Nature of Scale of Measurement in Statistics:

In teaching learning process as well as in the field of educational research measurement occupies an
important place. Measurement is a process through which observations are translated in to numbers.
The nature of measurement process produces the numbers. These numbers determines the
interpretation that can be made from them and the statistical procedures that can be meaningfully used
with them.

The first step in the measurement procedure is to define the objects, traits or the phenomenon under
measurement. For his purpose we have to classify the objects of our interest. We have to put them in to
different categories. But the simplicity of tins procedure seems to cause difficulty for students. People
spend much of their time on categorizing things, events and individuals. This process of classification
with measurement seems difficult.

Types of Scales of Measurement:

Stevens has classified the measurement as Nominal Scales, Ordinal Scales, Interval Scales and Ratio
Scales.

1. Nominal Scale:

The most primitive scales of measurement is the nominal scale. Nominal measurement involves the
placing of objects or individuals into categories which are qualitatively rather than quantitatively
different. Measurement at this level requires only that, one be able to distinguish two or more relevant
categories and know the criteria for placing the individuals or objects into one or other categories.

At this level the required empirical operation involves recognizing a given individual or object belongs to
a given mutually exclusive category or that it does not. The relationship between the categories is that
they differ in quality. It does not indicate that they represent more or less of the trait being measured.
Classification of students in to Section-A and B, Boys and Girls, Base ball players and Foot ball players,
Hindus and Muslims etc. constitute nominal measurement.

Sometimes numbers are used in the nominal measurement. Here numbers are assigned only in order to
identify the categories. The numbers are assigned arbitrarily to categories merely as labels or names.
Players in a team are assigned such numbers, telephones are assigned such numbers.

Groups may be given labels 1, 2 and 3 or A1, A2 or A3 Here all the members of a category are assigned
the same number and no two categories are assigned the same number. For example in preparing data
for a computer the numeral ‘0’ might be used to represent a male and ‘1’ for a female. Here the two
numbers have no mathematical relationship. Hence 1 is not greater than ‘0’.

Numbers in a nominal scale do not represent absolute or relative amount of any characteristic. They
merely serve to identify the member of a given category. In a nominal scale the identifying numbers can
never be arithmetically manipulated through addition, subtraction, multiplication or division. Those
statistical procedures only based on counting such as reporting the number of observations in each
category can be calculated. X2 (Chi- square) and mode can be calculated from the data on nominal
measurement.

2. Ordinal Scale:

Ordinal scale is the next higher scale of measurement. It indicates the relative position of the individuals
or objects with respect to certain attribute. But it does not indicate the distance between the-positions.
At this level the essential requirement for measurement is an empirical criterion for ordering individuals,
objects or events with respect to the attribute.

Ordinal measurement requires that the objects of a set can be rank-ordered on an operationally defined
characteristic or property. When a teacher ranks his students on certain characteristics such as their
social maturity, spelling ability, singing ability, leadership ability etc. an ordinal measurement occurs. In
an ordinal measurement the empirical operation involves only direct comparison of the objects or
individuals in terms of the extent to which they possess the attribute.

In this scale when numbers are assigned to the individuals or objects the only information considered is
the order of objects. Here the number or the rank only shows the order neither the difference nor the
ratio. So the ordinal numbers do not indicate absolute quantities; nor do they indicate that the intervals
between the numbers are equal.

When numbers 1, 2, 3 and so on are used in ranking there is no empirical distance between the rank of 1
and 2 and 2 and 3. It may be same, less than or greater than. There is simply no basis for interpreting the
magnitude of difference between numbers or the ratio of numbers.

Race is a good example of ordinal scale. In a race the runners are ranked as 1st, 2nd, 3rd and so on. Here
we can say that 1st individual was faster than the 2nd individual. But we cannot say how much faster he
was? And the difference between 1st and 2nd and 3rd may not necessary be same.

As the size of the intervals between the categories is not known so the statistical operations are limited.
Any statistical procedure that assumes equal intervals cannot be used in ordinal scale.

The major statistical procedures that can be worked out in ordinal scale are:

Median, Percentiles, Rank difference Correlation (ρ).

3. Interval Scale:

Interval scale is the next higher scale to the ordinal scale. It possesses the characteristics of nominal and
ordinal scale. “An interval scale is one that provides equal interval from arbitrary origin”. Interval scale
not only orders the individuals, objects or events according to the amount of attribute they represent
but also establishes equal intervals between the units of measurement.

For example we have measured four students on an interval scale and got the scores 80, 60, 50 and 30.
Here we can say that the difference between the 1st and 2nd is 20 and 3rd and 4th is 20. So the
difference between the 1st and 2nd is equal to the difference between 3rd and 4th.
The Fahrenheit and centigrade thermometers are examples of interval scale. On an interval scale both
the order and distance relationship among the numbers have meaning. We may assert that 50°C ~ 52°C
= 25°C ~ 27°C (~ denotes difference between). But we cannot say 50°C is twice as hot as 25°C. It is
because the zero point on an interval scale is not true zero point. It is an arbitrary zero point.

It is established by convention that the zero point on a psychological or educational measurement is

arbitrary. It is not fixed zero point. Therefore we cannot find or identify an individual with zero
intelligence or achievement. For example three students have scored 15, 30 and 45 on a statistics test.
We cannot say that 30 and 45 are twice or thrice of 15.

Thus because the ‘0’ point is arbitrary. In the interval scale multiplication and division are not
appropriate. However the difference between portions on an interval scale may be reported or the
numbers may he added.

Statistical procedures which are based on addition and sub-traction and procedures appropriate for
nominal and ordinal scales may be used in interval scale. Most of the common statistical procedures like
Mean, Standard Deviation (δ), Product moment correlation (r), Analysis of variance (ANOVA), Analysis of
co-variance (ANCOVA) etc. can be worked out from the data in the interval scale.

4. Ratio Scale:

Ratio scale involves the highest level of measurement. A ratio scale, in addition to the characteristics of
nominal, ordinal and interval scale possesses an absolute or fixed or natural zero point that has
empirical meaning. Ratio scale provides a true zero point as well as equal interval. Ratios can be formed
between any two given values on the scale.

The example of ratio scale is the yardstick used to measure length in inches or feet. Almost all the
physical measurements like Metre, Litre, Kilogram etc. are ratio measurements. The origin in this scale is
an absolute ‘O’ corresponding to no length at all. In an interval scale a ‘O’ score in Mathematics does not
mean zero knowledge in mathematics but ‘O’ length in a ratio scale means no length at all.
So that it is possible to state that a stick of 8 feet long is twice as long as a stick of 4 feet. It is possible
with ratio scale to multiply or divide each of the values by a certain number without changing the
properties of the scale. For example we can divide 2000 grams by 2 to convert the measurement to 2 kg.
In educational measurement only a few variables come under ratio scale. These variables are largely
confined to motor performances. All types of statistical procedures are appropriate with a ratio scale.