6.

CHARACTERISTICS
OF
MEASUREMENT
TECHNIQUES
-Reliability, Validity Etc.
What do mean by Measurement?
 It implies that we are using the Numbers to describe
a specified features of a given Product/Service.
 I.e.-in quantitative terms.

 This is true in case of physical sciences,But in case

of social sciences(marketing, sales, advertising
etc), measurement is considered in qualitative as
well as quantitative terms.
 [Link] response for a recently introduced ball
point pen.
What do mean by Measurement—cont--
 Here we do not measure buyer/product
but,we measure certain characteristics
of buyer/product like buyer’s
preference, usage rate, income, attitude
[Link] product’s size, colour, it’s
suitability for the purpose etc.
 Hence, measurement is the assignment
of No. of characteristics of an object or
event as per rules.
 Types and
characteristics of Measurement scales.
 [Link] Scale:Here the [Link] as a
label to identify a person, object or
event.e.g.a table showing the graduate and
post-graduate students, hostellers and
others,and their total
 [Link] Scale:as the name implies are
ranking [Link] of two or more
households as per their annual income.
 Types and characteristics of
Measurement scales:—cont--
 [Link] scale:It has additional strength
which is determination of the equality of
[Link] measurement of
temperature, both the centigrade and
Fahrenheit scales belongs to this type.
 [Link]=5/9([Link]-32)
 [Link] Scale:It has order,distance and unique
origin and most superior than all scales.e.g.9 kg
and 45 kg are in the ratio of 1:[Link] we convert it
into gms, we obtain the same ratio.
 So these are in the increasing order of
sofistication from data analysis point of view.
 Types and characteristics of
Measurement scales:—cont--

Scale Number Marketing Permissible

system Phenomenon statistics
Nominal Unique defin. Classification: Percentages,
of Nos.(0,1,..9) Brands,male- Mode,Binomial
female,Stores test,Chi square
types,sales test.
territories
rankings
Ordinal Order of Attitude,Prefer Percentiles,Me
numbers ence,Occupati dian,Rank
(0<1<2--<9) on,Social class order
correlation,Two
way annova
 Types and characteristics of
Measurement scales:—cont--
Scale Number Marketing Permissible
system Phenomenon statistics
Interval Equality of Attitude, Range,Median,
differences Openions, Standard
(2-1=7-6) Index deviation,t test,
Numbers. f test etc.

Ratio Equality of Sales(units/ Geometric

ratios Rs)Units mean,
(5/10=3/6) produced , Harmonic
Cost, No. of mean,Coefficie
customers -nt of variation.
 Difficulty of Measurement:
e.g. Conducting a survey on existence of
medical facilities in certain locality of the city.
Measurement in the area of social
sciences is comparatively more difficult
than physical sciences.

Physical Sciences Social Sciences

Length Attitude
Weight Preference
Volume Happiness
Time Creativity
 At this stage we must know the
SOURCES OF ERROR.
 [Link]
(status/education/social class,nature of job)
 [Link]
(situation in which an interview is being held
influences measurement.)
 [Link]
(method used by the interviewer)
 [Link] instrument of Data collection.
(type of questions and the questionnarie-
lengthy,ambiguous,language,layout,space for
answers,printing,font size,quality of paper etc.)
 There are two important criteria
for ensuring good scale
 RELIABILITY  VALIDITY
[Link]-Retest [Link] Validity
reliability [Link] Validity
[Link] Alternative [Link]
forms reliability Validity
[Link]-Half [Link]
reliability Validity
Reliability :[Link]-Retest Reliability
 It involves repeated measurement of the
same respondent/group using the same
scaling technique under same
conditions at different times
 The scores of the same type of test with
same respondent/group at different
times are correlated
 If the correlation is low, then the
reliability is too less.
Reliability:[Link]-forms Reliability
 It involves the same respondent being given
a set of two forms.
 The two forms are not identical but
considered equivalent.
 The result of these two forms are compared
to ensure, whether there is a considerable
difference between the two scores of the
forms & reliability is checked accordingly.
 It is more time consuming and expensive too.
Reliability:[Link]- Half Reliability
 Split-half technique is used in case of multi-
item instrument.
 It is another version of alternative forms
technique.
 It involves the splitting of multi item
instrument into two equivalent groups
 Then item response of the two groups is
correlated and reliability is estimated
accordingly.
 A high degree of correlation indicates that
there is similarity or homogeneity among the
items.
 Validity:[Link] Validity
 It implies that the content of the scale corresponds
to the content of attitude system and that they are
comprehensive enough to cover the full range of
the attitude.
 A researcher should define the problem clearly,
identify the item to be measured and evolve the
suitable scale for the purpose.
 Despite this, critics are there on the existence of
content validity in the given scale since it is
dependent on the judgement of the researcher
which vary from individual to individual.
 To avoid this, it’s better to approach a group of
knoledgeable person.
 Validity:[Link] Validity
 It is based in theoretical consideration.
 With this test we can measure the attitude
indirectly.
 Here the researcher postulates the nature and
the extent of association between the attitude
and the other specified variables.
 Then researcher examines whether this
relationship exists. If not then there could be
two explanations (a) The scale might be
invalid(b) His theory might be deficient.
 [Link] of an individul in a society and their
association.
Validity:[Link] Validity
 Itsignifies how best the researcher can
guess the future performance, from his
knowledge of attitude score.
 The obtained score is correlated with
earlier predicted score. If they are
closely associated then the scale is said
to have predictive validity.
 E.g opinion questionnaire which is a
basis for correct forecasting of the
demand of a product.
Validity:[Link] Validity
 Here , an attitude scale on one variable can
be used to estimate the score on another
variable.
 [Link] may decide the social status of
respondents on the basis of their attitude
towards savings. In this example the attitude
scale has also the criterion measure are
done at the same time.
 A high degree of concurrent validity may
sometimes not actually correct since
collection of one set of data may influence the
collection of another set.
 Criteria for good scale--
cont-----

 Apart from reliability and validity, there

are following other factors responsible
for good scale:
 Practicality
 Economy
 Convenience
 Interpretability. Etc.
THANK YOU--------------