Ch07 Measurement and Scaling

After reading this chapter, the student should be able to:
1. Introduce the concepts of measurement and scaling

and show how scaling may be considered an extension
of measurement.
2. Discuss the primary scales of measurement and

differentiate nominal, ordinal, interval, and ratio scales.
3. Classify scaling techniques as comparative and

noncomparative, and describe the comparative
techniques of paired comparison, rank order, and
constant sum.
Copyright © 2015 Pearson Education Copyright © 2015 Pearson Education Chapter 7 - 2
4. Describe the noncomparative scaling techniques,

distinguish between continuous and itemized rating scales,
and explain Likert, semantic differential, and Stapel scales. 7. Discuss the considerations involved in implementing
scales of measurement in an international setting.
5. Discuss the variations involved in constructing itemized
rating scales. 8. Explain how social media can be used to implement
comparative and noncomparative scaling techniques.
6. Discuss the criteria used for scale evaluation and explain
how to assess reliability and validity. 9. Understand the ethical issues involved in selecting scales
of measurement.
Copyright © 2015 Pearson Education Chapter 7 - 3 Copyright © 2015 Pearson Education Chapter 7 - 4
Step 3 enlarged
Step 3: Formulating a Research Design

Step 1: Defining the Problem
• Secondary & Syndicated Data Analysis
Step 2: Developing an Approach to the Problem • Qualitative Research
• Survey & Observation Research
Step 3: Formulating a Research Design • Experimental Research
• Measurement & Scaling
Secondary & Syndicated Survey & Observation Questionnaire & Form Sampling Process & Preliminary Plan of Data
Qualitative Research Experimental Research Measurement & Scaling
Data Analysis Research Design Sample Size Analysis
• Questionnaire & Form Design

Step 4: Doing Field Work/ Collecting Data • Sampling Process & Sample Size
• Preliminary Plan of Data Analysis
Step 5: Preparing & Analyzing Data
Step 6: Preparing & Presenting the Report

1
Measurement means assigning numbers or other Scaling involves creating a continuum upon which
symbols to characteristics of objects according to certain measured objects are located.
pre-specified rules.
Consider an attitude scale from 1 to 100. Each
One-to-one correspondence between the numbers respondent is assigned a number from 1 to 100, with
and the characteristics being measured. 1 = Extremely Unfavorable, and 100 = Extremely
Favorable.
The rules for assigning numbers should be
standardized and applied uniformly. Measurement is the actual assignment of a number
from 1 to 100 to each respondent. Scaling is the
Rules must not change over objects or time. process of placing the respondents on a continuum
with respect to their attitude toward department stores.
Age in Years
Description 120 Very Old Description
By description we mean the unique labels or descriptors that are used to
designate each value of the scale. All scales possess description.
Golden
Years
Order
By order we mean the relative sizes or positions of the descriptors.
80
Order is denoted by descriptors such as greater than, less than, and
equal to. Seniors
60
Distance Middle Order
The characteristic of distance means that absolute differences between Aged
40
the scale descriptors are known and may be expressed in units.
Young
Distance
Adults
Origin 20
The origin characteristic means that the scale has a unique or fixed Youth
beginning or true zero point.
Origin 0 New Born
Copyright © 2015 Pearson Education Chapter 7 - 10

Primary Scales  In a nominal scale, the numbers serve only as labels or

tags for identifying and classifying objects.
Ratio Highest level of  When used for identification, there is a strict one-to-one
measurement
Scale correspondence between the numbers and the objects.
Interval
Scale
 The numbers do not reflect the amount of the
Ordinal characteristic possessed by the objects.
Scale
Lowest level of Nominal
measurement Scale
 The only permissible operation on the numbers in a
nominal scale is counting.
2
 An ordinal scale is a ranking scale in which numbers  In an interval scale, numerically equal distances on the
are assigned to objects to indicate the relative extent to scale represent equal values in the characteristic being
which the objects possess some characteristic. measured.
 It permits comparison of the differences between objects.

 Can determine whether an object has more or less of a
characteristic than some other object, but not how much
more or less.  The location of the zero point is not fixed. Both the zero
point and the units of measurement are arbitrary.
 Any series of numbers can be assigned that preserves  It is permissible to shift the origin, for example by adding
the ordered relationships between the objects. a constant, and that will result in an equivalent scale.
 It is not meaningful to take ratios of scale values.

 A ratio scale possesses all the properties of the nominal, Scale

ordinal, and interval scales. Nominal Numbers
Assigned Finish
to Runners 7 8 3
 It has an absolute zero point.

Ordinal Rank Order Finish
of Winners
3rd 2nd 1st
 It is meaningful to compute ratios of scale values. place place place
Interval Performance
Rating on a 8.2 9.1 9.6
 Ratio scales allow only proportionate transformations
0 to 10 Scale
that involve multiplication by a positive constant.
Ratio Time to Finish 15.2 14.1 13.4
in Seconds
Marketing
Primary Scale Basic Characteristics Common Examples NOMINAL SCALE ORDINAL SCALE INTERVAL SCALE RATIO SCALE
Examples
Nominal Numbers Social Security Brand numbers;
Identify and classify objects numbers; numbering store types; sex
of football players classification Preference Ratings $ spent ¢ spent
No.
Store Preference Ranking in last 2 in last 2
Ordinal Numbers indicate the relative Quality rankings; Preference months months
1-7 11 - 17
positions of the objects but rankings of teams in a rankings; market
not the magnitude of tournament position; social
differences between them class 1. Burger King 4 79 4 14 10 1000
Interval Differences between objects Temperature Attitudes; 2. McDonald’s 2 25 7 17 50 5000

can be compared; zero point (Fahrenheit, Celsius) opinions; index
is arbitrary numbers 3. Sub Way 5 82 3 13 0 0
Ratio Zero point is fixed; ratios of Length; weight Age; income; 4. KFC 3 30 6 16 25 2500
scale values can be costs; sales;
computed market shares 5. Pizza Hut 1 10 7 17 75 7500
3
Scaling Comparative scales involve the direct comparison of
stimulus objects. Comparative scale data must be
Techniques interpreted in relative terms and have only ordinal or
rank order properties.
Comparative Noncomparative
Scales Scales
In noncomparative scales, each object is scaled
independently of the others in the stimulus set. The
Objects are resulting data are generally assumed to be interval
Objects are
Evaluated
Relative to Evaluated scaled.
Other Objects Independently
Comparative scaling is also referred to as nonmetric
Paired Rank Constant
scaling. Noncomparative scaling is also called metric
Comparison Order Sum scaling.
Instructions
 In paired comparison scaling, a respondent is
We are going to present you with ten pairs of shampoo brands. For each pair, please
presented with two objects and asked to select one indicate which one of the two brands of shampoo in the pair you would prefer for
according to some criterion. personal use.
Recording Form
 The data obtained are ordinal in nature.
Jhirmack Finesse Vidal Sassoon Head & Shoulders Pert
 Paired comparison scaling is a widely used comparative Jhirmack 0 0 1 0

Finesse 1A 0 1 0
scaling technique. Vidal Sassoon 1 1 1 1
Head & Shoulders 0 0 0 0
 Paired comparison scaling is useful when the number of
Pert 1 1 0 1
brands is limited, but with a large number of brands, the Number of times preferred 3B 2 0 4 1
number of comparisons becomes unwieldy. A A 1 in a particular box means that the brand in that column was preferred over the
brand in the corresponding row. A 0 means that the row brand was preferred over the column brand.
B The number of times a brand was preferred is obtained by summing the 1s in each column.
Instructions
 In rank order scaling, respondents are presented with Rank the various brands of toothpaste in Brand Rank Order
several objects simultaneously and asked to order or order of preference. Begin by picking out 1. Crest
the one brand that you like most and
rank them according to some criterion. 2. Colgate
assign it a number 1.
Then find the second most preferred-brand 3. Aim
 Possible that the respondent may dislike the brand
and assign it a number 2. 4. Mentadent
ranked 1 in an absolute sense.
Continue this procedure until you have 5. Macleans
 Furthermore, rank order scaling also results in ordinal ranked all the brands of toothpaste in order
of preference. 6. Ultra Brite
data.
The least preferred brand should be 7. Close Up
assigned a rank of 10. No two brands 8. Pepsodent
should receive the same rank number. The
Distances or differences between ranks are not criteria of preference is entirely up to you. 9. Plus White
defined, i.e., not meaningful in a rank order scale. There is no right or wrong answer— 10. Stripe
Just try to be consistent.
4
Instructions
 In constant sum scaling, respondents allocate a Below are eight attributes of bathing soaps. Please allocate 100 points among
constant sum of units, such as 100 points, to attributes of the attributes so that your allocation reflects the relative importance you attach
a product to reflect their importance. to each attribute. The more points an attribute receives, the more important
the attribute is. If an attribute is not at all important, assign it zero points. If an
 If an attribute is unimportant, the respondent assigns it attribute is twice as important as some other attribute, it should receive twice
as many points.
zero points.
Form AVERAGE RESPONSE OF 3 SEGMENTS
 If an attribute is twice as important as some other Attributes Segment I Segment II Segment III
1. Mildness 8 2 4
attribute, it receives twice as many points. 2. Lather 2 4 17
3. Shrinkage 3 9 7
 The sum of all the points is 100. Hence, the name of the 4. Price 53 17 9
scale. 5. Fragrance 9 0 19
6. Packaging 7 5 9
7. Moisturising 5 3 20
8. Cleaning Power 13 60 15
SUM 100 100 100
Copyright © 2015 Pearson Education Chapter 7 - 25 Copyright © 2014 Pearson Education,

Copyright
Inc. ©
Publishing
2015 Pearson
as Prentice
Education
Hall Chapter 7 - 26
 In comparative scales, small differences between  Comparative scales result in ordinal data.
stimulus objects can be detected.
 There is inability to generalize beyond the stimulus
 Same known reference points for all respondents. objects scaled.
 Easily understood and applied.
 Involve fewer theoretical assumptions.
 Tend to reduce halo or carryover effects from one

judgment to another.
 In noncomparative scaling respondents evaluate only

one object at a time, and for this reason noncomparative Noncomparative
scales are often referred to as monadic scales. Scales
 Noncomparative techniques consist of continuous and

itemized rating scales.
Continuous Itemized
In noncomparative scaling, respondents employ Rating Scales Rating Scales
whatever rating standards seem appropriate to them.
Likert
Semantic
Differential Stapel
5
In a continuous rating scale, respondents rate the objects by
placing a mark at the appropriate position on a line that runs from  In itemized rating scales, the respondents are provided
one extreme of the criterion variable to the other. The form of the
continuous scale may vary considerably. with a scale that has a number or brief description
associated with each category.
How would you rate Sears as a department store?
Version 1  The categories are ordered in terms of scale position,
Probably the worst - - - - - - -I - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Probably the best and the respondents are required to select the specified
Version 2 category that best describes the object being rated.
Probably the worst - - - - - - -I - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -- - Probably the best
0 10 20 30 40 50 60 70 80 90 100  The commonly used itemized rating scales are the Likert,
Version 3 semantic differential, and Stapel scales.
Very bad Neither good Very good
nor bad
Probably the worst - - - - - - -I - - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - -Probably the best Itemized rating scales are widely used in marketing research
0 10 20 30 40 50 60 70 80 90 100 and form the basic components of multi-item scales.
The Likert scale requires respondents to indicate a degree of agreement

or disagreement with each of a series of statements about the stimulus
objects.
Strongly Disagree Neither Agree Strongly

disagree agree nor agree
disagree
1. Sears sells high quality merchandise. 1 2X 3 4 5
2. Sears has poor in-store service. 1 2X 3 4 5
3. I like to shop at Sears. 1 2 3X 4 5
 The analysis can be conducted on an item-by-item basis (profile

analysis), or a total (summated) score can be calculated.
 When arriving at a total score, the categories assigned to the

negative statements should be scored by reversing the scale.
The semantic differential is a seven-point rating scale with end points The Stapel scale is a unipolar rating scale with ten categories
associated with bipolar labels that have semantic meaning. numbered from -5 to +5, without a neutral point (zero). This scale is
usually presented vertically.
SEARS
SEARS is: +5 +5
Powerful --:--:--:--:-X-:--:--: Weak +4 +4
Unreliable --:--:--:--:--:-X-:--: Reliable +3 +3
Modern --:--:--:--:--:--:-X-: Old-fashioned +2 +2x
+1 +1
 The negative adjective or phrase sometimes appears at the left High Quality Poor Service
side of the scale and sometimes at the right. -1 -1
-2 -2
 This controls the tendency of some respondents, particularly those
-3 -3
with very positive or very negative attitudes, to mark the right- or
-4x -4
left-hand sides without reading the labels.
-5 -5
 Individual items on a semantic differential scale may be scored on  The data obtained by using a Stapel scale can be analyzed in the
either a -3 to +3 or a 1 to 7 scale. same way as semantic differential data.
6
Basic
Scale Examples Advantages Disadvantages
Characteristics
Continuous Place a mark on a Reaction to TV Easy to Scoring can be
Rating Scale continuous line commercials construct cumbersome
unless
computerized
Itemized Rating Scales
Likert Scale Degree of agreement Measurement of Easy to More time

on a 1 (strongly attitudes construct, consuming
disagree) to 5 administer, and
(strongly agree) scale understand
Semantic Seven-point scale with Brand, product, Versatile Difficult to cons-
Differential bipolar labels and company truct bipolar
images adjectives
Stapel Scale Unipolar ten-point Measurement of Easy to Confusing and

scale, -5 to +5, without attitudes and construct and difficult to apply
a neutral point (zero) images administer over
telephone
 Noncomparative itemized rating scales need not be

used as originally proposed but can take many different
forms.
Balanced Scale Unbalanced Scale
Surfing the Internet is Surfing the Internet is
 For example, the number of categories in a Likert scale
need not be 5, but could be 7. ____ Extremely Good ____ Extremely Good
____ Very Good ____ Very Good
 The number of categories could be odd or even.
____ Good ____ Good
 A Categories can be expressed by boxes, discrete lines, ____ Bad ____ Somewhat Good
or units, with or without numbers. ____ Very Bad ____ Bad
____ Extremely Bad ____ Very Bad
A variety of scale configurations may be employed to measure the comfort of Nike shoes. Some examples
include: 4)
Nike shoes are: Very Uncomfortable Somewhat Somewhat Comfortable Very

Neither
1) Place an “X” on one of the blank spaces… Uncomfortable Uncomfortable Comfortable Comfortable
Comfortable nor
Very Very
Uncomfortable
Uncomfortable Comfortable
2) Circle the number…

Very 1 2 3 4 5 6 7 Very
Uncomfortable Comfortable 5)
3) Place an “X” on one of the blank spaces…

Very Uncomfortable -3 -2 -1 0 1 2 3
Uncomfortable
Very Neither Very
Neither Uncomfortable nor Comfortable
Uncomfortable Comfortable nor Comfortable
Comfortable
Uncomfortable
Very Comfortable
7
Thermometer Scale
Instructions: Please indicate how much you like Smiling Face Scale
McDonald’s hamburgers by coloring in the thermometer. Instructions: Please point to the face that shows how
Start at the bottom and color up to the temperature level much you like the Barbie Doll. If you do not like the
that best indicates how strong your preference is. Barbie Doll at all, you would point to Face 1. If you liked
it very much, you would point to Face 5.
Form:
100 Like very much Form:
75
50
25 1 2 3 4 5
0 Dislike very much
Construct Scale Descriptors
Attitude Very Bad, Bad, Neither Bad nor Good, Good,

Very Good
Importance Not at All Important, Not Important, Neutral,
Important, Very Important
Satisfaction Very Dissatisfied, Dissatisfied, Neither

Dissatisfied nor Satisfied, Satisfied, Very
Satisfied
Purchase Frequency Never, Rarely, Sometimes, Often, Very Often
Scale  Reliability can be defined as the extent to which

measures are free from random error.
Evaluation
Reliability Validity
8
 In test-retest reliability, respondents are administered  The validity of a scale may be defined as the extent to
identical sets of scale items at two different times and the which differences in observed scale scores reflect true
degree of similarity between the two measurements is
determined. differences among objects on the characteristic being
measured, rather than systematic or random error.
 Internal consistency reliability determines the extent to
which different parts of a summated scale are consistent in  Perfect validity requires that there be no measurement
what they indicate about the characteristic being measured. error. That is, random error must be zero and systematic
 In split-half reliability, the items on the scale are divided into error must be zero.
two halves and the resulting half scores are correlated.
 The coefficient alpha, or Cronbach's alpha, is the average of

all possible split-half coefficients resulting from different ways
of splitting the scale items. This coefficient varies from 0 to 1,
and a value of 0.6 or less generally indicates unsatisfactory
internal consistency reliability.
 Validity consists of content validity, criterion validity, and  If a measure is perfectly valid, it is also perfectly reliable.
construct validity. In this case, there is no random or systematic error.
 Content validity is a subjective but systematic  If a measure is unreliable, it cannot be perfectly valid,
evaluation of how well the content of a scale represents since at a minimum random error is present. Thus,
the measurement task at hand. unreliability implies invalidity.
 Criterion validity reflects whether a scale performs as  If a measure is perfectly reliable, it may or may not be
expected in relation to other variables selected (criterion perfectly valid, because systematic error may still be
variables) as meaningful criteria. present.
 Construct validity addresses the question of what  Reliability is a necessary, but not sufficient, condition for
construct or characteristic the scale is, in fact, measuring. validity.
Step 1 : Marketing Research Problem Definition  In developing countries, the respondents might have
difficulty using interval and ratio scales. consumer
Step 2 : Approach to the Problem
• Specification of information needed
preferences in these countries are best measured with
ordinal scales.
Step 3 : Research Design
•Appropriate level of measurement and appropriate scales to measure each item of  The primary scales should be matched to the profile of
information
•Questionnaire Design: translation of the information needed to
the target respondents.
appropriate questions using the identified scales
Step 5 : Data Preparation and Analysis:

•Using appropriate statistical techniques compatible with the level of measurement
of the data
9
 Pan-cultural scales, designed to be free of cultural  Of the scaling techniques considered, the semantic
biases, are used in international research. differential has been applied with the greatest
consistency in results across countries.
 An approach to developing pan-cultural scales is to use
descriptors the respondents create themselves. The end  Rating scales also are used to construct indexes used to
descriptors used to anchor the scale are particularly make global comparisons.
prone to different interpretations.
In some cultures 1 may be interpreted as best,
 Additionally, the scale numbering may have different while in others it may be interpreted as worst,
meanings. In such cases, it might be desirable to avoid regardless of how it is scaled. Scale end points
numbers and to just use boxes that respondent can and the verbal descriptors should be used in a
check. manner that is consistent with the culture.
 All the primary scales and all the comparative scales can  All the noncomparative scales can be easily
be easily implemented in social media. implemented in social media.
 An analysis of social media content can shed light on the  An analysis of social media content can provide
level of measurement that is appropriate in a given guidance on whether continuous or itemized rating
project and provide guidance on the type of scaling scales should be used. It can also aid the researcher in
techniques to use. making appropriate rating scale decisions.
 Specific measures have been developed to evaluate  It is feasible to employ more than one scaling method to
social media sites based on information that is publicly measure a given construct.
available, e.g., number of friends or followers, number of
comments, and media citations.
 Knowingly using inappropriate scales raises ethical  The researcher should establish the reliability and
questions. validity of scales.
 It is the obligation of the researcher to obtain the data  The researcher has a responsibility to both the client and
that are most appropriate given the research questions. respondents to ensure the applicability and usefulness of
the scales.
 The researcher’s should not deliberately bias the results
by building that bias into noncomparative scales.
The researcher should not bias the scale by biasing
the wording of the statements, the scale descriptors,
or other aspects of the scale.
10
The four primary types of scales can be described by the The different comparative and noncomparative scales can
acronym FOUR: be represented by the acronym SCALES:
F igurative: nominal scale S emantic differential scale
O rdinal scale C onstant sum scale
U nconstrained zero point: interval scale A rranged in order: rank order scale
R atio scale L ikert scale
E ngaged: paired comparison scale
S tapel scale
11

Ch07 Measurement and Scaling

Uploaded by

Copyright:

Available Formats

You might also like

Ch07 Measurement and Scaling

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ch07 Measurement and Scaling

Uploaded by

Copyright:

Available Formats

After reading this chapter, the student should be able to:

1. Introduce the concepts of measurement and scaling

2. Discuss the primary scales of measurement and

3. Classify scaling techniques as comparative and

4. Describe the noncomparative scaling techniques,

Step 3: Formulating a Research Design

• Questionnaire & Form Design

Step 6: Preparing & Presenting the Report

Copyright © 2015 Pearson Education Chapter 7 - 10

Primary Scales  In a nominal scale, the numbers serve only as labels or

 It permits comparison of the differences between objects.

 It is not meaningful to take ratios of scale values.

 A ratio scale possesses all the properties of the nominal, Scale

 It has an absolute zero point.

Interval Differences between objects Temperature Attitudes; 2. McDonald’s 2 25 7 17 50 5000

 Paired comparison scaling is a widely used comparative Jhirmack 0 0 1 0

Copyright © 2015 Pearson Education Chapter 7 - 25 Copyright © 2014 Pearson Education,

 Involve fewer theoretical assumptions.

 Tend to reduce halo or carryover effects from one

 In noncomparative scaling respondents evaluate only

 Noncomparative techniques consist of continuous and

The Likert scale requires respondents to indicate a degree of agreement

Strongly Disagree Neither Agree Strongly

1. Sears sells high quality merchandise. 1 2X 3 4 5

2. Sears has poor in-store service. 1 2X 3 4 5

3. I like to shop at Sears. 1 2 3X 4 5

 The analysis can be conducted on an item-by-item basis (profile

 When arriving at a total score, the categories assigned to the

Likert Scale Degree of agreement Measurement of Easy to More time

Stapel Scale Unipolar ten-point Measurement of Easy to Confusing and

 Noncomparative itemized rating scales need not be

Nike shoes are: Very Uncomfortable Somewhat Somewhat Comfortable Very

2) Circle the number…

3) Place an “X” on one of the blank spaces…

Construct Scale Descriptors

Attitude Very Bad, Bad, Neither Bad nor Good, Good,

Satisfaction Very Dissatisfied, Dissatisfied, Neither

Purchase Frequency Never, Rarely, Sometimes, Often, Very Often

Scale  Reliability can be defined as the extent to which

 The coefficient alpha, or Cronbach's alpha, is the average of

Step 5 : Data Preparation and Analysis:

You might also like