2. Measurement
Production No.of shifts, overtime, total worth of stock of materials…
Marketing Customers, salesmen, brand loyalty …
Finance Investment plan, lending rate, credit period …
Personnel Hiring/firing, training programs, promotion, incentive
schemes …
- Systematic way of assigning numbers or names to objects and their
attributes according to some specified rules.
Easy Convert variable into Quantitative variable => Measured with
standard units/scales
Ex: Height,weight etc.
- Study male to female attendance ratio of a show
Difficult Qualitative/Abstract data [Qualitative Attitude, Perception,
preferences etc. Abstract Brand image, leadership style etc. ]
3. Scaling
Extension of measurement
“The procedure for the assignment of numbers to a property of objects in
order to impart some of the characteristics of numbers to the properties in
question.”
1. Judgement 2. Questionnaire
Classification of measurement scales:
1) Nominal
2) Ordinal
3) Interval
4) Ratio
4. Nominal Scales
Categorical scale
Assigning number symbols to events for identification. Ex: Numbers on
jerseys, Registration number
Convenient way of keeping track of people, objects & events
Least powerful
No order, distance relationship
Used in surveys
Mode, Chi-square test
Ex: What is your PAN card number?
5. Ordinal Scale
Ranking Scale – Used for ranking in most market research studies
Ex: consumer preferences, perception etc.
Places events in order
Greater than or less than
Real distance between adjacent values may not be equal
Mode, median, Percentile & quartile measures, Correlation(some extend),
non-parametric tests
6. Interval Scales
Cardinal level of measurement
Have properties of order and distance
Intervals are adjusted in terms of some rule that has been established as a
basis for making the units equal
Standard units equal distance
Addition & substraction
Mean, Standard deviation, Correlation, T test, F test
7. Ratio Scale
Highest measurement scale
Attitude measurement- general use
All form of arithmetic operations
All statistical operations can be performed
Zero has real meaning
Facilitates comparison. Ex: A’ typing performance is twice as good as B.
8. Sources of error in measurement
Respondent Guess (Reluctance/little knowledge)
Transient factors(boredom, fatigue etc.)
Situation Any condition/ situation straining interview
Measurer Rewording & reordering questions/ behavior, style, looks /
Mechanical processing etc.
Instrument Wording, Printing mistakes, inadequate space for replies,
response choice omissions
If “errors” are not eliminated/ neutralized/ dealt with final result will be
contaminated
Instrument Researcher – completed / Subject completed
9. Process of developing measurement
tools
Concept development Major concepts to be studied (Ex: Company image)
Specification of concept dimensions Ex: product reputation, customer
treatment, social responsibility, corporate leadership
Selection of indicators For measuring specific questions, scales, or other
devices by which respondent, knowledge, opinion, expectation etc. are
measured
More than 1 indicator stability to scores improves validity
Formation of index Combining various indicators
- Provide scale values to the responses & sum up corresponding scores
10. Test of sound measurement
Test of Validity The extend to which a test measures what we actually wish
to measure
Content validity Adequate coverage of topic under study, Representative
sample of universe, Judgemental and intuitive – panel of experts
Criterion related validity refers to [usefulness of a test in ..]
Predictive validity – ..predicting some future performance
Concurrent validity – ..Closely relating to other measures
- Criterion must possess relevance, freedom from bias, reliability, availability
Construct validity- Assessment of suitability of measurement tool to measure
the phenomenon – panel of experts
11. Test of reliability Consistent results, accuracy and precision of a
measurement procedure
Valid instrument is always reliable but reliable instrument is not valid
Reliable No interference of transient & situational factors
2 aspects
1) Stability Consistent result Repeated measurements, same person,
same instrument
2) Equivalence How much error can get introduced by different
investigators or different samples of the items being studied
Improve reliability :
(a)Standardising conditions
(b)Carefully designed directions, trained investigators, broadening
samples
12. Test of Practicality
Economy
Budget, length of measuring instrument, data collection methods
Convenience
Easy to administer proper layout
Interpretable
Instructions during test, scoring keys, guides for using test & interpreting result etc.
Features of a good measurement tool/instrument:
Validity, measurability, reliability, unidimensionality, linearity, practicability, accuracy
Appropriateness of scales:
Full knowledge of population/universe
Factors – logically related & continued measurement measurable/scalable
13. Scale construction techniques
Statements must elicit responses which are psychologically related to the
attitude being measured
Statements need be such that they discriminate not merely between
extremes of attitude but also among individuals who differ slightly
- Trying to measure “expressed” opinion and draw inferences from about real
feelings and attitudes
14. Construction of Scales : 5 approaches
I. Arbitrary approach:
- Develops on adhoc basis
- Statements(unmistakable & suitable) based on topic select some Makes scales
- “+” Easy, inexpensive
- “-” Statements can’t be generalized, Not clear to respondent’ logic
II. Consensus approach:
- Statements selected by panel of judges
- Statements Topic, meaningful& clear, Level of attitude
- Attitude studies Differential scale (L.L.Thurstone)
15. III. Item analysis approach:
- Statistical technique for selecting / rejecting items of the test based on
difficulty value and discriminating power
- To select appropriate items, modification of items, difficulty value? ,
Discriminating power (capable & less capable respondents)
- Order of questions easy to difficult
- Summated scale(Likert scale)
IV. Cumulative scale approach:
- Respondents given a set of questions(Yes/ No)
- Cumulative scale (Guttmann scale)
1. Would you read the advertisement for admission fully?
2. Would you buy application from the institution after reading its ad for
admission?
3. Would you submit the filled – in application to the institution?
4. Would you sit for the written test of the institution?...
16. V. Factor analysis approach
- To measure human perceptions and preferences towards some stimuli(objects),
like products, organisations, places, events, brands etc., and position them in a
perceptual place
- Semantic differential scale & multi dimensional scaling
- Not widely used
-
17. Scaling techniques
1.Rating scales 2.Ranking scales
Dichotomous scale Paired comparison
Category scale Comparative scale
Likert scale Forced choice
Semantic scale
Numerical scale
Itemised rating scale
Graphic rating scale
Fixed/constant scale
Rating Judging an object against a specified criteria
Ranking Relative judgements against other similar objects
18. Dichotomous scale
- Question has only 2 possible responses Yes/No, True/False, Agree/Disagree
- Nominal scale
- Ex: I enjoy eating pizza at Dominos
--- True ---False
Do you like eating pizza at Dominos?
--- Yes ---No
Category scale
- Question has multiple possible responses to select single responses
- Nominal scale
- Ex: Which is your preferred brand in baby clothing?
Toffyhouse--- Babyhug--- Carters --- Mark&Mia--- Yellowduck --- Others---
19. Likert scale
- Rensis Likert – Summated scale
- Attitude measurement
- 5 point scale [6/7/8 point are Likert type scales]
- Ex:
-2 -1 0 +1 +2
- Single item(direct question) or Multiple item (indirect question)
- Steps in construction
- Merits & demerits
Semantic scale
- Use of words in scale – Bipolar attributes
- Attitudinal scale
- Good Bad
Etremely Quite Slightly Neither Slightly Quite Extremely
TV advertisement is
monotonous
Strongly
Disagree
Disagree Neither
Agree nor
diagree
Agree Strongly agree
20. - Semantic differential scale 7 point bipolar – To measure intensity of judgement –
Use of adjectives (two end terms ONLY)
Numerical scale
- Similar to semantic diff. scale
- use adjectives – Bipolar with numbers in between
- Useful Useless
5 4 3 2 1
Itemised rating scale
- Presents a series of statements(ordered progressively) from which respondent selects
one as best reflecting his evaluation
- Numbers and/or brief descriptions associated with each category(Ref. text)
- Ex: How well Mr. Jai get along with his fellow workers?
He is almost always involved in some friction with a fellow worker
He is often at odds with one or more of his fellow workers
He sometimes gets involved in friction
He infrequently becomes involved in friction with others
He almost never gets involved in friction with fellow workers
21. Graphic rating scale
- Various points are put along the line to form a continuum and the rater indicates
his rating by simply making a mark ( ) at the appropriate point on a line that runs
from one extreme to the other
- Ex: How do you like the product?
Like very much Like some what Neutral Dislike some what Dislike very much
Fixed/constant scale
- Distribute points across various items
- Ex: Distribute the total points among the various attributes of Pears soap
Colour
Smell
Size
Shape
Quality of
foam
Total points 100
22. Paired comparison
- Comparative judgement
- No.Of pairs = n(n-1)/2 [Statement 5; 5(5-1)/2=(5*4)/2=10 pairs]
Comparative scale
- Ordinal/rank order properties
- Direct comparison
- Brand, product or feature against another
Forced choice
- Ex: Given 10 brands. Rank first three
- Careless responses