The Relationship between the Number Sense and Problem Solving

Abilities of Year 7 Students

Jemmy Louange Jack Bana
Port Hedland Senior High School Edith Cowan University
<[email protected]> <[email protected]>

This paper reports on a component of a large yearlong study in three Year 7 classes in three
different schools. The aim of this research component was to determine the relationship
between students’ number sense and their problem-solving ability by means of paper-and-
pencil tests, classroom observations, and interviews of students and teachers. The results
revealed a strong correlation between these two aspects of school mathematics, with
important implications for classroom teachers.

This paper reports on one aspect of a much larger study conducted in three schools in
the Perth Metropolitan over a one-year period (Louange, 2005). The three schools were an
independent boys’ school, an independent girls’ school and a coeducational public school.
One Year 7 primary school teacher, identified by tertiary mathematics educators as being
an effective mathematics teacher, was selected in each school for the study. The main
study investigated the relationships between, teaching style, learning style, number sense,
and problem solving ability in the three Year 7 classes through testing, interviews of
students and teachers, and extensive periods of classroom observation over the year’s
mathematics lessons. The purpose of this paper is to determine the relationship between
number sense and problem solving as a facet of the larger study just described.

Background
Both number sense and problem solving are promoted as two of the major areas of
emphasis in mathematics education, as evidenced in major curriculum documents
(Australian Education Council, 1991; National Council of Teachers of Mathematics, 2000;
Curriculum Council of Western Australia, 2005). Thus, there is an urgent need to answer
questions such as, “How do they relate to each other in terms of how they are taught, learnt
and utilised in solving mathematical problems?” While there have been many studies
relating to the two aspects independently, there is a lack of research which focuses on the
relationship between the two.
Unfortunately, one possible reason for lack of a combined number sense and problem
solving research could be due to the difficulty of divorcing one from the other. Both the
terms ‘number sense’ and ‘problem solving’ have suffered many diverse definitions. The
controversy stems mainly from the fact that number sense is akin to common sense
(McIntosh, Reys, Reys, Bana, & Farrell, 1997), which is a necessary tool for solving any
problem. Compared to problem solving, number sense seems to have proved to be the most
difficult to define, as McIntosh et al. (1997) state:
Like common sense, number sense is a valued but difficult notion to characterise and has stimulated
much discussion among mathematics educators (including classroom teachers, curriculum writers,
and researchers) and cognitive psychologists. (p. 3)
Thus, while there seems to be general agreement by mathematics educators on a
definition of a problem, it is not possible to define number sense in such a straightforward
manner. A problem is generally regarded as a situation where both the solution and method

L. Sparrow, B. Kissane, & C. Hurst (Eds.), Shaping the future of mathematics education: Proceedings of the
33rd annual conference of the Mathematics Education Research Group of Australasia. Fremantle: MERGA.

360
of solution are not obvious. However, number sense is best described by a comprehensive
list of characteristics. Without listing them here, we would venture that number sense
could be considered as ‘the ability to make sound estimates’, since this ability necessarily
incorporates the list of number sense attributes that are usually identified.
In the publication edited by McIntosh and Sparrow (2004) mathematics educators from
the UK, Europe, the USA and Australia who are recognised as leaders in the issue of
number sense, provide considerable insight for researchers, curriculum developers and
mathematics teachers at all school levels. There is an abundance of other material by
mathematics educators and researchers on the teaching of number sense, with publications
by Anghileri (2000) and Askew (2002) being two examples. Considerable research has
been undertaken on problem solving, and the teaching of mathematics through problem
solving continues to be a focus of mathematics educators (Schoen & Charles, 2003; Nisbet
& Putt, 2000). However, investigations into the links to number sense are lacking. Despite
their evident connection, problem solving and number sense are not necessarily
synonymous because while number sense is inherent in problem solving, many problems
are solved without recourse to number sense (Hiebert et al., 1997). Most definitions of
number sense incorporate a sense of problem solving (Denvir & Bibby, 2002), which
serves to show that it is virtually impossible to separate the two. Although in these
definitions the intention weighs more towards number sense inherent problems, it could
also be that number sense ability is intricately linked to mathematics problems, which are
devoid of number sense (Anghileri, 2000). Yet there is a lack of research to elucidate the
relationship between problem-solving ability and number sense.
In the context of this present research any problem which necessarily requires
knowledge and skill in number, to arrive at an acceptable resolution, will involve number
sense; while any question for which the solver has no immediate and apparent way of
solving will constitute a problem (Thiessen & Trafton, 1999; Reys & Yang, 1998). This
study aims to explore what relationships exist between Year 7 students’ number sense and
their problem-solving ability.

Methodology
As indicated above, the subjects were three Year 7 primary classes in three different
schools, each of which had a full-time teacher identified by a number of tertiary
mathematics educators to be an effective mathematics teacher. The study involved both
quantitative and qualitative approaches. Pre-tests in both number sense and problem
solving were administered to all students at the beginning of the school year, and the same
tests were given again at the end of the year. On both occasions the tests were administered
over two days, with one test each day in the normal mathematics time-slot. Each item was
read to the class to help overcome any possible reading difficulties. During the year, 30
mathematics lessons were observed in each class, which was one lesson per week for each
teacher over most of the data collection period. Both teachers and students were
interviewed.
The number sense instrument consisted of 45 of the items used by McIntosh et al.
(1997) in their international study of number sense in three countries. A sample item is
included in the Appendix. The problem-solving instrument consisted of items selected
from problem-solving competitions conducted by the Mathematical Association of
Western Australia (2000, 2001, 2002). The problem-solving instrument was refined from a
pilot study and consisted of eight items – four of which involved number sense and four
that were devoid of number sense. A sample of each type is in the Appendix. Each problem

361
was presented on a single page and students were required to document all their processes
and the strategies used. It was considered most important that the process as well as the
product be taken into account for each problem. A scoring scale was adapted from Charles,
Lester, and O’Daffer (1987) that allocated a score of 0, 1, or 2 for each of the categories
Understanding the Problem, Planning a Solution, and Getting an Answer. These three
scores were then added to give an overall score for each of the eight problems.
A total of 64 students completed all the assessments and 45 of these were formally
interviewed. The 45 were selected in such a manner as to be representative both of the
three classes and also their performances on the two initial paper-and-pencil tests. This
interview involved each student working aloud to complete two problems of each type,
with follow-up questions by the interviewer as appropriate, and using stimulated recall.
Half the 64 students were also interviewed informally during the classroom observation
sessions over the one-year period. Semi-structured interviews were conducted with each of
the three teachers, and there were informal interactions on the occasions of the
observations of the mathematics lessons. Observations of significance to the study were
recorded during the observed mathematics lessons by the researcher. All data gathering
was undertaken by the one researcher.

Results
The results for both the pre- and post-tests of number sense (NS) were combined to
give one score. The same was done for the pre- and post-tests of problem-solving (PS)
performance to give one basis for comparing the two aspects. The combined scores for
number sense were categorised as High (HNS), Medium (MNS), or Low (LNS). The
combined scores for problem solving were categorised in the same way. Using the three
levels for each skill, the grid of nine cells in Table 1 gives a good insight into the
relationship between number sense and problem solving.
Table 1
Number and Percentage of Students within each Score Category for NS and PS
HPS MPS LPS Total

HNS 12 6 1 19
(18.8) (9.4) (1.6) (29.7)

MNS 5 15 6 26
(7.8) (23.4) (9.4) (40.6)

LNS 1 6 12 19
(1.6) (9.4) (18.8) (29.7)

Total 18 27 19 64
(28.1) (42.2) (29.7) (100)

Note: N = 64, with percentages of students shown in parentheses.

The diagonal of matching categories demonstrates a very strong relationship between

number sense and problem-solving ability, with 12 students rated high in both, 15 rated
medium in both, and 12 rated low in both aspects. There are 11 students rated high on one

362
aspect and medium on the other, and there 12 rated medium on one and low on the other,
so that the relationship for both these sets could be considered as fairly strong. This leaves
only two students who scored high on one and low on the other. These results clearly
demonstrate a strong relationship between the two aspects overall.
The problem solving instrument consisted of four number-sense-inherent problems
(NSIP) and four devoid-of-number-sense problems (DNSP). The 64 students were asked
whether they preferred NSIP or DNSP types. Figure 1 shows a scatter plot of the
relationship between problem-solving performance and number sense, and also gives each
student’s preference for either NSIP or DNSP types. A two-tailed Pearson Correlation was
applied to the pre- and post- PS and NS combined scores, resulting in quite a strong
correlation of 0.77 at the 0.01 level. The coefficient of determination indicates almost 60
percent shared variance, which implies that number sense helps to explain nearly 60
percent of the students’ scores on the problem solving test. Although the converse could
also be true, triangulation of data obtained from the various forms of data collected,
especially those from the interview involving stimulated recall with the solving of four
problems, show greater support for a theoretical framework in which problem solving
ability level depends more on number sense than vice versa. For instance, for the 45
students interviewed, there was a correlation between their PS scores at the interview and
their PS test performance scores (R = 0.31, p < 0.04). There was also a significant
correlation between the 45 students’ interview PS scores and their NS performance scores
(R = 0.55, p < 0.005).

110

100
Problem Solving Total Percentage Score

90

80

70

60
Preference for NSIP,
50
Prefers NSIP

40 Prefers DNSP

30 Likes Both
20 30 40 50 60 70 80 90 100

Number Sense Total percentage Score

Figure 1. Relationship between number sense and problem solving, with student preferences.

363
 Irrespective of the statistical calculations, Figure 1 clearly shows a strong correlation
between the number sense and problem-solving ability of the 64 Year 7 students. Since the
students preferences for solving either NSIP or DNSP or both seemed to be related to their
number sense and problem solving performance, the scatter plot presented in Figure 1 also
shows the distribution of the students’ scores according to the type of problems they
preferred to solve. Although there was no marked difference between the percentage of
students preferring NSIP (45%) and those preferring DNSP (38%) it was found that
student preference for solving NSIP was more closely related to number sense performance
(R = 0.69). Even more striking was the higher correlation of NSIP preference and problem
solving scores pertaining to NSI problems (R = 0.56) as opposed to a very low correlation
between NSIP preference and performance scores for solving DNS problems (R = 0.29).
These results are graphically supported in the scatter plot presented in Figure 1.
Student interviews revealed that the students considered number sense to be a
significant factor in problem solving. Although few used the actual term number sense, it
was clear from their descriptions that this was the inference. Table 2 lists the three major
factors considered by students to affect problem-solving performance. Lack of number
sense often created difficulties, as explained by one student:
I don’t think that I did not understand what I read. I understand all these words, but there
are calculations to be made, but I don’t know which calculation to do. I don’t always
understand what the numbers, what to do with the numbers.
Table 2
Main Factors Identified by Students as being Responsible for Poor Problem-solving
Performance (N = 64)
Factor Summary of students’ Count Percentage
most common answer
Lack of number Lack of understanding of number facts 45 70
sense and how to apply them

Lack of language Not understanding the language; not 36 56

proficiency being able to read properly

Mathematics Afraid to solve any mathematics 31 48

anxiety problems; lack of confidence

One issue arising from the discussion with the students was the need to work
mathematically. This was supported by the three teachers and was summarised quite
succinctly through Bob’s statement that “without number sense students would find it hard
to work mathematically”; and according to Amanda, “it is extremely difficult to work
mathematically if one has poor number sense … because this will make it even more
difficult to solve most problems”. Yet, as Chantal pointed out, the challenge to overcome
this obstacle “is a very big one, given that mathematics is not only about number sense, but
also about other concepts and mathematics sense”. This notion of “making sense of the
mathematics” was explained by Chantal as being “more prominent in making sense of
number, as it permeates all other strands of the mathematics curriculum”. Such a notion
was quite widespread in both practice ― through the learning experiences observed ― and
theory, as expressed by Amanda: “since most problems require number sense, students

364
with such ability have a great advantage over those with poor or no number sense, when it
comes to successfully solving a problem”. All three teachers and the majority (70 percent)
of students believed that lack of number sense is a probable major cause of poor
performance in solving mathematics problems. Clearly, the link between number sense and
problem solving is very significant. Bob reiterated this point when he stated that:
Number Sense is very much like problem solving in the sense that you have to read the problem, try
to understand it, plan a way to solve it and come up with a reasonably accurate answer. All these
performance components must be assessed in both number sense and problem solving if I am to
encourage the students to love working with numbers, and to make sense of what they have learnt.
All three teachers favoured an assessment method that took process as well as product
(the solution) into account, thus linking problem solving with assessment. Hence, it was
not surprising to learn that Chantal’s comment that “… number sense should be, or maybe
I should say must be assessed through problem solving, since it [number sense] involves
mainly how students make sense of the number components of a problem” was a view also
shared by the other two teachers. Bob’s view, that “assessing for number sense through a
problem-based method helps me not only to gauge the student’s content knowledge, but
also his thinking process and solution” was a prevalent one among all three teachers.

Conclusion
This study showed that there is quite a strong correlation between the number sense
and problem solving proficiency of Year 7 students. The evidence points towards a
relationship in which problem solving performance depends upon number sense
proficiency more than the latter depending on the former. The relationship is borne out, not
only in the results of the paper-and-pencil tests, but also from the views of both the
teachers and students. Teaching through a problem-based approach should be a priority for
every teacher of mathematics who endeavours to enhance his or her students’ number
sense problem solving proficiency. As pointed out by the NCTM Standards (2000), both
number sense and problem solving are crucial to the learning of mathematics. Number
sense and problem solving are linked through assessment, which incorporates
consideration of both the thinking process and the final solution by a student. The specific
relationship between number sense and the solving of problems, which are devoid of
number sense does need some further investigation. Finally, it is clear that teachers need to
ensure that problem solving is the focus of their mathematics programs, so that students are
always working mathematically.

References
Anghileri, J. (2000). Teaching number sense. New York; London: Continuum.
Askew, M. (2002). The changing primary mathematics classroom: the challenge of the National Numeracy
Strategy. In L. Haggarty (Ed.), Aspects of teaching secondary mathematics: Perspectives on practice,
(pp. 3-17). London: Routledge Falmer.
Australian Education Council. (1991). A national statement on mathematics for Australian schools.
Melbourne: Curriculum Corporation.
Charles, R. I., Lester, F. K., & O'Daffer, P. (1987). How to evaluate progress in problem solving. Reston,
VA: NCTM.
Curriculum Council of Western Australia. (2005). Curriculum framework: Progress maps. Perth: Author.
Denvir, H., & Bibby, T. (2002) Diagnostic interviews in number sense: One-to-one assessments mapping
children’s understanding of number. London: BEAM Education.

365
Louange, J. E. G. (2005). An examination of the relationships between teaching and learning styles, and the
number sense and problem solving ability of Year 7 students. Unpublished PhD thesis: Edith Cowan
University, Perth.
Mathematical Association of Western Australia. (2000). Have Sum Fun Competitions. Unpublished
document. Perth: Author.
Mathematical Association of Western Australia. (2001). Have Sum Fun Competitions. Unpublished
document. Perth: Author.
Mathematical Association of Western Australia. (2002). Have Sum Fun Competitions. Unpublished
document. Perth: Author.
McIntosh, A., & Sparrow L. (Eds.) (2004). Beyond written computation. Perth: MASTEC, Edith Cowan
University.
McIntosh, A., Reys, B., Reys, R., Bana, J., & Farrell, B. (1997). Number sense in school mathematics:
Student performance in four countries. Perth: MASTEC, Edith Cowan University.
National Council of Teachers of Mathematics. (2000). NCTM standards 2000: Principles and standards for
school mathematics. Reston, VA: NCTM.
Nisbet, S., & Putt, I. (2000). Research in problem solving in mathematics. In K. Owens & J. Mousley (Eds.).
Research in mathematics education in Australasia 1996 to 1999, (pp. 97-122). Sydney: Mathematics
Education Research Group of Australasia.
Reys, R. E., & Yang, D. C. (1998). Relationship between computational performance and number sense
among sixth- and eighth-grade students in Taiwan. Journal for Research in Mathematics Education,
29, 225-237.
Schoen, H. L., & Charles, R. I. (2003). Teaching mathematics through problem solving: Grades 6-12.
Reston, VA: NCTM.
Thiessen, D., & Trafton, P. (1999). Learning through problems: Number sense and computational strategies,
a resource for primary teachers. Portsmouth: Heinemann.

Appendix
Number Sense Item Sample

2 3
A None. Why?
How many different fractions are there between /5 and /5? _______________________

Circle your answer and then fill in the blanks. B One. What is it?
_______________________

C A few. Give two:

___________ and________

D Lots. Give two:

___________ and ________

Problem Solving Item Sample Including Number Sense

Peter, Paul and Pat divide $120 so that Peter gets three times as much as Paul, who gets half as much as Pat. How much
does Peter get?

Problem Solving Item Sample Devoid of Number Sense

Alan, Brett, Carol and Dianne went to basketball, cricket, hockey and athletics. Carol didn't go to basketball; Brett
couldn't go to cricket; the girl who went to hockey would like to have gone to cricket; and the person who went to
basketball was upset she couldn't go to athletics. Who went where?

366