Chapter-4

Measurement, Scaling and

Sampling
Ram Nath Neupane, PhD
Variables
• Variable is a symbol of an event, fact, characteristics, trait, attribute that
can be measured and assigned categorical values.
• Motivation, satisfaction, absenteeism, age, sex, income, etc.
• Numerical/value is assigned to the variables
• A variable is a symbol to which numerals or values are assigned- Kerlinger
• A variable is any quantity or characteristics which may
poses different numerical values or characteristics- Young
• Bivariate- men and women, employed and unemployed- o and 1 symbol
• Multivariate- highly absent, normally absent, and less absent in an
organization-1,2,3 symbol
Types
• Independent variable- A variable that brings change in other
variables
• Also called cause variable
• It influences the dependent variable in both positive and negative
way
• E.g. Employee satisfaction increases production of goods -
• If employee satisfaction increases, production of goods increases and
if employee satisfaction decreases, their performance also decreases
• Dependent variable: These variables are affected by the
change in independent variables
• Also called Effect variable
• Use of New technology increases the sale of the company
• Job satisfaction increases product of the company
Moderating Variables
• A variable that affects the relationship between independent and
dependent variable
• Remuneration (indep) - position/sex/age - job satisfaction (dep)
• Another example, examining the relationship between stress and job
performance, age might be a moderating variable.
• The relationship between stress and job performance may be
stronger for younger workers than for older workers, meaning that
age is influencing the relationship between stress and job
performance.
Intervening Variables
• The variable that modifies the relationship between dependent and
independent variables- intervening variable
• Rich person - better health care facilities- higher life expectancy-
• Renumeration- job satisfaction - performance
Measurement
• Assigning numbers or other symbols to any event, issue,
characteristics, traits, attributes, etc.
• Number is a symbol- provided as 1,2,3...or, I, II, III... used to
get information
• Measurement is the method of turning the series of qualitative fact
into quantitative series- Goode & Fact
• Measurement is the equipment of providing numbers to objects or
events according to rule- Steven
• If a researcher intends to find out the satisfaction level of employees
of an organization, they are given 1 (to very dissatisfied) and 5 (to
very satisfied) perception
Nature and characteristics of good
measurement
• Validity
• Reliability
• Practicality
Validity
• an ability of a measuring tool to measure what it intends to measure
• e.g: The tools that are used by the researcher for the collection and
analysis of data should be able to collect and analyze the data.
• A scale process its validity when it actually measures what it claims to
measure-(Goode & Halt)
Types of validity
• Content validity- representative sample of the universe , Adequate
coverage of the concept,
• Criterion-related validity: (success of measure for prediction or
estimation)
• Concurrent and predictive validity
• Construct Validity: If a measure confirms the predicated correlation
with other theoretical proposition then such measure possesses
construct validity.
Reliability
• Is related to the results of the research
• Act of generating stable and consistent results when instruments are
used in different samples and situation
• Stable and consistent results-possible through reliable data
• Methods for establishing reliability
• Test-retest method
• Alternative or parallel form method
• Split-half method
• inter rater method- several respondents rate in the in the same
instrument in similar way
Praticability
• Must be applicable in practice
• has the quality of:
• Economy: less expensive
• Convenience: clear instruction
• Interpretability: detail instruction for examination, administration,
scoring keys, and instruction, norms of reference groups, evidence of
reliability, etc.
Scaling
• A tool or mechanism by which researchers distinguish the variables
of interest to the study, in some form or the other
• it is the values, e.g. 1 for male respondent, 2 for female respondent,
0 for managerial employees, 1 for assistant managerial employee
Nominal Scale
• Useful in qualitative data
• lowest or least practice level of measurement
• Nominal scale is the scale which is used only to classify or distinguish
the variables in various categories or group
• The numbers or alphabets used in nominal scale have no numeric
meaning
• eg. Religion: Hindu-1, Muslim-2, Christian-3, Buddhist-4
Ordinal Scale

• Those scales which are used to differenetiate variables into categories in

terms of order or rank
• more preciseness than nominal
• useful in qualitative research rather than in quan
• strongly agree, agree, disagree, strongly disagree
Causes rank
political instability
povery
weak remuneration
Interval scale

• Has the quality of nominal and ordinal +

• Interval scale can specify the distance - SUCH DISTANCE CAN BE
ORDERED
• data have equal interval
• it measures the magnitude of differences in the variables
• IT MAY CREATE THE ARBITRARY ZERO
• it describes both position and distance but not value.
• For example, the 20-degree difference between 10 and 30 Celsius is
equivalent to the difference between 50 and 70 degrees. However,
these variables don’t have a zero measurement that indicates the lack
of the characteristics. For example, zero Celsius represents a
temperature rather than a lack of temperature.
• Example: There are 5 ball pens A,B,C,D and E which are priced with
adding Rs 1 on the former pencil. We can easily say that ball pen
A,B,C, and D are less expensive by Rs. 4, 3, 2 and 1 respectively than
that of pencil E but we can not say that ball pen A is how many times
expensive than of ball pen E.
Ratio scale
• most powerful of all four types due to zero origin
• the highest and most precise level of measurement
• Variables attributes can be rank ordered, the distance between them
is precisely measured and there is an absolute zero.
• For example, you can say that 4 children is twice as many as 2
children in a household. Similarly, 8 years is double 4 years of
experience.
Attitude measurement
• reaction made by a person on a particular event, product or issue-
attitude
• attitude of human being is developed on the basis of culture,
behaviour and experience
• social science researchers have developed various scales for the
measurement of attitude of the people- attitude measurement
Techniques for developing attitude scales
• Multiple choice; provide alternatives
• Rating: Ask them give the certain values to the given alternatives-
respondent estimate their values, i.e. quality, quantity, reliability, etc
• Ranking: Respondents are asked to rank the given alternatives as per
their priority
• Sorting: Respondents are given several alternatives and they have
to rank the alternatives on the basis of priority or value.
Scale construction for Attitude Measurement
Attitudes are constructs which are not open to direct observation. Therefore,
the common way of measuring attitudes is to study some aspect of behavior
and from that makes inferences about the attitudes that may be responsible for
behavior.
1. Arbitrary scale
• Researcher sometimes develops a new rational appropriate method if s/he
feels that subject of the study is abolutely new and is essential for
measurement of attitude
• popular in qualitative research
• should show the validity and reliability of such scale
2. Summated rating scale or Likert scale
• developed by Rensis Likert, popularly known as likert scale
• requires the respondents to indicate the degree of agreement and
disagreement on a series of statements
• For example,
• 1 fully disagree 2. disagree 3. undecided 4. Agree 5. Fully agree
3. Differential scale or Thurnstone scale

• developed by Robert Thurnstone to approximate the measurement in

equal appearing variables
• The researcher gathers a large number of statements that express
various points of view towards a group, event, subject,
• the selection of items is made by a panel of judges who evaluates the
items in terms of relevancy to the topic area,
• a mean score of the agreement or disagreement is calculated as the
attitude of respondent towards the topic.
The researcher selects a number of statements - usually about 20 -
which cover all the possible attitudes towards the topic of the survey. A
panel of "experts" then rate them according to their "favourableness"
towards the topic and place them along a scale, ranging from "most
unfavourable" through to "most favourable". Each statement should be
equally spaced along the scale. The statements are then incorporated
at random in a questionnaire and respondents are asked to indicate
their degree of favourableness towards the attitudes expressed.
To measure the employees’ attitude towards organization
Statements scale value (mean)
1 5.5
2 8.2
3 4.2
4. Cumulative or Guttman scale

• The Guttman scale is also known as cumulative scaling or scalogram analysis.

• It is an ordinal scale with a number of statements placed in a hierarchical
order.
• The order is arranged so that if a respondent agrees with a statement, they
will also agree with all of the statements that fall below it in extremity.
• eg: A score of 5 means the respondent is in agreement with all the statements
which indicates the most favourable attitude.
• The cumulative scale or Guttman scale measures to what degree a person has
a positive or negative attitude to something.
• gather a series of statement on which respondents express his/her agreement
or disagreement
Attitude Scales Commonly Used in Social Science/
Business Research
Likert scale
• used to measure attitude, value, norms and behaviour of individuals
or groups
• Here respondents are asked to give their opinion limiting with given
scales and they show their agreement or disagreement on the
statements provided
• eg : 1 strongly agree 2. agree 3 undecided 4 disagree 5 strongly
disagree
Semantic Differential scale
• A semantic differential scale is a survey or questionnaire rating scale
that asks people to rate a product, company, brand, or any 'entity'
within the frames of a multi-point rating option (usually 5 to 7 point
scale)
• It is based on the assumption that an object can have several
dimension of implied meaning.
ctd
• A great example of the Semantic Differential Scale is the Single Ease
Question (SEQ). The SEQ presents a 7-point scale with the ends
labeled very easy and- very difficult.
• The semantic differential technique involves presenting pairs of
bipolar, or opposite, adjectives at either end of a series of scales, such
as the following:
• Weak o o o o o o o Strong
• Ugly o o o o o o o Beautiful
• Cool o o o o o o o Warm
• Amateur o o o o o o o Professional
• Graphic rating scale
• respondents are asked to indicate their value to a particular item by
placing a mark at the appropriate point to express their opinion.
• eg, respondents are asked to mark at one point ranging from 0-10
(very bad- right- excellent) which presents the opinion of respondents
- positively or negatively
• Itemized rating scale
• Researcher provides a category of responses out of which
respondents select one that is most relevant
• How do you rate your interest in changing organizational policies?
• extremely poor not at all
• some what well very well
• Rank order rating scale
• Respondents are asked to rank the given items or product on the
basis of their priority- highest to lowest priority
• Other simple scales
• simple category questions- Yes or No
• multiple choice questions
• open ended questions
SOURCES OF MEASUREMENT PROBLEMS
• Related to Respondents: reluctant to answer, little knowledge,
tiredness, anxiety,
• Related to situation: situational factors- rapport bulding, natural
environment, incentive, time , place,
• Related to measure: careless in making questions, rewarding or
reordering questions , mechanical processing, coding,
• Related to instrument: defective measuring instrument, use of
complex or unabiguous word or statements, inadequate spacing for
responses, exessive length,
Population, sample and sampling
• population also called universe refers to the total items or units of
information under the study
• sample- representative units of the study
• sampling- process of selecting few (each units) from larger population
• Act of designing a sample for purposeful research
• Sampling is necessary when there is a large and heterogeneous
population
• Sampling is some elements of population which helps to draw
conclusions about the entire population- Cooper and Schindler
Types of sampling method
• Probablity or random sampling
• simple random sampling
• complex random sampling (mixed)
• stratified random sampling
• cluster sampling
Non-random or non-probablity sampling
• Purposive sampling (judgemental)
• Quota sampling
• convenience sampling (as per the easiness of the researcher)
• self selecting- after announcing from mass media, one can join in the
group
• snow ball sampling
Sampling Errors
• A sampling error occurs when the sample used in the study
does not represent the entire population.
sampling errors
sampling errors
Non-sampling errors
• Non-sampling error refers to all sources of error that are
unrelated to sampling.
• Non-sampling errors are present in all types of survey,
including censuses and administrative data.
• They arise for a number of reasons: the frame may be
incomplete, some respondents may not accurately report data,
data may be missing for some respondents, etc.