Download as pdf or txt
Download as pdf or txt
You are on page 1of 13

02-Dec-20

Topic-9
I welcome all of you to the sessions on Measurement and

Scaling.

Out line of the session


Measurement and Scaling
•Meaning
•Levels of measurement or types of scales
•Mathematical comparison basis (nominal, ordinal, interval and ratio)
•Other scales (Thurstone, Likert, Guttman, and Bogardus)
•Criteria of good measurement (reliability, validity, and sensitivity)

Reference for this session


Sample questions from this session

1
02-Dec-20

Measurement and Scaling

I AGAIN welcome all of you to the sessions on

Measurement and Scaling.

Measurement; meaning

•Measurement means the process of assigning numbers to objects


or observations, the level of measurement being a function of the
rules under which the numbers are assigned.

•e.g. Low 1, Medium 2, High 3

•True 1, False 0

•Married 1, Unmarried 2

2
02-Dec-20

Scaling; meaning
• Scaling is progressively arranging series of items according to
value or magnitude into which an item can be placed according to
its quantification.

• Scaling is thus constituting a series of categories through marks.

• e.g. salary (monthly in BDT)


• Upto 10,000 10,000-20,000 20,000-30,000 …..

Types of scales; based on mathematical comparison


Based on mathematical comparison scales are of four types.
These are

1. Nominal
2. Ordinal,
3. Interval, and
4. Ratio

I am going to discuss these 4 types first. Then, I will discuss the


remaining 4 types.

3
02-Dec-20

Nominal scale

•Nominal scale classifies individuals into two or more categories, the


members of which differ with respect to the specific characteristics.

•However, the categories have no rank order. Implies provide no


information regarding order or magnitude.

•e.g. male and female, young and old, rural and urban, rich and poor.

Continued

•The nominal scale is often called classificatory scale.

•Useful for quantifying qualitative data.

•The rules for assigning numbers in this scale is very simple.

•e.g. male 1, female 2, young 1, old 2, …..

4
02-Dec-20

Ordinal scale
Ordinal scale ranks individuals along the continuum of the
characteristic being scaled, say, from highest to lowest, greatest to
least, first to last, and so on.

e.g. undecided 1, bad 2, good 3, better 4, best 5

The ordinal scale provides a rank order of categories and arranges


objects according to their magnitude in an ordered relationship.

However, it does not describe the distance or interval between the


objects.

Interval scale
• In interval scale, the distance between the numbers or units on the
scale is equal and the direction is known.

• e.g. salary (monthly in BDT)


• 10,000-20,000 20,000-30,000 30,000-40,000….

• This scale provides more powerful measurement than that of


ordinal scale and robust statistical measures (e.g. mean, standard
deviation) can be used in interval scale.

5
02-Dec-20

Ratio scale
Highest level of measurement
An interval scale with an absolute zero (0) point
e.g. Length (in cm) 0-10, 10-20, 20-30
Subsumes of above all three levels of measurement
Explains proportion of one value to another
e.g. the ratio of female crime to male crime is 1:19, i.e. for every 5
female criminals there are 95 male criminals.

Comparison of four scales; a synopsis


Criteria Nominal Ordinal Interval Ratio

Properties of Naming Naming and ranking Naming, ranking and Naming, ranking, equal
measurement equal intervals intervals and zero point

Nature of measurement Categorical Ranking Scoring Scoring

Examples Gender: Male and Income: Upper, middle IQ of A is 25% higher IQ of A to B is 1:1.5
female and low than B.
Residence: Rural and
urban
Nature of underlying Discrete Discrete or continuous Continuous Continuous
construct

Mathematical functions None None Addition and Addition, subtraction,


subtraction division, multiplication

Central tendency Mode Median Mean Mean

Statistical tests X2 test, Lambda test, U-test, Spearman’s p, Pearson’s r, t-test Pearson’s r, t-test
Phi Gamma

6
02-Dec-20

Types of scales; based on other basis


•On other basis scales are of four types.

•These are

1. Thurstone
2. Likert
3. Guttman, and
4. Bogardus

Now I am going to discuss these 4 types in short.

Thurstone scale
• Thurstone scale – A type of composite measure constructed in accord
with the weights assigned by “judges” to various indicators of some
variables.

• It allows composing a large number of statements by the researcher


concerning the attitude to be measured.

• The statements/items are then given to a number of judges to order


them in a continuum (generally from 1 to 11)

• It also allows to review the significance of each separated item and to


calculate median value for each item.

7
02-Dec-20

Likert scale
• Likert scale – A type of composite measure, designed to improve the
levels of measurement in social research through the use of
standardized response categories to determine the relative intensity
of different items.

• Ask an opinion question on a 1-to-5, 1-to-7, scale etc.

• It is a bipolar scale; has a neutral point and scale ends are at opposite
positions of the scale starting point

• It allows summing up the value of each item that is checked, a total


score for each respondent is calculated.

Guttman scale
Guttman scale: respondent checks each item with which they
agree; constructed as cumulative, so if you agree to one, you
probably agree to all of the ones above it in the list.

Cumulativeness implies that the component items can be ordered


by ‘degree of difficulty’ and that the respondent who reply
positively (or favorably) to a difficult/complex item (question) will
always respond positively to less difficult items or vice versa.

This is a method of discovering and using the empirical intensity


structure among several indicators of a given variable.

8
02-Dec-20

Bogardus scale
• Bogardus Social Distance Scale – A measurement technique for
determining the willingness of people to participate in social relations
– of varying degrees of closeness – with other kinds of people.

• In this scale there is a logical structure of intensity inherent among the


items.

• e.g. in case of relationship between upper class Hindu and


Dalits/Tribal in India different type of associations, say, as office
colleagues, as neighbors, as friends, and as marriage partner.

• If a person is willing to accept a Dalit as marriage partner he/she must


be willing to have associations of lesser-intensity.

Criteria of good measurement


• Three important criteria of good measurement are

• reliability,

• validity, and

• sensitivity

9
02-Dec-20

Reliability
• Reliability – is that quality of measurement method which
suggests that the same data would have been collected each time
in repeated observations of the same phenomenon.

• Reliability is the consistency of measurement, or the degree to


which an instrument measures the same way each time it is used
under the same condition with the same subjects.

• There are several methods of reliability testing. These are i) Test-


retest; ii) internal consistency; iii) split-half; and iv) equivalent
form.

Validity
• Validity means the ability to produce findings that are in agreement
with conceptual or theoretical values.

• It can be of face, content, criterion, and construct validity.

• Important here is to know the difference between reliability and


validity.

• An example can be of helpful. Lets say 70% of the respondents in


survey and again in retest answered in favor of using a particular
product of a particular brand. Hence, the response is reliable.
However, it is still may not be valid. As because may be due to high
price actually a lower percentage (lets say 40%) is practically using
that product of that brand. Hence, 40% is valid response.

10
02-Dec-20

Reliability and Validity -- Relationship


The relationship between reliability and validity can be confusing
because measurements and research can be reliable without being
valid, but they cannot be valid unless they are reliable.

 For a study to be valid it must be consistently (reliability) do


what it purports to do (validity).

 For a measurement to be judged reliable it should produce a


consistent score.

Continued
 For the research study to be considered reliable each time it is
replicated it too should produce similar results.

It is my belief that validity is more important than reliability


because if an instrument does not accurately measure what it is
supposed to, there is no reason to take into cognizance the result
even if it measures consistently same.

11
02-Dec-20

Sensitivity
Sensitivity refers to instrument’s ability to accurately measure
variability in responses. A dichotomous response category such as
‘agree or disagree’ does not reflect attitude changes.

A more sensitive measure with numerous items on the scale may be


needed.

e.g. 5-point scale (strongly agree, agree, neither agree nor disagree,
disagree and strongly disagree) increases as scale’s sensitivity.

By assigning ‘zero’ (0) to third type of answer and adding +2 and -


2 is one step more to clarify attitude accurately.

Reference for this session


Ahuja, Ram, Research Methods, Rawat Publications, Jaipur and New
Delhi, Second Reprint, 2005, PP 329-350.

&

Kothari, C. R., Research Methodology: Methods and Techniques,


Second Edition, New Age International Publishers, New Delhi, 1990, PP
69-94.

12
02-Dec-20

Sample questions from this session


1) What is measurement? Explain the major criteria of a good
measurement.

2) What is scaling? Comparatively discuss the types of scale based


on mathematical comparison.

3) Discuss any three of the following scales with example:


a) Thurstone
b) Likert
c) Guttman, and
d) Bogardus

Thanks 

13

You might also like