UNITS OF MEASUREMENTS

MEASUREMENT SCALES
When the entire possible variable in education is considered, it soon becomes
apparent that all measurement is not the same. In the science achievement
example, the variable teaching method is simply a categorical variable, we know
that the methods are different and can be given different names. On the other
hand, if the dependent variable of the score on a science achievement test is
considered, the measurement is not simply categorical; there is some kind of
ordering and quantifying of the scores.

Essentially, there is a hierarchy of measurement scale-with

four general categories, based on the amount of information contained in a score.
The lowest type of scale is the nominal scale, which simply categories without
order.sex of the individuals are a variable that MEASURED ON A NOMINAL
SCALE.

Consider a variable such as attitude toward school. If one

individual indicates a highly favorable attitude toward school and another
individual indicates a neutral attitude, we not only know that they are different,
but we can also order the individuals on the degree of favorable in their attitude
toward school. Thus, besides having a difference, we also have order.A variable so
MEASURED IS ON AN ORDINAL SCALE.
FRED N. KERLINGER

Suppose that we measure a variable such as is or performance on an achievement

rest in reading or mathematics. If three individuals have scores of 105,110, and
115, respectively, the difference of five points between the first two individuals is
considered equivalent to the difference of five points between the second two
individuals. Thus gives not only difference and order but also a unit of equal
different established in the measurement. This level of MEASUREMENT IS CALLED
EQUAL-UNIT OR INTERVAL SCALE.
 FRED N. KERLINGER

“A variable such as the weight of different quantities of applies is an example of a

fourth level of measurement. If we have two bags of apple, a 25-pound bag and
or 50 pound bag, we say that the 50 pound bag weights twice as much as the 25
pound bag. We can say ‘’twice’ become having no apples is zero point in quantity
of apples. Thus, we have not only different, order and an equal unit, but also a
comparison in term of the ratio of one observation to another.hence, this level of
MEASUREMENT IS CALLED RATIO SCALE. If an individual’s scored no points on an
in test one would not say that the individuals is completely lacking in intelligence.
To establish a ratio scale must be have a true zero point.”
FRED N. KERLINGE

SCALES

A scale is a numbers whose properties determine imperial properties of all objects

to which the numbers are assign.

OR

Scale is a device by which we measure things. THERE ARE DIFFERENT SCALES FOR
MEASURING DIFFERENCE PHENOMENON. FOR EXAMPLE HEAT IS MEASURED BY
THERMOMETER. Moreover, there are things that are scalable i.e. capable of being
measure and those are not scalable. Thus a scale is measurement.

We can also say that ANY SYSTEM OF OPREATIONS THAT WILL ASSIGN
NUMBERS IS KNOWA AS MEASURMENT. It helps to have some internally
consistent plan for the development of a new measure. The plan is known as a
scaling model and the measure which results from exercising THE PLAN IS
KNOWN AS A SCALE.
IN NOTES

MEASUREMENT POSTULATES
A postulate is an assumption that is an essential pre-requisite to carrying out
some operation or line of thinking. In this case, it is an assumption about the
relation between the objects being measured. Although postulates are usually
assumed to be true in measurement it is necessary to test the postulates
whenever possible. More than three postulates are really necessary in order to
make it possible to equal objects to rank order them and to add them.
FRED N. KERLINGER P434

THE THREE POSTULATES CAN BE WRITTEN-

1. EITHER (a=b) OR NOT, BUT NOT BOTH

2. IF [(a=b) AND (b=c)], THEN (a=c)

3. IF [(a>b) AND (b>c)], THEN (a>c)

1) FIRST POSTULATES SAYS:

a is either equal to b or not equal to b, but not both. “This postulate is necessary
for classification; we must be able to assert either that one object is the same in
characteristics as another or that it is not the same. In measurement ‘the same’
does not necessarily mean complete identify. If we wish to assign boys to social
class categories, we might use the criterion of father’s occupation and/or
residence. The criterion must be sufficiently unambiguous to make classification
possible, that is, to satisfy the condition the postulates states.

2) THE SECOND POSTULATES SAYS:

If an equal b, and b equal, then an equal c’’. If one member of a universe is the
same as another member is the same as the third member, then the first member
is the same as the third member. This postulate enables a researcher to establish
the equality of set members on characteristics by comparing objects. More
important, if the postulate is satisfied objects. More important, if the postulate is
satisfied, object no ordinarily amenable to observation may be assigned to subset
of a universe. For example, suppose we wish to assign individuals to two
categories. “prejudiced” and “unprejudiced” or we may be able to assign him to
the category “prejudiced “on the basis of his response to another measuring
instrument which is highly correlated with a prejudice measuring instrument.
FRED N.KERLINGER P434

3) THE THIRD POSTULATES SAYS:

More immediate and practical importance for our purposes. It says, “If a > b and
b > c, then a > c” this is the tenability postulates. Other symbols or words can be
substituted for “>” “<” is at greater distance than, “is stronger than”, “proceeds”,
“dominates”, and so on. Most measurement is psychology and education
depends on this postulates. It must be possible to assert ordinal –or rank order
statement like, “a has more of a property than b, b has more of the property than
c, therefore a has more of the property than c “.
FRED N.KERLINGER P435

THE 1937 SCALE

DESCRIPTION
The scales differ from that of 1916 in many details, but it does not differ in its
essential and basic conceptions. As the authors themselves state. The revision
utilizes the assumption, methods and principles of the age scale as conceived by
‘binet’.they do,however,regard it as a better standardized and more useful scale
than its predcessors. The principles differences and modification follow.

The 1937 scale has two equivalent forms (land

m) each of which contains 129 test items as compared with the 90items in the
first Stanford binet. Items that proved unsatisfactory in the original were
eliminated and new ones were added.
FRANK S. FREEMAN
The scale was standardized on a carefully chosen and extensive group of subjects.
The base of the standardization population was broadband and its component
members are regarded as more representative of the population but only
American born white subjects were used in the standardization of the scale, the
total number being approximately 3000.The subjects were chosen from eleven
states in several widely separated areas of the country, and an effort was made to
have them from homes which, occupationally and socially would be
representative of the population at large.
FRANK S.FREEMAN P211

NEED FOR SCALING

In social phenomena most of the facts are qualitative in nature, and are not
subjected to direct measurement. The question now arrives as to what is after all
the need for providing quantities measurement. For this purpose the reader is
referred back to the first chapter of this text. There we have discussed at length
how one of the chief characteristics of a mature science is its capacity of being
quantatively and objectively measured .qualitative measurement are mostly
subjective in nature and may differ from person to person. they hardly give any
precise idea about the state of affairs, to take an example, we have heard people
talk about growing indiscipline, lack of morality ,rising prices and cost of living
etc.but all this does not convey a complete picture of the problem so long as it is
not expressed in quantities terms.

In social sciences too, efforts for quantities measurement

are being tried in the form of various kinds Sociometric scales .one false
assumption must be cleared at this stage .it is generally considered that social
data is quantities in nature and therefore any attempt to quantify it may result in
arbitrary, unstandardised, false and misleading measurement. This problem has
been dealt at length in the first chapter, but it will be useful to refer to it in brief
at this point. Social phenomena appears to be complex, intangible and therefore,
incapable of numerical expression because of our lack of complete knowledge
about it. As Science improves, more and more research is carried on, Factors
involved, shall be able to grade, classify and ultimately measure the
independently.
(DR. R. N. TRIVEDI P236)

PROBLEM OF SCALING

Once it is accepted on principle that quantities measurement is definitely

more superior, precise and adequate, the next problem is to prepare
reliable and valid scales for measuring them. The following problem may
have to be talked in this connection.
(DR.D.P.SHUKLA)

1) DEFINATION OF CONTINUUM

Before a scale can be prepared for measuring Social Phenomena, we have to see
that Phenomena in question is scalable. The scalability of Phenomena is
dependent upon continuum. The various factors to be measured rather than
stray facts hare & there. To take an example, suppose a scale measuring our
Social distance to various minorities to be prepared. Now our Social Measures or
farness can be expressed in a number of ways. All the expressions that are taken
as indicative of social distance should be such as can show gradually increasing or
reducing distance. The continuum thus depends upon the nature of Phenomena
to be measured and also the nature of indicating factors to be included.

(DR. R. N. TRIVEDI P236)

2) RELIABILITY

Besides forming a continuum the scale has to be reliable. A scale may be said to
be reliable, when it gives the same measurement under similar conditions. Thus
in the above example of social distance, the scale would be reliable if two persons
or groups who are similar in various relative factors show the same social distance
toward a particular minority. A scale that will show different results when applied
to the same persons at different times or different persons of the same category
cannot be said to be reliable.

Following three methods are generally used to test the reliability of a scale.

(DR. R. N. TRIVEDI P237)

RATEST METHOD

According to this technique the same scale is applied twice to the same
population and the results so obtained are compared. If the two results are
similar or the differences can be explained suitably the scale would be taken as
reliable.

MULTIPLE FORM

According to this method two forms of scale are constructed and alternately
applied to the same population.

SPLIT HALF METHOD

According to this method the scale is divided into two equal parts e.g. taking even
numbers on the other side.

3) VALIDITY
A scale can be said to be valid when it correctly measures, when it is expected to
measure, validity should not be confused with the reliability of the scale.
Reliability is achieved, when the scale is free from erratic measurements. It is
valid only when the measurement is real and correct.

GOODE and HATI have given following four criteria of validity of a Scale:

(DR. D.P.SHUKLA P 237)

LOGICAL VALIDATION

According to this principle a scale would be said to be a valid if it stands to our

commonsense or reason.

JURY OPINION

This is more reliable and frequently used method. In this case we do not rely upon
the commonsense of one person but on the judgment of a number of persons.

(DR.R.N.TRIVEDI P238)

KNOWS GROUPS

According to this method the scale is administrated among the person who are
known to hold a particular opinion or belong to a particular category, and the
result are then compared with the known facts.

(DR.D.P.SHUKLA)

INDEPENDENT CRITERIA

According to this method the scale is tested on the basis of various variables
involved. If all or most of the tests show the same result the scale would be said
to be valid.

4)PROBLEM OF WEIGHTS
The validity of a scale is increased if proper weights are given to various points
included in the scale .In the absence of weights the scale becomes equally
weighted and equal importance is attached to each item. If all the attributes
involved are no of equal importance, they must be provided with proper weights.

5) DIFFICULITIES OF SCALING

various difficulties are experienced in the formation of scale .In fact it is because
of these difficulties that scaling and quantities measurement is not still much in
use in social research although the tendency in constantly growing .The following
difficulties are mainly experienced in preparing a scale.

1. The social phenomena are very complex and cannot easily be quantitatively
defined.

2. There are no universally accepted measures of values.

3. The greater difficulty is faced because of the fact that social phenomena cannot
be put to laboratory type test. The social events have to be watched as and when
they take place.
DR R.N.TRIVEDI P239
 KINDS OF SCALE
Various scales commonly used in Social Research have been described in the
succeeding paragraphs.

POINT SCALE

In this scale a number of works or situations about which the opinion of the
respondent is to be elicited are selected. These words may be of any tupe viz,
dancing, birth control, prayer etc. The respondent is asked to cross out every
word that is more annoying than pleasing to him. One point to given to each
agreement or disagreement whichever is to be selected. Thus if agreement is to
be considered, each word which has not been crossed will be given one point.

Another type of point scale is also sometimes used. In this kind of scale two sets
of words are given, indicating both favorable and unfavorable opinions. The
attitude of a person is judged by the number of unfavorable items crossed and
favorable items left unscored.

The scales formulated for measuring social distance are of two types.

1). The Bogardus social distance scale

2). Sociometric measurements

(DR.D.P.SHUKLA P240)

THE BOGARDUS SOCIAL DISTANCE SCALE

This scaled was devised by Emory S. Bogardus. He formulated a large number of

relationship showing varying degree of social distances. He gave them to a
hundered judges to group them in seven categories indicating a gradually
increasing distances. Out of these seven groups one relationship most typical and
representative of the whole group was chosen from each lot. Thus following
seven relationship, in serial order towards increasing social distances were
selected. A specimen of seven categories with name of Races, as used by
Bogardus is given below.

Aschedule was supplied to each person from whom information was desired. The
following instructions given:-

1. Remember to give your first feeling reaction in each case.

2. Give your reaction to each race as a group. Do not give your reaction to
the best or the worst members that you have known, but think of the
picture or stereo type that you have of the whole race.

(DR. R. N. TRIVEDI P 241)

SOCIOMETIC MEASUREMENT

This measure of social distance was adopted by J. L. Moreno and Helen H.

Jennings, to measure the degree of attraction and repulsion between individuals
within a small group. The method is also known as Sociometery. Helen Jennings
has defined sociometery as “Means of presenting simply and graphically the
entire structure of relations existing at a given time among members of a given
group.

(DR. P. P. SHUKLA P 243)

 METHODS OF SCALING

1. THARUSTON’S TECHNIQUE

2. LIKER’S TECHNIQUE

THARUSTONS TECHNIQUE

These two methods are used for scaling attitude tests

An attitude is a dispositional readiness to respond to certain situations, persons or

objects in a consistent manner which has been learned and has became ones
tipical mode of response.

In this approach, the person express his agreement or disagreement with a

number of statements relevant to the issue. On the basis of his response, he is
assigned a score. The score, the person gets indicates his positions on the issue.
The scores reveals his attitude towards the issue.

Attitude scales focus on an individual thoughts, perceptions, feelings and

behavior towards a cognitive objective.

Thauston’s Technique: This technique of scaling attitude test is known as

method of equal appearing intervals. It is based on eleven point scale. In this
method, the statements both favourable and un favouring are a particular
problem, question or institution are obtained from a group of selected writers,
other experts and laymen and then these statements are edited and classified on
a 11 point scale by a large number of judges. The rating is done by placing each
statement in one of the 11 piles, according to the degree of favourableness or
unfavourableness of each item w.r.t the question at hand. It means that each of
the judges is requested to place in the first pile the statement, which he thinks are
most unfavourable to the psychological object in the second pile that he
considers slightly less disagree then the first one and like this in the eleven pile
the statement he consider most favourable. The 6 th position on this continuum is
defined as neutral. The 1st pile is statement considered as most unfavourable,
carries a score of 1 whereas, the last pile statements carry a value of 11. The
median of the judged locations for a item is its scale value. The statements or
irrelevant continuum are eliminated.

Before including the statements in the final scale, each is analysed for its
consistency with general attitudes found by the total scale e.g. on a scale to
determine attitudes towards church, if it is found that many persons having an
unfavourable, then that item is considered irrelevant and is discarded. Ambiguity
of an item is determined by the spread or range of judges ratings in the original
eleven point scale, given in terms of qualitative ( quartile deviation )and if an
items quartile is “high” it is eliminated.

In this method, the respondent checks those statements with which he agrees, his
score being the median of the scale values of the items he has marked. Thrusten
held that scales constructed for different attitudes by this method permit direct
caomparison of the scores of any attitudes to measured. But the validity of such
comparison however, has been questioned because :

a) The defined “neutral point” of diff. attitudes are not necessarily the same.
b) The intervals are not demonstrable equal, they are only equal appearing.

This method is useful only if the strict comparability of scores is not assumed.
Thruston and his students developed a series of scales, each consisting of
statements from extremely favourable to extremely unfavourable. The topics
included in these scales deal with attitudes toward Megroes, Chinese war,
Censorship, the Bible, Patriotism and Freedom of speech e.g. the following
statements are from the scale of attitudes towards the church. The scale value of
each is given in bracket, low values being favourable, with a range from 0 to 11.

i) I find the services of church both restful and insoiring (2.3)

ii) I think church is a parasite on society (11)
iii) I believe what the church teaches but with mental reservations (4.5)
iv) I think the teaching of the church is altogather too superficial to have
much social significance (8.3)
v) I believe in religion but I seldom go to church(5.4)
vi) I believe the church is the greatest institution in America
today(1.7)