CHAPTER 9

MEASUREMENT AND SCALING:
FUNDAMENTALS AND COMPARATIVE SCALING
OPENING QUESTIONS
1.
2.
3.
4.
5.
6.
7.

What is meant by measurement and scaling? Can scaling be considered a part of
measurement?
What are the primary scales of measurement, and how do we differentiate among them?
How can scaling techniques be classified, and what are the various comparative scaling
techniques?
How do measurement and scaling relate to the various steps of the marketing research
process?
What considerations are involved in implementing the primary scales of measurement in an
international setting?
How does technology improve measurement and scaling?
What ethical issues are involved in selecting scales of measurement?

AUTHOR’S NOTES: CHAPTER FOCUS
This chapter provides an introduction to primary scales of measurement: nominal, ordinal,
interval, and ratio. Scaling techniques are classified as comparative and noncomparative. The
comparative techniques consisting of paired comparison, rank order, constant sum, and Q-sort
scaling are discussed. The noncomparative scales are discussed in Chapter 10.
The discussion of scaling techniques is much more extensive as compared to competing
texts. Two chapters are devoted to covering in detail the type of scales used in marketing
research. Examples, including instructions for administering these scales, have been provided.
This chapter could be taught by focusing on the opening questions, sequentially. Greater
emphasis could be placed on the concept of measurement and scaling (Opening Question 1), the
primary scales of measurement (Opening Question 2), and comparative scaling techniques
(Opening Question 3). In addition, material on conducting international marketing research
(Opening Question 5), ethical considerations (Opening Question 7), and Internet and computers
could be discussed.
CHAPTER OUTLINE
1. Overview
2. Measurement and Scaling
3. Primary Scales of Measurement
i.
Nominal Scale
ii.
Ordinal Scale
iii. Interval Scale
iv. Ratio Scale
4. A Classification of Scaling Techniques
5. Comparative Scaling Techniques
i.
Paired Comparison Scaling
ii.
Rank Order Scaling
iii. Constant Sum Scaling
116

1. 4. Opening Question 3 117 . and compare intervals or differences--for example. and income. rather. 9. 10. 13. they are used only for identification--for example. Opening Question 2 * Discuss and illustrate the primary scales of measurement. 11.4 and Table 9. 8. age. Interval scale: numbers are used to rank objects such that numerically equal distances on the scale represent equal distances in the characteristic being measured. height. or social security numbers. Distinguish the two concepts by noting that measurement precedes scaling in test construction. street names.2 provide a framework and examples for explaining scaling. rankings of teams for the NCAA Basketball tournament. The numbers in a nominal scale do not reflect the amount of a characteristic possessed by the objects. Examples include time and temperature. 7. 12. 3.6. Ratio scale: this is used to identify or classify objects. Ordinal scale: this is a ranking scale in which numbers are assigned to objects to indicate the relative extent to which some characteristic is possessed. Measurement is the assignment of numbers or other symbols to characteristics of objects according to certain prespecified rules. Nominal scale: this is used only as a labeling scheme where numbers serve only as labels or tags for identifying and classifying objects. socioeconomic status. 2. and quality rankings. Figure 9. It is then possible to determine whether an object has more or less of a characteristic than some other object--for example. numbers on baseball players’ uniforms. Scaling is an extension of measurement where it involves the generation of a continuum upon which measured objects are located. rank order the objects. Relationship of Measurement and Scaling to the Marketing Research Process Summary Illustration Using the Opening Vignette International Marketing Research Technology and Marketing Research Ethics in Marketing Research Summary Key Terms and Concepts Acronyms TEACHING SUGGESTIONS Opening Question 1 * Explain the differences between measurement and scaling.

3. The researcher must identify an appropriate level of measurement (nominal.9. or Stapel). 118 . Discuss the relationship of measurement and scaling to the marketing research The relationship of measurement and scaling to the previous and subsequent steps of the marketing research process is described in Figure 9. If available. an approach to the problem is developed (step 2).6 for an example of paired comparison scaling.8 for an example. If the measurement level is ordinal. The data obtained is ordinal in nature. semantic differential. Opening Question 4 * process. This should be highlighted as each of the scales is discussed in turn. bring examples of different scales to class to show to students. Noncomparative scales . Paired comparison scaling: here a respondent is presented with two objects at a time and asked to select one object in the pair according to some criterion.a direct comparison of stimulus objects is elicited. Comparative scales . 2. two brands may be compared along a dimension such as quality. See Figure 9. Based on this definition. See Figure 9. or constant sum). This is frequently used in marketing when comparisons of products or brands are being made. interval or ratio) for each item of information needed. See Figure 9. chits. one brand is rated on a scale independent of other brands. The marketing research problem is defined in step 1. rank order. See Figure 9.7 for an example of rank order scaling. Measurement and scaling are part of the research design (step 3). the researcher generally selects one of the comparative techniques (paired comparison.the respondent provides whatever standard seems appropriate to him/her. * Describe the different comparative scaling techniques. If the data are interval. If an attribute is twice as important as some other attribute it should receive twice as many points. it is possible to assign zero points. ordinal.5 for the hierarchy of scaling procedures. Begin by recalling that all comparative scaling techniques involve a direct comparison of stimulus objects with one another. 1. the researcher selects one of the noncomparative techniques (continuous or itemized rating scale: Likert. dollars. stickers. Constant sum scaling: respondents are required to allocate a constant sum of units such as points. or chips among a set of stimulus objects with respect to some criterion. only one object is evaluated at a time.* Distinguish the two broad scaling measures. Specific instructions are provided that if an attribute is not at all important. This is commonly used to measure preferences for brands as well as the importance of attributes. thus. Begin by stating the two types of scales: comparative and noncomparative. In this case. Thus. Rank order scaling: respondents are presented with several objects simultaneously and asked to order or rank them according to some criterion.

g. could ask consumers to rate their preferences for wearing jeans on specified occasions using a seven point interval scale. they should be analyzed correctly. After the data have been collected. these data cannot be easily used in multivariate analysis. The researcher has the responsibility to use the appropriate type of scales to get the data needed to answer the research questions and test the hypotheses. the use of binary scales (e. consumers in Papua New Guinea could be shown a pair of jeans and simply asked whether or not they would prefer to wear it for a specific occasion (e. Opening Question 6 * Discuss how computers and software can be used to implement comparative scaling. without compromising test reliability.g. Database managers allow researchers to develop and test several different scales to determine their appropriateness for a particular application.). However. and ratio scales to determine consumer preferences among competing alternatives. In particular. are quite used to providing responses on interval and ratio scales. For example. preferred/not preferred) is recommended. To examine differences in the personality characteristics and relate them to other consumer behavior variables. if one product is clearly winning.. interval scale data are needed. Pulse/MPC by Pulse Analytics enables multiple paired comparison analysis. It allows testing to end early. Conclusions based on the misuse of statistics are misleading. however. constant sum. when shopping. Up to 30 brands or variables can be analyzed simultaneously. whereas ratio scales are the most complex. working. Computers can also be used to construct and administer rank order. When designing a questionnaire. such is not the case in less developed countries. EzPair by Barry Cohen can design paired comparison scale and paired comparison product tests using statistical quality control techniques.Likewise. Preferences can. If ordinal scaled data are collected. Respondents in many developed countries. the researcher must translate the information needed to appropriate questions using the identified scales. therefore. the researcher must select an appropriate scale for an information item that is to be measured on a nominal or ordinal level. the researcher should only use those statistical techniques that are compatible with the measurement level of the data. due to higher education and consumer sophistication levels. When analyzing the data (step 5). if personality characteristics are measured using ordinal scales. relaxing on a holiday. etc. be best measured by using ordinal scales in less developed countries. also part of the research design. Opening Question 7 * Discuss the ethical concerns of scaling. nominal scales are the simplest to use. Using the personality 119 . Levi Strauss & Co. statistical procedures developed for use with interval or ratio data should not be used.. Opening Question 5 * Identify the measurement and scaling issues in International research. It projects pairwise comparison data onto a share-of-preference scale. EXAMPLE: While measuring preferences for jeans in the United States. From the view point of the respondents.

comfort. taste. It allows testing to end early. and touch comparisons are difficult to implement. as they are the biggest consumers of the products offered by Home Depot. BE AN MR! AND BE A DM! It should be noted that a variety of answers are appropriate. 120 . Microcomputers have been used to administer paired comparison scales in taste tests. Various characteristics need to be measured to get a holistic picture. if after data collection the client wishes to know how the users and nonusers differed. When the researcher lacks the expertise or the computer software to compute these statistics. such as local newspapers.  Use local media. the researcher should treat these data correctly and use nonmetric techniques for analysis (discussed in Chapter 15). brand name.  Target families. An initial study needs to be conducted to identify the characteristics that need to be measured based on their relevance to customer satisfaction. Database managers allow researchers to develop and test several different scales to determine their appropriateness for a particular application. Ordinal scales can be used to rank order the stores based on customer satisfaction. All the primary scales of measurement can be implemented on the Internet. smell. EzPair by Barry Cohen can design paired comparison scales and paired comparison product tests using statistical quality control techniques.example above. ethical dilemmas arise. Be an MR!: Home Depot Nominal scales can be used to assign numbers to the stores for identification purposes. Be an MR!: Coach Consumer preferences for leather good are determined by factors such as finish. Internet Emphasis * Identify software programs relevant to measurement and scaling. if one product is clearly winning. to advertise. where 1 implies very poor and 5 implies excellence. durability. however. Ratio scales that measure the amount of money spent on leather goods during a specified time period can also be used. The use of intervals would be a good idea since preferences can be rated from 1 to 5. Be a DM!: Home Depot Marketing Strategies:  Seasonal Promotions  Use the power of Internet to reach out to customers. price. etc. without compromising test reliability. Either an outside statistician should be hired or the relevant software should be obtained. visual. The same is true for the commonly used comparative scales. Paired comparisons involving verbal. The ones given here are merely illustrative. The process of implementing comparative scales may be facilitated by searching the Internet for similar scales that have been implemented by other researchers. or auditory comparisons can be implemented with ease.

Price versus quality tradeoff can be used for segmentation. Respondents in the United States are sophisticated enough to make such comparisons and provide such information.). Demographic variations in preferences.  Identify the values and qualities that are relevant for the type of styles. 3.  Understand driving forces behind purchase decisions. Be an MR!: Gap Students should be encouraged to make the online search. however. formal wear. It just gives a preference order.  Identify the values and qualities that are relevant for the type of use. Be a DM!: Lexus  Study the way customers intend to use the cars.  Input for the range of products to be offered (to cater to various segments). the simple form of paired comparison should be used where the respondents merely indicate which style of clothing in a pair they prefer. In rural Nigeria. their relative values and order. so response rates might be better. Identify which qualities are valued by consumers.Be a DM!: Coach 1. 121 . Rank order scaling should be used in the United States where casual clothing is compared to other styles of clothing (e.  Demographic variations in intentions and interests.  Understand the target segments better. Find qualities associated with most successful products. Constant sum scaling gives a precise and holistic picture of preferences. 2. Be a DM!: Gap  Study the way customers intend to wear the clothes of different styles. Rank order scaling gives a relative comparison but fails to reveal the value of each quality.g.. Be an MR!: Lexus Purchase Intentions for luxury car are determined by factors such as: Social Status Comfort Daily Commute Convenience Security     Rank Order Scaling or constant sum scaling are two comparative scaling techniques which could be used. 4.product fit. etc.  Analyze segment. part wear. Rank order scaling is easier for the customer to fill in.

Responses would likely vary across classes. how many had no preference between Coke and Pepsi? Answer: About 20 %. EXPERIENTIAL LEARNING Coke or Pepsi? This exercise is designed to give students experience in conducting an actual taste test. Demographic variations in preference for different styles. 2. This will be repeated for each person when tasting from the second. Of those who correctly identified the two identical samples of cola. How many correctly identified the two identical samples of cola? Answer: 50 per cent of the class. and took four Dixie cups back to their seats containing the sample for Pepsi. Which cola drink would be the most popular among your friends in a taste test? Using the Actual Taste Test for the KitKat candy bar as a guide. and take a swallow of water. Make sure the drinks are chilled to the same temperature. After tasting from each of the three cola sample groups. taste the cola. also. Note the respondents loyal to Coke were seeking this taste because they knew Coke was one of the tested samples. votes were taken as to which colored cup corresponded to which cola brand. develop a script for testing for preference between two cola drinks: Coke and Pepsi. The following answers occurred in an actual taste test conducted in a marketing research class in October 2004. 1. and the third group Pepsi). In sum. and the third groups. the second group Coke. (The first group might be Pepsi. Understand the target segments better. The answers will likely vary from student to student or group to group. and poured in three groups of identical cups. which cola was preferred . it will be important for half of the respondents to have a rotated order of presentation: (The first group Coke. have the respondents complete the survey form you created similar to that in the KitKat piece included in this chapter. 3. Each person will select a cup from the first group. the redundant sample of Pepsi or Coke. 4. Coke. the second group Pepsi. To avoid a biasing effect from the order of presentation of the samples. An execution technique that worked well was this: using different colored board markers to put the same colored dot on Dixie cups holding the same contents. and the third group Coke). The students should download the KitKat Taste Test from the Web site of this textbook and use it as a guide to develop a script for testing for preference between two cola drinks. eat a cracker. Of those who correctly identified the two identical samples of cola. Every person in the class then lined up. who would you say is the winner in your taste test: Coke or Pepsi? Or is it too close to tell? 122 . Then. and bottled water.   Understand driving forces behind preference for different styles.Coke or Pepsi? Answer: Coke by a 2:1 margin.

because loyal Coke drinkers were actively seeking Coke. (a) Spreadsheet Example Example Question Preview Visa MasterCard American Express Please indicate the number of cards you own : Please indicate your account credit limit : Please indicate your current account balance : (b) Select to Add a Question: Single Multiple Answers Which credit card do you most often use ? Visa MasterCard American Express Discover Card (c) Side-by-side Matrix 123 Discover Card . not disclosing which brands would be sampled. In hindsight.Answer: Coke. and then having respondents take the test would likely bring different results. Surveyz! Question Types and Corresponding Scales Part I.

Types of question formats available using Surveyz. which include users. 4 . etc.com: a) Analyses using criteria to specific subgroups. b) Cross tabulation of two variables at time that allows for the examination of two variables 124 .Highest.Lowest): Visa MasterCard American Express Discover Card Part II 1. demographic groups. Multiple Choice Multiple Choice Battery Open Ended Text Rank Order Constant Sum Text only (no response) Types of analyses available using Surveyz.com: a) b) c) d) e) f) 2.Please evaluate our performance on When purchasing a product how important are each of the following attributes: Very the following attributes: Very Somewhat Somewhat Excellent Good Good Fair Poor Important Important Unimportant Unimportant Easy to use Comfortable Reasonably Priced (d) Rank Order Rank the following credit cards in order of preference (1 .

125 . The ones given here are merely illustrative. D 2. The Marketing Management Decision 1. harmonic mean. consumer priorities and perceptions. The Marketing Research Decision 1. While both methods will give a relative order of brand preferences. however. Hence. while in an ordinal scale the numbers are used as a ranking device. Although both nominal and ordinal scale data can be used for counting operations. and coefficient of variation can be applied to analyze ratio scale data. Command files are also prepared and exported. 4. they need to launch newer models (more colorful) to attract the young men. Rank order provides an easy comparison of brand preferences. B. the numbers serve only as labels or tags for identifying and classifying objects. They should be aiming youngsters who are serious about athletic shoes and affluent customers with a greater propensity to spend. A. the constant sum method has the advantage that it provides a measure of the degree of preference. 3. Marketing decisions can be taken based on these facts.c) Advanced reports which contain complete summary reports with tables and charts 3. The primary scales of measurement are nominal. and ratio. it is not meaningful to take ratios of scale values. C. 2. While they are successful with the affluent group. In this case. Measurement is the assignment of numbers or other symbols to characteristics of objects according to certain prespecified rules. WHAT WOULD YOU DO? It should be noted that a variety of answers are appropriate. Celebrity (such as Nike’s NBA campaign) endorsements and increased promotional spending should be able to win the attention of this segment. ordinal. To get an idea of the absolute preference. it is meaningful to take ratios of scale values. The advantages of a ratio scale over an interval scale is that the origin is fixed. the magnitude of difference in preference between brands will not be obvious using these methods. 5. Statistics such as the geometric mean. Scaling techniques can reveal brand preferences. C and E 2. The implications of having an arbitrary zero point in an interval scale means that any positive linear transformation of the form y = a + bx will preserve the properties of the scale. The differences between a nominal and an ordinal scale are that in nominal scales. New Balance is a premium brand aiming the high end market. itemized rating scales should be used. ordinal scales permit the use of statistics based on centiles. REVIEW QUESTIONS 1. Hence. Types of formats in which results can be exported? Results may be saved into an Excel file in preparation for importing into SPSS. interval.

126 . a respondent is presented with two objects at a time and asked to select one object in the pair according to some criterion. In paired comparison scaling. 7. or chips among a set of stimulus objects with respect to some criterion. 9.6. the order in which the objects are presented may bias the results. Thus. stickers. and they bear little resemblance to the marketplace situation involving multiple alternatives. violations of the assumption of transitivity may occur. 8. A comparative rating scale involves the direct comparison of stimulus objects with one another. The constant sum is a more refined ranking scale in that it allows fine discrimination among stimulus objects without requiring too much time. In a constant sum scale. dollars. the respondents are required to allocate a constant sum of units such as points. Paired comparison scaling is useful when the number of brands is limited because it requires direct comparison of brands. chits. the disadvantages of paired comparison scaling is that with a large number of brands the number of comparisons become unwieldy.

The URLs of several marketing research firms are given on the web site for the book. Interval . Pepper ____ 4. 2. 7-Up ____ Interval: How likely are you to buy each of the following four brands of soft drinks when you next buy soft drinks? Not at all Very Likely Likely 1. 4.walmart. the higher the scale response. Ordinal: Please rank order the following four brands of soft drinks in terms of your intentions to purchase them when you next buy soft drinks. Ordinal . Coke 1 2 3 4 5 6 7 2. Pepsi ____ 3. Ratio . Brand Rank 1. The students should be encouraged to find examples of the four primary scales discussed in this chapter. It is important that the students relate the scales to the marketing research context (problem) and the managerial context (management decision problem). Ratio . This exercise is designed to give the students first hand experience in constructing an actual ranking scale. The Target and Wal-Mart Web sites can be visited at www.targetstores.The scale represents increasing and uniform degrees of agreement. f. The brand that you are most likely to purchase should be assigned a rank 1 and the brand you are least likely to purchase a rank 4. 3. 2. Pepsi ____ 3. Coke ____ 2.com. Dr. Brand $ Amount 1. a. Pepper ____ 4. 7-Up ____ INTERNET AND COMPUTER EXERCISES 1. b. Coke ____ 2.The more minutes spent studying. The characteristics identified should be similar to those given in Chapter 2.An absolute 0 is assumed. c. Pepper 1 2 3 4 5 6 7 4.An absolute 0 is assumed.APPLIED PROBLEMS 1. 7-Up 1 2 3 4 5 6 7 Ratio : Please indicate the amount of money you are likely to spend on the following four brands of soft drinks when you next buy soft drinks. Dr. d. Ordinal. in the section on Relevant Characteristics: quality of merchandise. Pepsi 1 2 3 4 5 6 7 3.A simple matching has occurred.Rankings display a relationship of order between the scale items. e. Dr. variety and 127 . Nominal .com and www.

yet economic or marketing considerations may negatively impact sales. prices. Pertinent to this discussion is the issue of what is being measured. in the auto market consider a Porsche. Products can be perceived very positively. 2.g. however. Another issue is that none of the brands considered may be desirable. service of store personnel. If we are measuring attitude toward the brand. Hence. layout of store. it can be very highly rated. One paired comparison is needed for each characteristic. The discussion should be related to one or more of the concepts discussed in this chapter.. convenience of location. even the top ranked brand may not be desirable. Attitudes and intentions do not perfectly correlate with behavior because there are additional factors affecting behavior. returns and adjustment policy. Any one of readings can be selected. 128 . a high price may be charged which eliminates most buyers.assortment of merchandise. GROUP DISCUSSIONS 1. e. and credit and billing policies.