3 4 5 (very boring)
sex
women 1 2 11 4 2
men 2 6 7 4 2
lÀ Merit
Merit students 3 6 10 2 1
non-Merit students 0 2 7 5 3
usefulness (very useful) 1 2 3 4 5 (a waste of time)
sex
women 1 6 6 5 2
men 2 6 7 5 1
lÀ Merit
Merit students 2 10 6 4 0
non-Merit students 1 2 6 5 3
Q18 Students' feelings about mathematics at university
This question presented the same problems encountered for Q9. Many of the
students mentioned the more theoretical approach in the university course when
compared to theft school course, but most of the responses were not particularly
revealing.
Q19 Attitude towards the components of the mathematics course
 153

difficulty (very easy) 1 2 3 4 5 (very difficult)

algebra
women 0 4 8 7 1
men 1 4 9 7 0
analysis
women 0 0 6 9 5
men 0 4 7 8 2
calculus
women 0 6 11 3 0
men 0 3 12 6 0
interest (very interesting) 1 2 3 4 5 (very boring)
algebra
women 1 8 7 3 1
men 3 8 8 1 1
analysis
women 0 4 5 6 5
men 2 5 3 9 2
calculus
women 3 8 9 0 0
men 2 10 5 2 2
usefulness (very useful) 1 2 3 4 5 (a waste of time)
algebra
women 0 6 11 3 0
men 3 7 7 3 1
analysis
women 0 4 10 5 1
men 2 8 3 7 1
calculus
women 0 11 9 0 0
men 2 9 5 4 1
Q20 Change of opinion towards mathematics since coming to university
Just under a third of the students who said their opinion had changed mentioned the
more theoretical aspect of university mathematics compared to the treatment of the
subject at school. It was difficult to classify this type of comment as a positive or
negative statement and therefore responses of this nature were categorised as neutral.
students with
women men Highers A levels
No change of opinion 5 7 10 2
Change of opinion 15 14 19 10
Type of change: positive 1 3 4 0
negative 8 6 6 8
neutral 6 5 9 2
Q21 Attitude towards Applied Mathematics 1
Five students were exempted from the course. N women= 16, N men=20.
Ali: Non-Physical Applied Mathematics
Bh: Physical Applied Mathematics
 154

difficulty (very easy) 1 2 3 4 5 (very difficult)

Ah
women 1 10 5 0 0
men 6 9 5 0 0
Bh
women 0 3 2 5 6
men 0 1 4 9 6
interest (very interesting) 1 2 3 4 5 (very boring)
Ah
women 1 3 9 3 0
men 3 11 3 2 1
Bh
women 3 1 4 5 3
men 1 6 4 6 3
usefulness (very useful) 1 2 3 4 5 (a waste of time)
Ah
women 0 9 4 3 0
men 2 12 6 0 0
Bh
women 1 6 2 6 1
men 2 5 9 4 0
Q22 Aspirations concerning jobs involving mathematics
One man did not know whether he wanted a job using mathematics. Since Q42 asked
about the students' career plans, the responses to Q22(b) were not examined in any
detail. They consisted mainly of references to jobs in the financial field (accountancy
or actuarial work).
women men
Q22(a)
Students thinking they might like a
job using mathematics 18 16
Students thinking they would not like
a job using mathematics 2 4
Q22(c)
Students thinking they would get .a
job using mathematics 16 13
Students thinking they would not get
a job using mathematics 3 3
Did not know whether they would get a
job using mathematics 1 5
Q23 Students' assessment of the importance of the mathematics lecturer in
determining their enjoyment of the course
women men
Students finding the lecturer:
very important 5 10
fairly important 13 7
unimportant 2 4
Q24 Students' assessment of the important qualities of a lecturer
The students were asked to indicate their responses to this question on a pre-printed
 155

card.
very fairly unimportant
important important

Enthusiasm women 11 9 0
men 16 - 4 1
Clarity women 19 1 0
men 20 1 0
Interesting presentation women 15 5 0
of topic men 11 10 0
Pleasant manner women 4 15 1
men 3 14 4
Confident lecturing style women 10 10 0
men 17 4 0
Easy to gain access to in case women 6 11 3
difficulty men 6 10 S

Helpful attitude women 12 8 0

men 12 9 0

Being good at research women 1 5 14

men 1 5 15
Understanding the problems the women 18 2 0
students might have men 13 7 1

Q25 Students' views on obtaining help or information from staff

One women said she did not find it easy to obtain help or information from staff, 14
women and 13 men said they did, and 5 women and 8 men said they had never tried.
Q26 Students' interactions with staff outside class and tutorial times
The majority of students (17 women and 17 men) had never gone to see a lecturer
outside class times. The numbers who had never seen their tutor outside tutorials
were similar (15 women and 17 men). In view of this, there did not seem much point
in analysing the responses to the other parts of the question.
Q27 Feedback from staff
17 women and 11 men said they would like more feedback from staff on how well
they were doing and the staffs real opinions about their strengths and weaknesses.
Q28 Students' opinions on how the course is taught
The students were asked to indicate their replies on a pre-printed card. The
following table shows the numbers of women, men, Merit and non-Merit students
who agreed with the comments on the course. 1A Merits were used for this analysis:
22 students out of 39 obtained 1A Merits (two women were direct entrants to
Mathematics 2A and had not taken Mathematics lA).
 156

Merit non-Merit
women men students students

Impersonal compared with school 18 13 13 16

Challenging 17 19 20 14
Presented as too abstract 13 10 8 14
Not enough examples or explanations 17 8 11 12
Encourages exploration of the subject 7 9 8 8
Uninteresting 7 7 3 9
Presented too quickly, not enough
time given to assimilate 16 10 11 13
Q29 Type of solutions preferred
15 women and S men said they preferred written solutions, 3 women and 7 men
preferred an oral presentation of solutions, while 2 women and 6 men expressed no
preference.
The reasons given for preferring one method over another varied too much to permit
meaningful analysis. On the whole, reasons for preferring the first method involved
the advantage of having the solutions in front of one and being able to work at one's
own pace. The advantages of the second method were more detailed explanations
and being able to ask the tutor to clarify difficult points.
Q30 Use of books for coursework
Only 2 women and 3 men did not use books for the coursework. Of those who did, 2
men said they did not find them helpful.
Q31 Links between the lectures and the exercises
The object of this question was to ascertain whether the students saw the exercises in
the context of the lectures, as educational reinforcement in a sense, or whether the
theoretical knowledge transmitted in the lectures was not seen as directly applicable
to the exercises. However, the responses to the question implied that the students
were unsure about the point of the question and therefore they were not analysed in
any detail.
Q32 Factors influencing students to attempt exercises
The table below shows the numbers of men and women saying that the listed
considerations would influence them to attempt a question on a mathematics
worksheet. The students were given pm-printed cards to indicate their responses on.
women men
It looks easy 17 11
You think you could do it 19 18
It is a hand-in question 20 18
It was recommended by a lecturer or tutor 16 13
It looks interesting 14 17
It looks challenging 9 9
It seems important for the course 17 18
You usually attempt all questions 7 9

Q33 Reasons for not having completed questions on worksheets

The table shows the numbers of women and men agreeing to the various masons for
not having completed a question (Q33(à)). For Q33(b), 8 women and 10 men
thought that they could have completed the question if they had more time.
 157

women men
You think you've spent enough time on it 11 16
You're not interested in it 12 10
You don't think it's important for the
course 8 8
Q34 Reactions on not being able to complete mathematics exercises
The table shows the numbers of women and men who said they had the following
reactions on not being able to complete questions on worksheets.
women men
Forgetting about it 3 5
Looking up solution sheets 18 - 19
Asking classmates 14 15
Asking tutors or lecturers 9 15
Q35 Attitude towards performance in mathematics examinations
For Q35(a), 18 women and 20 men said they had sometimes felt that they had not
done as well in an examination as they would have liked. Q35(b) did not yield any
information of interest. The table below shows the numbers of women, men, Merit
and non-Merit students agreeing with the listed reasons for not doing well in
examinations (Q35(c)). Students who answered 'no' to Q35(a) were not asked this
question and the replies were indicated on pre-printed cards. 1A Merits were
considered for this analysis and so the total number of Merit and non-Merit students
does not include one direct entrant to 2A who said she had not done as well in some
examinations as she would have liked.
Merit non-Merit
women men students students
A difficult examination 12 12 10 13
Not having done enough work 12 16 16 11
Lack of ability 9 2 3 8
Not having studied the particular
topics in the exam 10 10 9 10
Bad luck 3 7 3 6
Not feeling well 2 0 1 1
Not being interested 6 2 4 4
Q36 Guessing at answers in an examination
18 women and 14 men said that they sometimes guessed at answers in a university
mathematics examination.
Q37 Going over questions after an examination
The object of this question was to obtain an indirect measure of how important
formal examinations were to the students. The assumption was that students who
went over questions after the event would be the ones who found examinations and
examination results more important. 14 women and 16 men said they went over
examination questions afterwards.
Q38 Comparing performance in mathematics with others
The question was intended as an indirect measure of a competitive attitude towards
performance. However, it could probably also be interpreted as indicating
uncertainty regarding ability and was therefore not considered as particularly useful.
 158

15 women and 18 men said they compared their performance in mathematics to their
classmates'.
Q39 Factors which students find encouraging
The table shows the numbers of women, men, Merit and non-Merit students saying
they felt encouraged by the following things. IA Merits were used for this analysis
and the students asked to indicate their responses on a pre-printed card.
Merit non-Merit
women men students students

Past successes 19 17 19 15
Comments by members of staff 16 16 18 12
Doing better than classmates 13 14 18 7
The hope of doing well in the future 16 19 18 15
Succeeding at something generally
seen as difficult 19 19 20 16
Seeing relationships between
different parts of a subject :14 13 17 9
Getting the gist of a subject
as a whole 19 20 21 16

Q40 Performance
Since the numbers involved in this survey were small, it was decided to assess
performance by whether a Merit was obtained. Assessing achievement by
considering examination results rendered the data rather sparse. Since Merit passes
take into account performance at the class examinations as well as the degree ones,
they have the advantage of reflecting performance in several examinations over the
whole year. For Mathematics 1A, 8 women and 14 men got Merit passes, of which 2
women and 11 men obtained 1st class Merits. (2 women in the sample had not taken
Mathematics 1A). For 2A, 5 women and 11 men passed with Merit, of which 2 and 7
respectively were 1st class.
Q41 Students' expectations regarding their performance
Since the students appeared more willing to guess at what grade they might get in the
Mathematics 2A degree examinations than whether they might obtain a Merit pass
for 2A, letter grades were used in order to study the patterns of overestimating and
underestimating performance. The students were asked what grade they thought they
might obtain and usually gave a numerical grade. This was converted into a letter
grade and compared with the actual grade obtained using the Departmental system
(A: 75% and over, B: 65%-74%, C: 55%-64%, D: 50%-54%, E: 45%-49%, F:
35%-44%, G:34% and under). When the estimated performance covered a band
(such as 60%-70%), the lower grade was chosen. Two women and one man said
they did not know what grade they might obtain and were therefore not included in
the analysis.
women men
Underestimating obtained grade 4 3
Obtaining grade predicted 4 7
Overestimating obtained grade 10 10
Q42 Career aspirations
 159

women men
Finance 8 7
Did not mention a career 10 6
Teaching 2 3
Research 0 4
Other 2 1
Management 0 1
Computing 1 o
 160

ffl.Ill. Questionnaire 3
running number l0. Could you please indicate on the card how important you feel the following
considerations were in influencing your duels jun to do a mutt's degree?
I vetyi.sport;itIt 2 tairly i,aportjiit 3 unimportant

I. sex 1 male Being good at maths at school (10)

2 female Finding maths easy

Finding maths interesting
1-diat degree are you doing? I BSc -
Thinking a maths degree would be useful for your career
2M (Li1 Knowing people who had taken a maths degree
Encouragement from teachers or other people
Was your last school I Scottish
2 English Were there any other reasons why you chose maths? 0 no
3 N. Irish 1 yea.
4 other 11111 Unprompted comments;

4. Did you do I Ilighere

2 A-levels
3 other type of exam - ('i) liii
Prompted comments: if yes, what were they?
5. What Highers/A- levelS did you do and what grades did you get?
ABODE
ABODE -4

ABODE
(If all responses are negative) Why did you do a maths degree then? There -4
A B C D E
must be some reason.
ABODE

6. 'fiat 515 Maths papers were (a) offered by your school I II 111 IV V
taksn by you I Ii III IV V
IL. Did you cu,,sider doing unyrhiug elsa? 0 110
What did you get?
1. What university courses have you taken? - lyss (ll)D
Unprompted cosnents:

S. How would you rate maths at school on a one to jive scale for
difficulty with I very easy and 5 very difficult Prompted consents: if yes, what else did you think of doing?
interest with 1 very interesting and S very boring
usefulness with I very useful and S a waste of time

9. How did you feel about it in 5 enerat at the time?

12. IThy did you choose to do the BSc rather than the MA? (or vice - versa)
Prompt: What was it about the course in particular that you
enjoyed/found intere.ting/etc?
 19. How would you rate the different components of Maths 2A for
13. Have you found chat schooL gave you adequate preparation to do a maths degree
difficulty with I very easy and 5 very ,Iiffics,It
at university? 0 no
I yea (13) Li interest tth I very and S very boring
use Cu Incas wit It I very useful and S a waste of time
uinprompted coassents:

difficulty interest uaefuIn1

algebra

calculus
Prompted comments: If no, in what way was it inadequate?
analysis

20. Has your opinion of maths changed since you came to univera ity? 0 no
I yea (20)
Unprompted comments:
14. How many were in your maths class and what was the proportion of girls for
no I girls
0-grade/0-- leve1
Higher/A-level Prompted consents: If yea, in what way has it changed?
$15 Maths (specify paper)

15. Is the proportion of women in the university maths degree course what you 0'
expected it to be? 0 no
1 yes (15)

lflsat do you think it is? *21. 110w would you rate the two halves of the Applied Matha I course for
It no, what had you expected? dsffi culLy with I very easy and 5 very difficult
interest wiLls 1 very iut'areating and 5 very boring
If tinder 401, why did you think there would be fewer women?
use Lu I 'tess wi U. I very useful and S a waste of t lao

difficulty interest I aefulneaa

16. How would you rate Maths IA in general for

::
(a) difficulty with 1 very easy and S very difficult (16) ft
(b) interest with I very interesting and 5 very boring
(c) usefulness with I very useful and 5 a waste of time 22. (a) Do you chunk you would like a job using maths? 0 no
I yea (22)
17. Now would you rate Maths 2A in general for if so, wIser Sort of job?
(a) difficulty (17)
Do you think you will get a job using maths? 0 no
(b) interest
(c) usefulness I You
Unprompted comments:
18. How do you feet about the university maths course as a whole?
lJnprompted comments:
Prompted cosumenta: It no for (a) and yes for (c) why du you think you will
if you dot,' t rustily want otto?
Prompted consents (write down prompt): If sea for (a) and no for (c), why not?
 23. low i.aoortant do you find the maths lecturer in determining how much Do you gene,- a I ly find it he I phil? 0 liD

you enjoy a particular topic? Is he or she very important (l), fairly I Ye5
important (2) or unImportant (3)?
LIII lIlilirileli ted cli4lelle lit a:

.24. Could you please indicate how important you feel it is for a oaths
lecturer to have the following qualities?
I vary important 2 fairly important 3 unimportant
Prompted comsle,lts: if no, why not?
Enthusiasm
Clarity I- I
Interesting presentation of topic t1
Pleasant canner 11
Confident lecturing style
11111 (b) Do you ever go and see your tutor outside tutorials? 0 no

LII
-

Easy to gain access to in case of difficulty

Helpful attitude
1_J
I I low often?
- I yes

King good at research [-1 Do you generally find it helpful? 0 no

Understanding the problems you may have 11111 Unprompted coments:

1 yea LIII
Are there any other qualities you feel are important? 0 no

Unprompted comments:
lyes LIII -

Prompted comments: If no, why not? ON

U)

Prompted comments; if yes, which ones?

Would you like! more feedback from the staff on how you are doing and
on their real opinions about your strengths and weaknesses? n no
1 yes (21) [III
*28. could you indicate whether you agree or disagree with the following
Do you find it easy to obtain help or information from the staff in
comments on how the course is taught? 0 disagree I agree
the Mathematics Depart sent? 0 no
1 yea Impersonal compared to school (28)
(25)
Unprompted commeois: Challenging
Presented as too abstract
Not enough examples or explanations
Encourages- exploration of the subject
Uninteresting --

Prompted casements: if no, why do you find it difficult?

Presented too quickly, not enough time given to assimilate

Do you have any other comments on the course? 0 no

- 1 yes
Unprusipte d Lueseent 5:
(a) Do you ever go and see your lecturer outside class times? 0 no
1 yes (a)
(26) Prompted comments-. If yes, what are they?
How often?
29. Do you prefer good written solutions or good oral presentation of
ft luoks snterust log
soi ut ions in tutorsa Is? 0 writ ten solution. --
It l,'kscl,:, I lcngiiq
I oral solutions --
It sue,,,a itvh.t niL Inc the cow-au
2 no preference (29) _J
Vu,, usual ly utts'euut alt the questtons
Unprompted comments:
Are there any other reasons you can thin!; of which would influence you
Co try a question? 0 no -

1 yes [J
Unprompted commerce
Prompted comments: If 0 or I, why do you prefer that method?

Prompted comments: If yes, which ones?

30. Do you use books for the the cour saworkI 0 no

I yes (30) Lii
Do you find them helpful? 0 no
i yes ElIl (a) 4hen you give up a question you've attempted without completing it,
is it generally because
Unprompted coents;
you think you've spent enough time on it? (a)

you're not interested in it?

you don't think it's important for the course?

Prompted comments; If no. why not? (b) For ouestionm you don't complete, do you generally feel that you
If yes, in what way do you find them helpful? could have completed them if you had more time or less other work -

to do?
U
What do you tend to do whe,, you deride you cannot do an exercise on
a e,uths workslueui t? Do you
31. Do you find that the exercises and the maths course conplement each other? forget about it? (34)
0 no - 1 yes
Unprompted comments:
(31) LIII look no the solution sheets?
ask your classmates?
ask your tutor or lecturer? -

what did you use to do at school?

Prompted comments: If no, why not?
If yes, in whet way do they complement each other?

35. (5) Isv, you ever felt that you did not do as well in an exam as you
*32. Could you read this list and indicate which of the following considerations
would have liked to? This includes your ma in school maths exams
would influence you to attempt an exercise on a maths worksheet? 0 no
as well as your class and degree maths exams at university.
1 yes
(b) Which ones?
It looks easy (32) -
You think you can do it -
It is • hand-in question
It was recommended by a lecturer or tutor
 *(c) Could you indicate the reasons you did not do well? 0 no I yes 41. IThat do you think you might get in the March 2A exam?

The exam was hard FII (c) Mint do you think you might get in the ZA degree exam?
You hadn't done enough work - -

Lack of ability Do you expect a Merit for 2A? 0 no I yes (61)

You hadn't studied the particular topics in the exam If so, what class?
Bad luck
You weren't feeling well 42. Do you have any idea what you went to do after you graduate?
You weren't interested o no i yes
Unprompted constants:
(62) LIII
Have you ever guessed at an answer in a university maths exam
without being sure? 0 no 1 yes (36) 1111
Do you go over exam questions after a naths exam to see what you Prompted constants: If yes, what?
did wrong? (") LIII
Do you tend to compare your performance in maths to that of others
in your class? (38) LII
*39, Could you indicate whether you feel encouraged by these things?

Past successes (39)

Comments by lecturers, tutors or Directors of Studies
-5
Doing better than your classmates a"
C)?
The hope of doing well in the future
Succeeding at something generally seen as difficult
Seeing relationships between different parts of a subject
Cetting the gist of a subject as a whole

Is there anything else that encourages you? 0 no 1 yes

Unprompted constants:
[1111

Prompted comments: If yes, whet?

40. 'Stat did you get for

tintha IA class exams: Dec

liar

degree exam

Did you get a Merit? 0 no L yes

If so, what cLass?
(40) LI1
Did you expect to get it?
'I :ti .5 I. no,
10 21

low import ant do you feat the to! towing con. Idaret toils ware In iflfltieitc I
your decision to do a math, degree? Please circle the respunse you feel
How would you rate the two halves of this Applied Maths I routes for
applies the best.

fairly difficulty with (1) being very easy and (5) very difficult
very
Being good at the at school unimportant
Important important
interest with (I) being very interesting and (5) very boring
very fairly
Finding maths easy treat i.p.LLO.L unimportant
I. usefulness with (1) being very useful and (5) a waste of tias
very fairly
Finding maths interesting important important
unimportant
difficulty I interest usefulness

Thinking a maths degree would be useful for very fairly

unimportant Ala
your career important important
non-mechanical
component
very fairly
Knowing people who had taken a maths degree unimportant
Important important - Bh
very fairly mechanical
Encouragement from teachers or other people unimportant
important important component

19 24
I
How important do you feel it is for a maths lecturer to have the following
qualities? Please circle the response you feel applies the best.

Emit Ito ai em very fairly

Important wilsalport alil
important
Clarity very fairly
How would you rate the different components of the Maths 2A course for important unimportant
important
difficulty with (I) being very any and (5) very difficult Interesting presentation of topic very fairly
important important unimportant
interest with (I) being vary interesting and (5) very boring very fairly
Pleasant manner on impor taiii
important important
usefulness with (I) being very useful and (5) a wsste of time
Confident lecturing style very fairly
important unimportant
important
difficulty interest usefulness
Easy to gain access to its case of difficulty very fairly
important important unimportant
algebra
Helpful attitude . very fairly
important unlmport alit
important
calculus
Being good at research very fairly -

important unlmportalit
important
analysis
Understanding the problems you may have very fairly -

important unImportant
important
28 - 35(c)

Would you agree or disagree with the following co,ents on how the mail's course is taught?
G,us iduriuls clue times WIAeo you felL that you did UOL do as we' I in a si:itlimi exaam as
Impersonal compared to school agree disagree
you would have liked to. would you say that it was because

Challenging agree disagree

The exam was hard yes no

Subject presented as too abstract agree disagree

You hadn't done enough work yea no

Not enough examples or explanations agree disagree

Lack of ability yea no

Encourages exploration of the .object agree disagree

You hadn't studied the particular topics in the exam you no

Uninteresting agree disagree

Bad luck yea no

Presented too quickly, nor enough time given

agree disagree You weren't feeling well yes no
to assimilate

You weren't interested yes no

C'
32
39

iasich of the following con.iderationa would influence you to attempt a maths exercise
on a worksheet? Please reed down the list of suggested reasons and circle the appropriate
Do you find you feel encouraged by
response for each.

yes no Past successes yes no

It looks easy

yes no Consents by lecturers, tutors or Directors of Studies yes no

You think you can do it

yes no Doing better than your classmates yes no

It's a hand- in question

It was recommended by. lecturer or tutor yea no The hope of doing well in the future yes no

It looks interesting yes no Succeeding at something generally seen as difficult yes no

it looks challenging yes no Seeing relationships between different parts of

a subject yes no

It seems important for the course yes no

Getting the gist of a subject as a whole yes no

You usually attempt all exercises yes no

 S.

IV. Entrants to Mathematics degrees

The table on the next page shows the entry figures for Mathematical Sciences
degrees for each university for the years 1985-1987 inclusive. Entrants to Computer
Science degrees are not included.
 169

women total entrants

Aston 4 14
Bath 67 238
Birmingham 74 222
Bradford 33 74
Bristol 58 236
Brunel 40 126
Cambridge 115 697
City 30 72
Durham 59 216
East Anglia 25 ill
Essex 42 164
Exeter 104 302
Hull 55 160
Imperial (London) 35 161
Keele 9 33
Kent 30 118
King's (London) 39 125
Lancaster 23 77
Leeds 111 281
Leicester 39 129
Liverpool 96 250
London School of
Economics (London) 9 20
Loughborough 23 74
Manchester 75 355
UNIST 42 112
Newcastle 73 209
Nottingham 58 209
Oxford 169 619
Queen Mary College
(London) 67 '229
Reading 37 100
Royal Holloway and
Bedford (London) 73 169
Salford 58 171
Sheffield 105 325
Southampton 98 332
Surrey 45 112
Sussex 58 175
University College
(London) 40 126
Warwick 111 425
York 71 210
Aberdeen 30 69
Dundee 21 46
Edinburgh 79 193
Glasgow 86 207
Heriot-Watt 63 201
St Andrews 49 122
Stirling 6 19
Strathclyde 75 160

Source: Universities' Statistical Record

 170

V. Degree results for Scotland and England

The following table shows the numbers obtaining each class of Mathematical
Sciences degrees for the years 1985-1987 inclusive. Computer Science degrees are
excluded.
The figures also exclude the Cambridge results due to the difficulty of classifying its
degrees. The USR returns from Oxford for 1985 had been misclassified, and so the
Oxford results are for 1986 and 1987 only.

1st 2.1 2.2 3rd Pass/Ordinary

England women 320 584 788 375 81 N=2148
men 665 1151 1319 788 293 N=4216

Scotland women 48 76 88 49 71 N=332

men 100 105 110 53 105 N=473

Source: Universities' Statistical Record

 171

VI. Published papers

TEACHING MATHEMATICS AND ITS APPLICATIONS Volume 6 No. 4, 1987 157

Women and Mathematics: A Different Pictures

ELIZABETH J. P. FRASER and SHEILA CORMACK
Department of Mathematics, University of Edinburgh

DURING the summer of 1986, the Centre for students of 1: 1 while the national average is
Mathematical Education at Edinburgh Univers- 2:12 In other words 50 per cent. of mathematics
ity conducted a survey of all students who undergraduates at Edinburgh University were
entered Edinburgh University between 1978 and female compared with the national average of 30
1981 to do a mathematics degree, including per cent. for female mathematics undergradu-
those who subsequently obtained a mathematics ates. This figure was obtained by considering
degree after a change of course. The object of those students whose course on entry to
the research was to establish whether or not university had been a BSc or MA mathematics.
there were significant sex-related differences in This did not include mathematical physics which
mathematical achievement at university level as has a predominantly male entry. Due to course
there appear to be at secondary school level. A changes, the proportion of female mathematics
reference bank of previous research was compiled graduates was 48 per cent. The main results are
and used as a starting point for the analysis of summarised as follows:
data from the survey.
The results were consistent in some respects Mathematics graduates who Males, Females
with those of previous studies. Using a score percent. percent.
system to compare Higher or A-level results, we obtained a first class Honours degree 17 14
found that the modal score for males was the obtained a first or upper second 43 39
obtained a lower second or third 31 36
maximum 15 while that for females was 13 (22 obtained an Ordinary degree 26 25
per cent. of males scored 15 with 16 per cent. of Males Females
females doing so). There was no sex-related entrantsto a mathematics degree course
difference for the mathematics score, but this who failed to obtain a degree

was probably due to the nature of the sample

since we were concerned with mathematics As a comparison, the national average for
students in particular. Males took more science mathematics graduates obtaining first class
subjects than females, both at school and Honours degrees is 10per cent. for females and
university level, with 76 per cent. of males taking 14 per cent. for males.
6 science subjects during their first 2 years at The main differences between Edinburgh
university compared with 52 per cent. of University and the national averages stem from
females. When considering school results, we the fact that almost 50 per cent. of mathematics
found that in general those who took 2 or more graduates at Edinburgh University were female,
science subjects in addition to mathematics compared to the national average of 30 per cent.
obtained better mathematics scores. This trend These differences are represented by the
was also found to a lesser extent at university barchart (Fig. 1) which illustrates the impact of
level where there was a correlation between the the greater proportion of females on the
number of science subjects taken and the class of percentages of males and females awarded each
degree obtained. grade of degree.
When considering the male/female ratio of As yet we do not have enough information to
undergraduate mathematics students and the establish whether these results are typical of
distribution of degree class, we found that our Scottish universities in general, but preliminary
figures were not typical of the national average. data from secondary schools indicate that there
We found a male/female ratio of mathematics is a difference in achievement trends between
© The Institute of Mathematics and Its Applications 1987
158 TEACHING MATHEMATICS AND ITS APPLICATIONS Volume 6 No. 4, 1987

First Class Other Honours Degree Pass Degree

Female Male Female Male Female Male

Edinburgh 6.7 8.8 29.3 29.7 12.0 13.5

University

Great
Britain

1
9.9
I 22.

5 1 53.5

Fig. 1. Percentage of mathematics degrees awarded

I I
I3.6
7.6

Scottish schools and those in England and they are both prerequisites for university and
Wales. It therefore seems reasonable to assume therefore the greater percentage of girls sitting
that other Scottish universities would have Higher mathematics might explain the lower
similar enrolment and achievement patterns to male/female ratio at Edinburgh University
those found at Edinburgh University. compared to all UK Universities.
Data from the Department of Education and It seems probable that later specialization in
Science (DES 1977-80) and the Scottish Scottish schools accounts for the difference in
Education Department (SED 1981-84) show female representation: Scottish students gene-
that there is a smaller sex-related difference in rally take 5 or more Highers while English
0-grade mathematics pass levels in Scotland students usually take 3 A-levels. The lack of
than there is for 0-level mathematics in England intense pressure to specialise at an early age
and Wales. In Scotland 29 per cent. of male S4 (roughly 14 for English students) may encourage
pupils and 26 per cent. of female S4 pupils pass more Scottish girls to continue with mathematics
0-grade mathematics while in England and rather than drop it in favour of "easier" or
Wales 28 per cent. of male and 21 per cent. of "more feminine" options.
female school leavers pass 0-level mathematics. Studies in America' have indicated that
This is because more girls in Scotland actually there is little sex difference in school mathe-
sit the mathematics examinations compared matics scores when the scores are weighted
to England and Wales where girls give up according to the extent of mathematical studies
mathematics at a relatively early stage. 39 per pursued. This implies that there is a positive
cent. of female S4 pupils sit 0-grade mathe- correlation between mathematical achievement
matics and only 27 per cent. of female leavers sit and the amount of mathematical and scientific
the 0-level examination. 44 per cent. of leavers background. Since girls tend to drop mathe-
sitting 0-level mathematics are females com- matics and hard science subjects more often than
pared to 48 per cent. for Scotland. The same boys, we would expect to find that girls obtained
patterns are evident when considering Higher lower scores compared to boys. The tendency to
and A-level mathematics: 40 per cent. of girls avoid scientific subjects is generally considered
and 58 per cent. of boys taking Highers sit the to be due to pervasive gender stereotyping
mathematics paper. For A-level mathematics, the coupled with the pressure for early specialisa-
figures are 20 per cent. for girls and 47 per cent. tion. The pressure occurs at a period when girls
for boys. 44 per cent. of those taking Higher become increasingly aware and concerned about
mathematics and 27 per cent. of those taking their femininity or lack thereof. Since mathe-
A-level mathematics are females. The pass rates matics is considered a fairly masculine subject, a
for mathematics Highers or A-levels show no girl who is worried about her femininity and
sex-related difference but do show an wishes to "fit in" would be likely to drop
examination-related difference: the pass rate is mathematics in favour of some more "feminine"
64 per cent. in Scotland and 74 per cent. in subject or simply underachieve in whatever
England and Wales. The grade distribution mathematics she continued to study. She thereby
shows sex-related differences at both Higher and conforms to the traditional female model. So far
A-level with boys being more likely to obtain our results concur with this line of research, but
A's than girls who cluster about the B level. more study is needed to ascertain whether our
While A-levels and Highers are not equivalent results are typical of Scotland in general.
 TEACHING MATHEMATICS AND ITS APPLICATIONS Volume 6 No. 4, 1987 159

References
University Grants Committee, 'University Statistics Elizabeth J. P. Fraser graduated MA in
1982-83, vol I," 1983. mathematics from Edinburgh University in 1986
The Royal Society and Institute of Mathematics and its She held a vacation scholarship August-
Applications 'Girls and Mathematics," 1986. September 1986 and is currently a research
Fox, L., Brody, L,, and Tobin, D., Editors, "Women and assistant in the Department of Mathematics at
the Mathematical Mystique," 1980.
Edinburgh University, working on Women and
Mathematics.
Sheila Cormack graduated MA in mathematics
from Cambridge University in 1967 and obtained
her PhD in topology from Manchester University
in 1971. She has been a lecturer in mathematics
at Edinburgh University since 1972 and Director
of Studies since 1978. Her research interests
include Women and Mathematics.

Puzzles, Challenges and Investigations

These are designed by Hugh and Joyce Porteous. Please send solutions to Hugh Porteous, Department
of Mathematical Sciences, Sheffield City Polytechnic, Pond Street, Sheffield Si 1WB. We would also
welcome further suggestions but please send your solutions.

Cubism ABEECCDDD
Write down any positive ,integer in standard MFEEGGD
FFFEHSR
denary (base ten) form: Calculate the sum Fit I H 6
of the cubes of its digits to obtain a new I J J I H
3 K .1 H
positive integer. Repeat this process twenty 3 fr H
times. There are just fifteen possible final K K
answers. What are they?
Fig. 1
Solitairy confinement Notice that each group of five is adjacent
The familiar game-of Solitaire is played-on a but that no two blocks are the same shape.
33 hole board in the form of two crossing She now wishes to plant up a second plot in
3 x 7 rectangles. Each move consists of the same way, except that the 11 patterns
jumping a piece over a neighbouring piece, must not repeat ones used already in the
horizontally or vertically, into a vacant hole, first plot (mirror images are not allowed
and removing the piece jumped over. One either). How can she do this?
hole is vacant at the beginning and the 16. Obtuse integers
object of the game is to leave just one piece It is well known that triangles with sideü of
on the board at the end. Can this game be length 3, 4, 5 or 5, 12, 13 have a right angle.
played on a simple 5 x 5 square? It is not so well known that those with sides
15.- Pent-up gardener of length 3, 5, 7 or 7, 8, 13 have an angle of
A gardener has planted a triangular plot 1200 . Find other examples of this,
with 55 plants, five of each of 11 different preferably all those with sides of integer
varieties as shown in Fig. 1. length less than 100.

© The Institute of Mathematics and Its Applications 1987

J. R. Statist. Soc. A (1992)
155, Part 2, pp. 241-258

Female Participation in Mathematical Degrees at English and Scottish

Universities
By G. C0HENI and E. J. P. FRASER

University of Edinburgh, UK

[Received September 1989. Final revision April 1991]

SUMMARY
Among entrants into mathematical degrees in English and Scottish universities in 1985-87
the proportion of women varied greatly between universities and was generally higher in
Scotland than in England. Among English universities those with higher A-level require-
ments had lower proportions of women mathematicians, probably as a direct result of the
smaller proportion of women available with the proper qualifications for entry. The
speculation that women might be discouraged from studying mathematics at particular
universities that are technologically orientated or have higher entry requirements is
examined with quite crude data but not supported. Sex differences in mathematics degree
results are small but show a pattern found in other studies, namely greater variation in
attainment for men than for women.
Keywords: MATHEMATICS DEGREES; SEX DIFFERENCES

1. INTRODUCTION

Among numerous current educational concerns the shortage of well-qualified

graduates with a mathematical training must surely be counted as important.
Particularly severe is the problem of attracting such graduates into teaching. A related
concern is the level of entry of females into scientific and mathematical degrees. It is
widely believed that higher participation by girls in such disciplines would not only
help to alleviate current and prospective shortages but also, by counteracting negative
perceptions, enhance career opportunities for girls and improve the equality of
opportunity between the sexes.
A considerable amount of research has been conducted into differences at school
level between boys' and girls' attitudes to and ability in mathematics and their
examination performance (Shuard, 1981; Girls and Mathematics Association, 1984;
Chipman etal., 1985; Burton, 1986; Walkerdine and Girls and Mathematics Unit,
1989; Department of Education and Science, 1989a). Far less research has been
devoted to sex differences in mathematics at university level. This paper examines
some of the statistics available regarding variation between universities in entry
requirements for mathematics and in the proportions of mathematics students who
are women. The interpretation of these statistics presents difficulties but some
interesting relationships emerge. In particular, we have compared the English and
Scottish universities, since an initial motivation for the research was the observation
of a rather high female participation rate in mathematics at Edinburgh University,

tAddress for correspondence: Department of Mathematics and Statistics, University or Edinburgh, James Clerk
Maxwell Building, The King's Buildings, Mayfield Road, Edinburgh, EH9 3JZ, UK.

© 1992 Royal Statistical Society 0035-9238/92/155241 $2.00

242 COHEN AND FRASER [Part 2,
and this prompted the speculation that the broader-based System of school education
in Scotland might be a factor in encouraging girls to continue their studies in mathe-
matics. Investigation of this hypothesis would be best carried out through studies of
individuals. Here, although we have discussed some hypotheses, we have largely
confined ourselves to the presentation of the statistics available to us.

2. PERFORMANCE IN SCHOOL EXAMINATIONS

Research into sex differences in mathematical activities has been conducted from
the nursery through to the fourth year of secondary school and beyond (Walkerdine
and Girls and Mathematics Unit, 1989). It has been argued that by the fourth year of
secondary school many girls have been put off mathematics by a complex of social
and psychological processes in the classroom, while boys, even of little ability,
recognize that they will need mathematics for their careers and are willing to study it
further. As we are concerned here with university entry, we start by considering
performance at General Certificate of Education (GCE) 0-level and Scottish
Certificate of Education (SCE) 0-grade, leaving aside the indubitably important
matter of differential achievement and sex imbalance among those who do not take
these examinations. We recognize that the introduction of General Certificate of
Secondary Education and Scottish Standard grade examinations make the statistics
presented of somewhat historical interest, but the phenomena underlying the sex
differences revealed are unlikely to have been swept away by recent educational
reforms.
Table I shows the percentage of girls among the entrants to school mathematics
examinations and among those passing at various grades. In the mid-1980s, almost
half the entrants for GCE 0-level mathematics were girls, but girls were markedly
under-represented in the top two grades. In Scotland this effect was much smaller,
although still present, at 0-grade and only by the stage of the Higher grade examina-
tion was the pattern of female under-representation similar to that at 0-level in
England and Wales. (For those unfamiliar with the Scottish system, it should be made

TABLE 1
Female participation and performance in school mathematics, 19841

Subject % girls % girls among those with passes at (he following grades:
among entrants A B C A-B A-C A-E

GCE 0-level mathematics 47 36 43 48

SCE 0-grade mathematics 49 45 48 50
GCE A-level pure and 30 25 28 29
applied mathematics
SCE Higher mathematics 45 36 41 45
GCE A-level further 23 15 22
mathematics
SCE CSYS paper It 31 33 37
(calculus)t

tSources: Royal society and Institute of Mathematics and its Application (1986) (GCE papers) and special tabulations
provided by the Scottish Examination Board (SCE papers).
tScottish Certificate of Education, Certificate of Sixth Year Studies: of the five mathematics papers, paper 11 has by
far the largest entry.
19921 FEMALE PARTICIPATION IN MATHEMATICAL DEGREES 243
clear that, whereas the English 0-level and Scottish 0-grade are approximately
comparable examinations, taken at a similar stage of the school career, the Scottish
Higher grade examination is taken after a year of post-0-grade study, usually of
about five subjects, and is recognized by the Scottish universities as an entry qualifica-
tion for their four-year degree course.) By the GCE A-level stage, girls are severely
under-represented among entrants and, although there is little sex difference in
overall pass rates, fewer girls than boys achieve grade A. At further mathematics A-
level the representation of girls is again reduced among entrants and the percentage of
girls achieving grade A is again lower than for boys.
Some evidence regarding the change in the participation of females in mathematics
between 0-level and A-level is contained in a survey carried out in England and Wales
by the Office of Population Censuses and Surveys (OPCS) (1987) into young people's
intentions to enter higher education. It was found that, of those taking any A-levels at
all, the percentage taking mathematics A-level (defined as the 'take-up' of mathe-
matics A-level) was higher for girls at independent schools than for girls at maintained
schools, whereas for boys the reverse was the case. It was suggested that the greater
prevalence of single-sex schools in the independent sector could be associated with this
effect. It was also found that, among boys, the take-up of mathematics A-level was
not associated with social class, but for girls there was a lower take-up among those
from social classes III-V. There was a similar association with parental education:
among girls, those with one or both parents lacking qualifications after the age of 18
years were less likely to take up A-level mathematics, but among boys only those
whose parents both had degrees were any more likely than others to take up A-level
mathematics.
Respondents in this survey were also asked why they had chosen not to take A-level
mathematics after having sat 0-level mathematics. Sex differences were not great, but
girls stated more frequently than boys that they thought A-level mathematics would
be too difficult, this difference being apparent in both independent and maintained
schools and further education colleges, and at all levels of total A-level passes (Office
of Population Censuses and Surveys (1987), Tables 6.7 and 6.8). More girls than boys
also gave lack of enjoyment or interest in mathematics as a reason for not taking it at
A-level. However, the survey found that lack of interest was related to high academic
achievement in terms of total A-level passes, whereas expecting difficulty in mathe-
matics was related to low achievement. The survey concluded that brighter students
were more likely to be put off mathematics by lack of interest than by expected
difficulty whereas for weaker students the reverse was true, equally for boys and girls.
Low female participation at A-level is not confined to mathematics; physics and, to
a lesser extent, chemistry have many more male candidates over females. Conversely,
there are many more females than males in the entry for biology and arts subjects such
as English and French. The result is that girls are less likely to combine A-level mathe-
matics with other science subjects: among 1987 school-leavers with an A-level pass
in mathematics, an estimated 53% of the boys had A-level passes in three or
more science subjects compared with only 39% of the girls (Department of Educa-
tion and Science (1987), Table C16, based on a 10% survey of school-leavers).
We do not know whether this difference has any association with grades achieved,
but it could be that the reinforcement of mathematics in their other A-level studies
helps boys to achieve better at the higher grades as shown in Table 1. Notwith-
standing an increasing quantitative element in biological science and some social
244 COHEN AND FRASER [Part 2,
sciences the picture is that, for many girls who take mathematics as one of their A-
levels, mathematics is less central to their current and intended future studies than it is
for boys. The OPCS survey found that, of girls who had taken an A-level in
mathematics, statistics or computer science and had gone on to higher education,
20 010 were studying medical subjects and a further 27% were studying other pro-
fessional or vocational subjects, but only 13% were studying mathematical sciences or
physics. The corresponding figures for boys were 5%, 18 07o and 27% with a further
29 010 studying engineering or technology (Office of Population Censuses and Surveys
(1987), Table 5.9).
This situation is of course not static, although we cannot here discuss trends over
time. We merely remark that female participation in mathematics at A-level increased
steadily throughout the 1980s: among secondary school pupils taking A-level mathe-
matics in their seventh or later year, the percentage of girls rose from 29 016 in 1981 to
34 076 in 1988 (Department of Education and Science, 1989b). (There appears to have
been a slower change in Scotland. In 1988 the percentage of girls among leavers with a
Higher grade pass in mathematics was 45.3% (Scottish Education Department, 1990).
This is not quite comparable with the 45% (actually 44.6%) figure for 1984 in Table 1,
as that relates to passes in a given year rather than leavers in a given year, but it does
suggest a fairly static position in Scotland.) Among those passing A-level mathema-
tics, the percentage of girls rose from 22 076 in 1974 to 26 010 in 1980 and 29 076 in 1985
(Department of Education and Science, 1985). In addition, the current oversupply of
medical graduates and the perceived shortage of mathematics and physics graduates
must surely be affecting patterns of choice of subject.

3. FEMALE ENTRY INTO UNIVERSITIES AND ENTRANCE REQUIREMENTS

To study sex effects on entry into university mathematical degrees, a good strategy
would be to look at individual applicants' A-level grades, the entrance requirements
of the universities that they apply to, the levels of offers made and whether or not they
are accepted, incorporating additional university-specific variables to account for
'attractiveness' or 'prestige'. However, individual students' grades are confidential
and the Universities Central Council on Admissions (UCCA) will not even release
statistics from individual universities without permission. For these and other
reasons, the data studied here are not ideal.
We obtained data from the Universities Statistical Record showing the numbers
of male and female entrants into mathematical science degrees at English and
Scottish universities in the period 1985-87. (By an oversight, the University of Wales
was excluded from the data request.) Table 2 shows the percentage of females among
entrants to these degrees (combined degrees and statistics are included but computer
studies are excluded). There is a wide variation in the proportion of women, from
47% at Strathclyde to 16% at Cambridge. Six of the first 12 universities in the table
are Scottish, which suggests a higher propensity for girls to study mathematics in
Scotland. Table 3 compares English and Scottish universities with regard to female
entrants into mathematical and computing degrees. Whereas for computer studies
there is very low female participation in both countries, it is clear that in other mathe-
matical degrees there is substantially higher female participation in Scotland.
Statistics degrees have a particularly high female participation rate (40 07o in England:
owing to difficulties with subject classification, the figures for female participation in
19921 FEMALE PARTICIPATION IN MATHEMATICAL DEGREES 245
TABLE 2
Percentage of women among entrants into mathematical degrees, English and Scottish Universities
1985-871

University % women No. of University 010 women No. of

entrants, entrants,
1985-87 1985-87

Strathclyde 47 160 Stirling 32 19

Dundee 46 46 Sheffield 32 325
Bradford 45 74 University College London 32 126
London School of 45 20 Heriot-Watt 31 201
Economics Loughborough 31 74
Aberdeen 43 69 King's College London 31 125
Royal Holloway and 43 169 Southampton 30 332
Bedford College Lancaster 30 77
(London) Leicester 30 129
Glasgow 42 207 Aston 29 14
City University (London) 42 72 Queen Mary College 29 229
Edinburgh 41 193 London
Surrey 40 112 Nottingham 28 209
Leeds 40 281 Bath 28 238
St Andrew's 40 122 Keele 27 33
University of Manchester 38 112 Durham 27 216
Institute of Science Oxford 27 619
and Technology Essex 26 164
Liverpool 38 250 Warwick 26 425
Reading 37 100 Bristol 25 236
Newcastle 35 209 Kent 25 118
Salford 34 Ill East Anglia 23 III
Exeter 34 302 Imperial College London 22 161
Hull 34 160 Manchester 21 355
York 34 210 Cambridge 16 697
Sussex 33 175
Birmingham 33 222 Total 31 8795
Brunel 32 126

Degrees in mathematics, statistics and combined degrees with a mathematics or statistics specialization (source:
Universities Statistical Record).

statistics degrees in Scotland are not reliable).

We did not obtain comparable data on entrants for non-university higher
education. However, the Council for National Academic Awards (CNAA) gives
statistics for first-degree graduates in mathematical sciences from non-university
institutions in the whole of the UK (see Table 7, later). These show that in 1986-88
women comprised 18% of the Honours graduates in computing and 35% of the
Honours graduates in mathematics, statistics and combined degrees (see, for
example, Council for National Academic Awards (1989)). The corresponding figures
for graduates from English and Scottish universities (1985-87) are 13 07o and 34%.
Given the higher female participation in SCE mathematics examinations compared
with GCE examinations (Table 1), it is unsurprising to see a continued higher female
participation at university. Indeed the 29% of girls among those passing A-level at
grades A-C matches remarkably closely the 29.6% females among entrants to English
university mathematical degrees, as does the 41% females among the SCE Higher
passes at grades A-B with the 40.2 010 females among entrants to Scottish university
mathematical degrees. Since we know from Table 1 that fewer girls are available with
246 COHEN AND FRASER [Part 2,
TABLE 3
Percentage of women among entrants into university mathematical and computing degrees, 1985-871

Mathematics, statistics Computer studies

and combined degrees women Total entrants
women Total entrants

English universities 29.6 7778 9.2 3498

Scottish universities 40.2 1017 11.2 651
Total 30.8 8795 9.5 4149

tSource: Universities Statistical Record, special tabulations for 1985-87

A-level or SCE passes at the top grade, it is natural to ask whether the variation
between universities within each country may be related to entry requirements.
Actual (as opposed to minimum) university entry requirements vary from year to
year and are no doubt treated flexibly as many other factors may affect acceptance.
However, some information may be gleaned from Association of Commonwealth
Universities (1988), which gives for each university mathematics course the 'typical'
A-level grades asked for by the university in their offers to students (e.g. ABB) and an
indication of the range of A-levels actually accepted in the intake for the previous
year. Unfortunately, the guide is not available before 1988; however, using data from
the 1988 guide, which relates to the 1986 entry, we can make some comparisons for
English universities. A similar exercise could be carried out for Scottish universities by
using the Highers requirements given in the annual entrance guides published by the
Scottish Universities Council on Entrance (SUCE); however, for several reasons
discussed below, this analysis is not very satisfactory.
Table 4 shows that, among English universities that required a grade A in their
typical offer for mathematics entrants in 1986, 24.8% of the entrants were women,
whereas among those that only asked for a grade B 34.3 07o of entrants were women. It
is probable, though not necessarily always the case, that for entry to a mathematical
degree the highest grade required in a university's typical offer is required in the A-
level mathematics paper. A comparison of Tables 1 and 4 shows that the 25% women
among those passing at grade A in mathematics A-level in 1984 matches remarkably
closely with the 24.8% women among entrants to universities that required a grade A
in 1986. Similarly, the 15 010 women among those passing at grade A in further
mathematics matches rather closely the 16 076 women among entrants to Cambridge,
which has the highest entry requirement (grades AAA). (The percentage of women
among those achieving grade A in further mathematics is based on small numbers and
is subject to appreciable random variation from year to year, so that this match is
somewhat fortuitous.) The discrepancy between the 34.3% women among entrants to
universities requiring a grade B and the 30% women among those passing at grade B is
not large and can be attributed to variations between offers and intakes as well as the
fact that many who obtain a grade B in A-level mathematics will undertake a degree
other than mathematics (as indeed will some who obtain grade A).
1985-87 was a period in which there were sufficient university places in
mathematical degrees for nearly all suitably qualified applicants—i.e. supply
exceeded demand and there was relatively little competition for places. This probably
19921 FEMALE PARTICIPATION IN MATHEMATICAL DEGREES 247
TABLE 4
Association of percentage of women among entrants into university mathematical degrees with highest
grade required in typical offers

Highest grade Universities women among entrants

required (1985-87)
(1986 entry)

A Bath, Bristol, Cambridge, Durham, East Anglia, Hull, 24.8

Leicester, Loughborough, Manchester, Nottingham,
Oxford, Reading, Warwick, Imperial College London

Aston, Birmingham, Bradford, Brunel, City 34.3

University, Exeter, Keele, Kent, Lancaster, Leeds,
Liverpool, University of Manchester Institute of
Science and Technology. Newcastle, Salford, Sheffield,
Southampton, Surrey, Sussex, York, King's College
London, London School of Economics, Royal
Holloway and Bedford College, Queen Mary College
London, University College London

(C Essex 26.0)

tOrades are taken from Association of Commonwealth Universities (1988), Table 15. For Loughborough and
Sheffield the course with the largest entry was used. The figures are percentage women among entrants into these
universities for all mathematical science degrees, including combined degrees, but excluding computer studies.

accounts for the remarkably good agreement between Tables 1 and 4.

Further analysis may be carried out if we adopt the common A-level point Scoring
system of S for a grade A, 4 for a grade B, 3 for a grade C, etc. Then for each university
we may calculate the total point score of the typical offer and the middle value of the
range in total point scores of the intake. The correlations of each of these with the
percentage of women among entrants (not weighted by size of entry) are respectively
—0.49 and —0.45. As Fig. 1 shows, the former correlation is quite strongly
influenced by the outlying value for Cambridge: if this is omitted the correlation
drops to - 0.40 (which is still larger than the correlation that would be obtained in
random sampling from a normal population with correlation 0: p ' C 0.02 for a two-
sided test).
In contrast with this, for computer studies degrees there is little relationship
between offer scores and percentage of women among entrants. This probably
reflects the fact that no university requires a grade A in mathematics for entry into
computer studies. It can be deduced from Table 1 that the percentages of women
among candidates gaining grades B and C in A-level mathematics are very similar, so a
variation in entry requirements between universities is not likely to influence the rate
of admission of women into computing degrees.
The relationship between actual Higher grade entry requirements, or typical offers,
and the entry of females into mathematics is more difficult to examine in Scotland for
various reasons. Firstly, a comparison between universities is complicated. Some
Scottish universities publish entry requirements by faculty only; some publish
minimum requirements in the SUCE guide but also publish 'competitive entry
standard' leaflets; some 'strongly recommend' Certificate of Sixth Year Studies
(CSYS) qualifications and soon. In terms of total point score for Highers (3 for grade
248 COHEN AND FRASER [Part 2,

0)
0
0
C
0,
0
-D

0
0

0
E
0
-r
0
F

0)
C
C
F
C

C
0
F
0

Point score or P-level oFFer

Fig. 1. Percentage of women among entrants to mathematical degrees versus point score of the A-level
offer (1986) (highest grade required: x , A; 0, 8; Es, C)

A, 2 for grade B and 1 for grade C) there seems little variation between the Scottish
universities, and no more than the variation between different mathematical degrees
within the same university. (For example, in 1987 Strathclyde asked for 8 points for
mathematics and 9 points for mathematical sciences; Heriot-Watt asked for 7 points
for mathematics and 9 points for actuarial mathematics and statistics; Edinburgh
asked for 8 points but strongly recommend CSYS qualifications; Aberdeen and St
Andrews asked for.8 points; Glasgow had science faculty entrance requirement of
ABBB (9 points) or BBBBB (10 points) including three science or mathematics
Highers for which the acceptance rate as opposed to the required rate was apparently
only BCC. An attempt to distinguish Scottish universities in terms of A or B require-
ments for mathematics was equally unrewarding. Some universities equate an A-grade
in mathematics and no other science pass to a B-grade in mathematics accompanied
by another science pass.) Secondly, some Scottish universities attract substantial
numbers of English entrants with A-level qualifications, and these need to be
considered separately. Thirdly, in many cases Scottish universities know applicants'
Highers results at the time of making offers.
Table 5 shows the variation in entry of females to mathematical degrees in Scottish
universities by domicile. (In retrospect, an analysis by SCE or GCE qualification
might have been more appropriate than by domicile; however, the tabulations
requested from the Universities Statistical Record were by domicile.) In the vast
19921 FEMALE PARTICIPATION IN MATHEMATICAL DEGREES 249
TABLE 5
Percentage of women among entrants into mathematical degrees, by domicile, Scottish universities,
1985-87t

University Scottish domicile Non-Scottish, UK domicile

% women Total entrants °h women Total entrants

Aberdeen 44 63 33 6
Dundee 50 26 40 20
Edinburgh 37 136 51 57
Glasgow 41 203 50 4
Heriot-Watt 31 155 33 45
St Andrews 33 58 47 64
Stirling 27 II 33 8
Strathclyde 47 157 33 3

Scotland 39 809 43 183

Degrees in mathematics, statistics and combined degrees with a mathematics or statistics specialization (source
Universities Statistical Record, special tabulations, 1985-87).

majority of cases the entrants living in Scotland have SCE qualifications while the
entrants living in England have A-level qualifications. Although the percentage of
women among entrants living in Scotland varies from 27% at Stirling to 50Wo at
Dundee, the larger universities do not vary so much; in fact the variation between the
eight universities is not particularly large in comparison with random variation
(p=0.07 on a x 2 -test). Only St Andrews, Edinburgh and Heriot-Watt attract
substantial numbers of entrants from outside Scotland, and in Edinburgh and St
Andrews the female proportion is very high compared with the English universities,
even taking account of their relatively low A-level requirements. It is also high in
relation to the percentage of women among Scottish entrants to these two universities.

4. OTHER FACTORS AFFECTING WOMEN'S ENTRY

We considered whether there might be other factors that explain the variation
between universities in the percentage of women entering mathematical degrees.
Potential applicants' perceptions of a university and their decision to make an
application could depend on factors such as the proportion of women students at the
university, the proportion of science or technology students and the availability and
cost of accommodation. Some applicants may be drawn towards new universities or
campus universities, others may prefer older redbrick universities. Having made a
choice of universities to apply to, the transition to entrance to a particular university is
affected not only by the grades achieved in A-levels or Highers but also by many
individual factors: interview experience, distance from home and so on. The question
is whether there is sufficient sex difference in factors such as these, among applicants
to mathematical degrees, to contribute to an explanation of the variation exhibited in
Table 2.
One hypothesis we had initially was that universities perceived to be particularly
scientifically or technologically orientated might be more off-putting to girls than to
boys. Of many possible measures of the degree of scientific and technological
250 COHEN AND FRASER [Part 2,
orientation of a university we chose to use the percentage of graduates in pure and
applied science (Universities Statistical Record, 1988). Keele University was excluded
because of the large number of multidisciplinary graduates. The alternative of using
the percentage of current undergraduates in pure and applied science was less satis-
factory because many degrees were classified as interdisciplinary, particularly in their
earlier years (Universities Statistical Record (1987), Table 16). We use the designa-
tions 'Woscience' for our measure and 'Wowomen' for the percentage of women
among entrants to mathematical degrees (excluding computer studies). Fig. 2 shows
that in England there is a positive association (correlation r=0.43). Thus, contrary
to our initial hypothesis there is a weak tendency for female participation in mathe-
matics to be higher in the more technologically orientated universities. As can be seen
in Fig. 2, this tendency is chiefly due to a few of the technological universities,
namely Bradford, Surrey, City University and the University of Manchester Insti-
tute of Science and Technology. For the English A universities alone the correlation
is close to 0 (r = —0.07), whereas for the B universities alone the positive correlation
is quite marked (r = 0.51). In Scotland this correlation is large (r =0.69) but is
heavily dependent on the inclusion of Stirling, which has a tiny number of entrants
to mathematics and is also low on Woscience, and on whether the definition of
Woscience is taken to include medical subjects. (r= 0.45 if Stirling is omitted; exclud-
ing medical subjects from %science gives r=0.38 with Stirling and r= 0.02 without

0,
a,
a,
L
0,
01)
- o

0
0

0
C
C
-c
0
E

C
a,
1)
0

20 28 36 4' 82 60 68 78 84
Z Graduces in Science

Fig. 2. Percentage of women among entrants to mathematical degrees versus percentage of graduates
in pure and applied science (highest grade required: x , A; 0, B; z, C)
19921 FEMALE PARTICIPATION IN MATHEMATICAL DECREES 251
Stirling.) Heriot-Watt University is an outlier here as it is highly science orientated,
but has a low %women value; however, Strathclyde has the highest value for both
%women and %science. The four older Scottish universities fall in between on
both measures.
Two sources of data are available. Universities Statistical Record (1987), Table 21,
gives graduates from each university in a sixfold classification: arts; social studies;
pure science; applied science; medicine; dentistry and veterinary science; multi-
disciplinary studies. Graduates whose employment destinations were unknown (9%)
are not classified. Universities Statistical Record (1988), Table 23, gives graduates
from each university in a 16-fold subject classification. This table is unfortunately not
published for the earlier years, which would be more appropriate for our purposes.
Analyses were carried out using both tables and the results were very similar. Alter-
native analyses were also carried out with various measures of university 'science
orientation'—in particular with medical and allied studies both included and
excluded. The correlation between Wowomen values and %science values was positive
in all cases, but somewhat lower when ¼ science was defined in terms of pure science
graduates only. Correlations were also generally lower when English and Scottish
universities were taken together. The correlations for Scottish universities alone were
sensitive to the inclusion or exclusion of medical subjects in the Woscience variable.
The figures reported are based on the data from Universities Statistical Record (1988)
with ¼ science including medical subjects and the denominator excluding multi-
disciplinary subjects.
We thought it interesting to carry out a multiple regression of Wowomen on both
total A-level point score of offers and Woscience. These two explanatory variables are
themselves virtually uncorrelated (r= 0.04), and together they explain nearly half the
variation between English universities (R 2 = 44 076).
The fitted equation is
Wowomen = 39.3 - 1.66(offer score) + 0. 1 8(%science).
((-ratios) (-3.63) (+3.00)
It is quite possible that the variable which we have chosen to represent the orientation
in science of a university is in reality a proxy for some other factor influencing
women's propensity to study mathematics. As indicated above there are a great many
possibilities, and we would not wish to read too much into our results; nevertheless,
they suggest that a scientific or technological image need be no drawback for a
university attempting to attract women into its mathematical degrees.
It is possible to make a crude examination of the association of accommodation
factors with the variation in the variable Wowomen. For each university Association
of Commonwealth Universities (1988) lists the percentage of first-year students in
various types of university-owned accommodation and the cost of such accommoda-
tion. Values are listed for seven different variables related to accommodation, and the
correlations of these with the Wowomen variable in Table 2 were examined. We also
examined the correlations of these variables with the overall percentage of women
among 1987 graduates in all subjects. The variable which gave a consistent and
reasonably strong negative correlation with both the percentage of women among
mathematics entrants and the percentage of women among all graduates was the cost
of accommodation in halls of residence with meals provided (r1 = — 0.35 for %women
in mathematics; r2 = — 0.42 for overall Wowomen). Not surprisingly, the London
 252 COHEN AND FRASER [Part 2,
colleges and City University tend to have high costs (except for the London School of
Economics which has almost the lowest cost), but City University and Royal
Holloway and Bedford College have a high value of Wowomen despite their high costs.
If London colleges and City University are omitted there is still a negative correlation
with cost (r1 = — 0.38) but it is heavily influenced by Cambridge which is an outlier on
both variates, and r1 = — 0.16 if Cambridge is omitted as well.
If this cost variable is added into the multiple regression above (London having
been excluded since it is impossible to assign it a unique cost) it has a small and non-
significant coefficient. Thus, there is no strong evidence that accommodation factors
affect the participation rate of females in mathematical degrees once the other factors
considered earlier are accounted for. None of the other accommodation variables
contributed significantly to the regression and their simple correlations with Wowomen
in mathematics varied quite markedly from their correlation with Wowomen in all
subjects.
Different universities have different prestige and if girls tend to lack self-
confidence in mathematics in comparison with boys (Joffe and Foxman, 1986;
Walkerdine and Girls and Mathematics Unit, 1989), then some of the universities with
high entry requirements that are generally thought to be strong in mathematics
(Cambridge, Oxford, Imperial College, Warwick and Manchester) may perhaps be
off-putting to girls whose qualifications or potential qualifications make them
possible candidates for these universities. Unfortunately we had no data on the rela-
tive numbers of male and female applicants to individual universities, but at a late
stage in this study we obtained information on total numbers of applicants to the A
and B groups of universities and the Scottish universities.
Table 6 summarizes some of the information in a form comparable with Table 4.
The differences between the percentages of women among all applicants, applicants

TABLE 6
Percentage women among applicants and accepted candidates for mathematical degrees, English and
Scottish universities 1985-87

University group I Values of °k women for the following groups I:

All applicants First-choice applicants Acceptances §

English A 28.5 27.7 24.8

(21214) (6191) (3806)
English B 32.9 35.8 34.2
(27101) (4133) (3911)
Scottish 40.7 42.8 42.3
(5631) (1170) (1050)
Highers 42.2 42.9 42.2
A.leveI 36.6 42.1 42.7
73.2 80.4 78.6

IFor English groups A and B see Table 4.

IBase numbers are given in parentheses. Source: UCCA special tabulations.
§Acceptances as tabulated by the UCCA differ slightly from entrants as tabulated by Universities Statistical Record;
seeTable 4. Entrants arecounted at December 31st. Acceptances here are onlythose with at least one A.level or Higher
grade.
§ § Candidates for Scottish universities with SCE Higher and GCE A-level qualifications respectively.
Wohighers: percentage of mathematical candidates for Scottish universities with Highers qualifications.
19921 FEMALE PARTICIPATION IN MATHEMATICAL DEGREES 253
whose first choice of university, in their UCCA application, was in the university
group shown in Table 6 and acceptances are not large. In particular there is no
evidence that women are put off applying to the universities with higher entry require-
ments, since the difference between the percentages of women among applicants to A
and B universities is less than the difference between the percentages of women among
acceptances (32.9 —28.5 = 4.4 and 35.8 —27.7 = 8.l compared with 34.2 —24.8 =
9.4). Indeed the higher percentage of women among applicants than acceptances for
the A universities might be taken as evidence of a greater ambition among girl
applicants than their subsequent performance warrants, although we would not wish
to draw such a conclusion from a small difference with very crude data. It is likely that
those selected girls who have sufficient ability, interest and support to prosper in
mathematics within the school system have rather different attitudes from the girls
who fail or lose interest in school mathematics, so that much of the research into sex
differences at school may not be directly applicable here. However, it is also quite
possible that these mathematically successful girls are still disadvantaged relative to
their male peers by social and educational processes similar to those so well
documented in the studies at earlier secondary levels.
The picture in Scotland from Table 6 is of almost perfect agreement for applicants
qualified with Highers, between the values of Wowomen among applicants and
acceptances. Comparing the applicants qualified with A-levels to Scottish universities
with those to the English B universities, there is some evidence that boys from England
are less likely than girls to put a Scottish university as their first choice. The official A-
level requirements for Scottish universities are generally low compared with those for
most English universities, but to say that this is a reason for the high percentage of
women among these first-choice applicants would be to support the hypothesis of lack
of self-confidence or underambition in females that was not supported earlier. Clearly
many other forces are at work in determining patterns of application and indeed the
sex effects are relatively small.
Paterson (1992) used data on individuals from a series of Scottish school-leaver
surveys to model the probability of applying to university, in any subject. The greatest
effects in the model were concerned with the levels of individuals' school examination
qualifications, but there was a statistically significant, though small, sex effect. When
adjusted for all other variables in the model, Paterson found that boy leavers had a
5% excess over girls in the fitted probability of applying to university. However, he
found no significant interaction between sex and school qualifications, either gained
or attempted. This supports our conclusion for mathematics specifically, that there is
no firm statistical evidence for the hypothesis of girls' lack of self-confidence, at least
among those qualified to apply to university.

5. PERFORMANCE IN DEGREES
As Scottish entry requirements are applied at a younger age than in England, when
discrimination according to specifically mathematical ability might be more difficult,
we might speculate whether this has any effect on relative performance in degrees in
the two countries.
The possible influence of the higher proportion of women entrants in Scotland on
the pattern of degree results could take several forms. One view would be that the
broad-based school education system and the younger entry point succeeds in enticing
254 COHEN AND FRASER [Part 2,
more women with the potential to obtain a good mathematics degree before they are
put off by specialization, or by social pressures which affect the image of mathematics
as an unfeminine subject. If so, presumably the better girls would displace the worse
boys in competing for entry, so that we might expect the sex effect in degree results (if
there is any) to be more in favour of women, or less against them, in Scotland than it is
in England. But if the same social and psychological factors that operate in school to
deter women from studying mathematics, or to reduce their ability to perform well,
also operate at university (maybe even more forcefully) then we would expect the same
or worse sex effect in performance in degrees in Scotland compared with in England.
Another view might be that entry requirements are weaker in Scotland, and that
degree performance would therefore be uniformly lower for both sexes than in
England.
A comparison between the two countries is complicated by the fact that three-year
Ordinary degrees are much more common in Scotland. However, the proportion of
women graduating with Ordinary degrees is almost the same as for men (Table 7) so
there is certainly no evidence that the relatively large number of women entrants in
Scotland is associated with a tendency for women to be less likely to gain admittance
to Honours degrees. The Ordinary degree differs between English and Scottish
universities. In England it is awarded to candidates whose final examination
performance is too poor to merit an Honours degree but not sufficiently bad for
outright failure. Table 7 shows that in England women are rather less likely to fail to
obtain an Honours degree than men. In Scotland, the Ordinary degree is awarded to

TABLE 7
Performance in mathematical degrees by sex in England and Scotland

No. of % Pass or No.of Honours 4b with the following

graduates Ordinary degree graduates classes.

University degrees in England and Scotland, 1985-87

(a) Mathematics, statistics and combined degrees
England, men 4658 7 4343 18 32 31 19
England, women 2227 4 2143 16 30 37 18
Scotland, men 473 22 368 27 28 30 14
Scotland, women 332 21 261 18 29 34 19
(b) Computer studies
England, men 2196 4 2099 12 37 36 15
England, women 292 5 279 tO 40 38 12
Scotland, men 542 18 443 12 42 36 tO
Scotland, women 129 17 107 II 43 36 9

CN.4A degrees in UK. 1986-88

(c) Mathematics, statistics and combined degrees
Men 1334 24 1016 10 32 40 14
Women 752 27 548 7 34 43 13
(d) Computing
Men 3000 23 2300 7 38 43 9
Women 683 27 499 8 35 46 7

tSources: Universities Statistical Record (special tabulation) and CNAA annual reports for 1986-87, 1987-88 and
1988-89.
19921 FEMALE PARTICIPATION IN MATHEMATICAL DEGREES 255
candidates who have taken a shorter and less specialized university course, either as a
result of failure (after the first two years of study) to gain admission into the Honours
course or, sometimes, as a positive choice.
The distribution of classes of Honours degrees is compared in Table 7, part (a), for
women and men doing mathematical degrees in each country. In England more
women obtain lower second-class degrees and slightly fewer achieve first- or upper
second-class degrees. The differences are not great, but there is a tendency for women
to perform less well than men, but to avoid the worst class. In a sense, men are more
extreme or variable in their performance. This has often been observed in studies of
sex differences in educational achievement at earlier ages (Willms and Kerr, 1987).
The same effect has been observed for degrees in many other subjects and many
possible explanations have been debated (Clarke, 1988; Rudd, 1988).
In Scotland women seem to perform worse than men with a lower proportion of
first- and a higher proportion of third-class degrees. However, the numbers are small
and the sex difference in degree results for Scotland is not statistically significant
(X2 = 7.53 on three degrees of freedom; p = 0.06). The Scottish women's distribution
is very much in line with the English distribution of degree classes; it is the Scottish
men's distribution that is, apparently, out of line. It is not clear why this should be,
and given the diversity of degrees included in this analysis (with possibly varying sex
distributions) we should not attach too much importance to this finding without
further work.
The differences in class of degree distributions for computer studies (Table 7, part
(b)) are not statistically significant either between sexes or between countries.
However, it is notable that many fewer first-class degrees are given in computer
studies compared with other mathematical degrees, and also slightly fewer third-class
degrees.
For comparison, in Table 7 we have included some data for establishments other
than universities. (Unfortunately a change in subject classification from 1985 to 1986
means that it is necessary to use the period 1986-88 for the polytechnics.) The number
of non-computing mathematical degrees awarded in polytechnics is relatively small
and there are many more Ordinary degrees and notably fewer first-class Honours
degrees than in the universities. The slight deficiency of first- and third-class degrees
awarded to females is still present. In computing degrees the polytechnics produce
about half the national total but with more non-Honours degrees than the
universities. The distribution of Honours degrees is notably more concentrated than
in the universities with fewer first- and third-class degrees, but there is no deficit of
first-class degrees awarded to females.
In Table 8 we compare the degree performance of graduates from the two classes of
English university distinguished in Table 4: those requiring a highest grade of A in
their typical offer for 1986 and those requiring a highest grade of B. The A universities
are dominated by Cambridge and Oxford, which have by far the largest mathematics
classes in the country. There is very little difference between men and women in the B
universities, but in the A universities women achieve notably fewer first- and more
lower second-class degrees. Perhaps the competitive atmosphere of mathematics
degrees in these high prestige universities is inimical to good performances by women.
It may also be that in their efforts to recruit more women from the small pool available
(Table 1) these universities offer places to women who are rather less well qualified at
A-level than their male peers.
256 COHEN AND FRASER [Part 2,
TABLE 8
Performance in mathematical degrees in England, by sex and entry requirements I

No. of % Pass or No. of Honours 3b with the following

graduates Ordinary degree graduates classes:

A universities
Men 2211 5 2102 22 38 25 15
Women 786 3 761 16 37 33 14

B universities
Men 2344 8 2152 15 27 36 22
Women 1380 4 1327 15 26 39 20

tFor groups A and B see Table 4

Clarke (1988) reported that at Warwick University in 1978-82, whereas 50% of

men entering mathematics degrees with an A-level point score of 15 subsequently
obtained first-class degrees, the corresponding figure for women was 35%. Clarke
discusses explanations in terms of social and institutional pressures and examiner
bias, but Rudd (1984, 1988) appears to maintain that there are simply more males in
the highest ability category. The special character of mathematics degree examina-
tions can easily prompt speculation about reasons for the underachievement of
women at first-class level.
For instance, there could be a tendency for the brightest women to obtain their
marks from more conscientious and time-consuming answers, whereas the brightest
men obtain very high marks with slightly sketchier answers which allow them more
time to amass marks. It would be very difficult to verify such speculation although it is
consistent with the finding of gender-specific feed-back to school pupils for neatness
in their mathematical work (Walkerdine and Girls and Mathematics Unit, 1989). A
comparison of different universities' examinations could shed some light. Research
by Fraser (1992) examines gender differences in mathematics undergraduates'
motivations and attitudes towards their work and examinations.
A comparison between the A and B universities is slightly complicated by the fact
that the undivided second-class degrees at Cambridge have here been arbitrarily
classified as upper second class. This accounts for the high percentage of upper
second-class degrees and the small percentage of lower second-class degrees in the A
category. (Also Cambridge graduates are classified by Universities Statistical Record
according to their best class in either part of the tripos.) However, even if all second-
class degrees are taken together, a comparison of the A and B groups reveals that the
A universities award more first- and fewer third-class degrees than the B group. If
there is some correlation between performances at A-level and at degree level, and if
the standard of difficulty of degree examinations at A and B universities is
comparable, this would perhaps be expected.

6. CONCLUSION
Lower female participation in A-level mathematics is directly related to the under-
representation of women in university mathematics degrees. The degree of sex
19921 FEMALE PARTICIPATION IN MATHEMATICAL DEGREES 257
imbalance is worse in the universities that require a grade A at A-level. It is
considerably less in Scotland where female participation in Higher grade mathematics
is greater than in A-level and entry requirements are generally less severe. Female
participation in computing degrees is generally much lower than in other
mathematical degrees, but entry requirements are not so high and there is less
variation between universities.
Variation between individual universities in the proportion of mathematics
entrants who are women is quite striking but there is no evidence, from the limited
data studied, that a perception of a university as scientifically or technologically
oriented is a factor which discourages women from entry. Nor is there any evidence
that prospectively suitable women are discouraged from applying to universities by
high entry requirements; the universities with such requirements would appear to
receive smaller proportions of women principally because a lower proportion of
women than men achieve the top A-level grade in mathematics. Among the entrants to
universities requiring a grade A at A-level, women have a lower chance of achieving a
first-class Honours degree in mathematics, a phenomenon which has been noted in
other subjects. it is debatable whether this represents a true difference in ability at the
highest mathematical level or is a result of complex social and educational factors
operating in university mathematics courses.

ACKNOWLEDGEMENTS
The data on entrants were kindly supplied by Miss C. Holme of the Universities
Statistical Record, and data on applicants were supplied by Dr L. E. Williams of the
UCCA. We would like to thank Dr Sheila Cormack for stimulating this work and Dr
Lindsay Paterson and Mr David Salmond for useful discussions.

REFERENCES
Association of Commonwealth Universities (1988) University Entrance 1988, the Official Guide.
London: Association of Commonwealth Universities.
Burton, L. (ed.) (1986) Girls into Maths Can Go. London: Holt, Rinehart and Winston.
Chipman, S. F., Brush, L. R. and Wilson, D. M. (eds) (1985) Women and Mathematics: Balancing the
Equation. Hillsdale: Erlbaum.
Clarke, S. (1988) Another look at the degree results of men and women. Stud. Higher Educ., 13,
315-331.
Council for National Academic Awards (1989) Annual Report 1988-89. London: Council for National
Academic Awards.
Department of Education and Science (1985) Statistics of Education—School Leavers CSE and GCE
1985. London: Department of Education and Science.
(1987) Statistics of Education—School Leavers CSE and GCE 1987. London: Department of
Education and Science.
(1989a) Education Observed 14: Girls Learning Mathematics. London: Department of Education
and Science.
(1989b) Statistics of Education—Schools 1988, Table A20. London: Department of Education
and Science.
Fraser, E. J. P. (1992) Gender effects in university mathematics education. MPhil Dissertation.
Department of Mathematics, University of Edinburgh.
Girls and Mathematics Association (1984) Bibliography. London: University of London Institute of
Education.
Joffe, L. and Foxman, D. (1986) Attitudes and sex differences—some APU findings. In Girls into Maths
Can Go (ed. L. Burton). London: Holt, Rinehart and Winston.
258 COHEN AND FRASER [Part 2,
Office of Population Censuses and Surveys (1987) Young People's Intentions to Enter Higher
Education. London: Her Majesty's Stationery Office.
Paterson, L. (1992) The influence of opportunity on aspirations among prospective university entrants
from Scottish schools, 1970-88. J. R. Statist. Soc. A, 155, 37-60.
Royal Society and Institute of Mathematics and its Applications (1986) Girls andMathematics. London:
Royal Society.
Rudd, E. (1984) A comparison between the results achieved by women and men studying for first degrees
in British universities. Stud. Higher Educ., 9, 47-57.
(1988) Reply to Clarke. Stud. Higher Educ., 13, 333-336.
Scottish Education Department (1990) Statist. Bull. 11/E411990. Scottish Education Department,
Edinburgh.
Shuard, H. B. (1981) Differences in mathematical performance between girls and boys. In Mathematics
Counts: Report of the Committee of Inquiry into the Teaching of Mathematics in Schools, Appendix
2. London: Her Majesty's Stationery Office.
Universities Statistical Record (1987) University Statistics, 1996-87, vol. 2, First Destinations of
University Graduates. Cheltenham: Universities Statistical Record. -
(1988) University Statistics, 1987-88, vol. 1, Students and Staff. Cheltenham: Universities
Statistical Record. -
Walkerdine, V. and Girls and Mathematics Unit (1989) Counting Girls Out. London: Virago.
WilIms, J. D. and Kerr, P. D. (1987) Changes in sex differences in Scottish examination results since
1976. J. Ear/yAdol., 7, 85-105.