An Empirical Investigation of the Validity of Selected Attitude and Activity Measures



Citation:

Albert V. Bruno and Edgar A. Pessemier (1972) ,"An Empirical Investigation of the Validity of Selected Attitude and Activity Measures", in SV - Proceedings of the Third Annual Conference of the Association for Consumer Research, eds. M. Venkatesan, Chicago, IL : Association for Consumer Research, Pages: 456-474.

Proceedings of the Third Annual Conference of the Association for Consumer Research, 1972      Pages 456-474

AN EMPIRICAL INVESTIGATION OF THE VALIDITY OF SELECTED ATTITUDE AND ACTIVITY MEASURES

Albert V. Bruno, University of Santa Clara

Edgar A. Pessemier, Purdue University

All investigators, particularly those working in the social sciences, should be concerned about measurement problems. Prior to data collection, judgment must be exercised in the design or selection of instruments. After data has been collected, the quality of the data must be appraised. In exercising prior judgments and examining the quality of collected data, the researcher is interested in reliability (accurate, consistent data) and validity (data that measures what it is designed to measure).

Historically, marketing researchers have not demonstrated great concern about the reliability or validity of the instruments which they have employed. Cost considerations, time limitations or the nature of the applied work usually have prevented practitioners from adopting the traditions of careful measurement characteristics of basic social science research. On the other hand, academic investigators in marketing and consumer behavior have been scandalously short of the resources required to investigate the range of important topics related to reliability and validity. To fill a small part of this void, we recently reported on the reliability of a series of studies conducted by researchers affiliated with Purdue University (Bruno & Pessemier, 1971).

In this validity study, we examine a similar body of research. These studies measure many important life-style, "psychographic" components of consumers' activities and attitudes. Naturally, the decision to use studies generated largely by the "Purdue group" of investigators was motivated by the local availability of raw data, source documents and useful unpublished results.

THE VALIDITY OF MEASUREMENTS

The validity of measurements made by an instrument depends on the scientific or applied purpose(s) of the user. One researcher may be interested in testing a descriptive theory about the common properties of classes of consumers or enterprises; another researcher may be interested in measurements which predict selected types of market behavior. In each case, validity must be judged against some appropriate criterion. The criterion and associated methods for judging validity can be conveniently classified into four important types:

Construct validity: This form of validity must be examined in light of adequacy and the results of three steps: the extent to which

(1) theoretically satisfying constructs can account for test performance

(2) hypotheses have been developed that flow from the theory or construct(s) and

(3) the hypotheses are subjected to empirical test.

Content validity: The sampling adequacy or degree to which the test is representative of the construct or properties being measured.

Concurrent validity: "Immediate predictive" validity indicating the common variability of the test outcomes (measurements) and the one or more criterion variables. The passage of time and/or the order of causality are not necessary elements.

Example: One section of an instrument may be designed to measure product-specific innovativeness and another section may contain a previously employed test of product-specific information seeking and transmitting. Since prior theoretical considerations provide insights about the expected common variability of these two tests, concurrent validity of the new measure of product-specific innovativeness can be examined in part by the correlation of the test outcomes.

Predictive validity: The common variability of the test outcomes (measurements) and one or more criterion variables when the passage of time and/or order of causality are necessarily considered.

Example: Individuals' brand preferences are scaled to predict their subsequent choice of brands. The measurement of brand preferences and the observation of brand choices are made at different times and in a natural time order. Note, however, that the time order would be reversed if one were trying to predict preference from choice. Finally, if changes in preference resulting from brand choice and/or use were of interest, then two preference measures and one choice or use experience would be required. In all these cases, however, the researcher is concerned with predictive validity.

In this paper we will examine essentially one type of validity, concurrent validity. Our reasons for doing so are practical. The discussion of the construct validity of each set of measurements would represent an impossible task, one in which each researcher has a better chance of success individually than we could hope to achieve in summary form. The interested reader is encouraged to consult the primary sources and to judge construct validity for himself. Because the nature of the constructs is not easily summarized, content validity must also be by-passed in this paper. Since the exploratory context of these studies reduces interest in the problem of-content validity, the 1099 is small.

Predictive validity is not seriously treated in this discussion because of inadequate data. In some cases, investigators employ self-reported prior behavior or expected future behavior. These cross-sectional measures do not constitute "hard" tests of predictive validity, particularly for the classes of measures which are examined here. Therefore, we will focus on concurrent validity, accepting the exploratory nature of much of the research and looking across broad sets of criterion variables of potential interest to marketing scholars and managers.

DATA BASES

Recently, an evolving area of data collection and research activity has been concerned with market-related measures of consumers' activities and attitudes. Although the evolutionary time has been short, a large number of related life-style, psychographic studies have appeared. The sheer volume of this work makes it impractical to summarize results within the confines of this discussion. Interested readers should consult recent articles by Hustad and Pessemier (Hustad & Pessemier, 1971) and Kassarjian (Kassarjian, 1971) for a critical review of the subject.

At this point in the evolutionary process, it is not clear whether activity and attitude measures (more narrowly described as life-style or psychographic variables) will mature into well-developed, widely used bases for consumer and market analysis or will play a limited, specialized role. What is clear, however, is the need for more extensive validity and reliability research in order to allow this emerging area to develop its natural theoretical and applied stature. We believe that Kerlinger's remark ". . . all psychological and educational measuring instruments must be critically and empiricaLly examined for their reliability and stability" applies equally to marketing and consumer research and the validity of the instruments employed in these fields (Kerlinger, 1965).

SPECIFIC STUDIES AND MEASUREMENTS

Although concurrent validity in a marketing setting concerns practical issues and outcomes, difficult problems in validation are associated with the selection of criteria. For example, if a marketing researcher is interested in enlarging the psychological profile of potential triers of a new product, some dimension of new product purchasing behavior must be employed as the criterion. However, if a marketing researcher is interested in developing measures which are associated with determining the susceptibility of consumers to persuasive communication, he may find measuring latent persuadability for a class of communications is an imposing task. In this case, the validity of the criterion variable(s) may be doubtful.

In the set of activity and attitude studies discussed in this paper, we examined criteria which are reported market behavior or widely accepted attitudinal indices such as brand preferences. The variables tested against these criteria are more exploratory in nature; they are drawn largely from emerging specialized segments of the personality, life style and value domains. The reliability of selected sets of these activity and attitude measures have been examined. The wider, overlapping set of studies used as data sources for this paper are summarized in Table 1. Since the investigators who conducted the investigations have described their data collection methods, these details will not be discussed here.

Classifying the measures that appear in the last column of Table 1 is a complex, elusive problem. The taxonomic scheme shown in Table 2 is adapted from DeBruicker, 1971. The first row is concerned principally with measures of properties that are highly enduring such as personality and demographic characteristics, environmental conditions and some ownership profiles. The second and third rows are concerned with consumer predispositions to action subject to influence and change. These attitudes are separated into perceptual (cognitive) and preference (affective) elements. Finally, the fourth row is concerned solely with manifest behavior. The columns are arrayed along a continuum of properties ranging from the personal factors to the properties of individuals related specifically to products and brand interactions. The concurrent validity reported for variables of interest in each of the studies considered have been classified by the criteria utilized. In turn, these reported have been assigned to one or more of the cells in Table 3. The reader will note that, in several cases, the concurrent validation activities of each author or groups of authors encompass multiple categories.

ILLUSTRATIVE DETAILED FINDINGS

Since much of the material which follows deals with research presented in summary form, it is useful to present three small recently completed segments of analysis from the 1970 Purdue Consumer Behavior Research Project (1970 PCBRP) to illustrate the general problems of criteria selection and measurement which appear throughout activity and attitude studies.

In the 1970 PCBRP, a number of questions were concerned with aspects of self-confidence, desire for change, risk taking, the personal appeal of a new product, and similar topics. These attitudinal questions were factor analyzed. One of the extracted factors was labelled "New-product Proneness." Four principle questions had factor loadings greater than .55.

I like to try things just because they are new.

I really enjoy beating everyone to a new product.

I often look for new things to try so I can stay ahead of my friends and acquaintances.

I investigate new products to make sure I am always using the most efficient ones.

TABLE 1

SUMMARY OF CHARACTERISTICS OF STUDIES

TABLE 2

A TAXONOMY OF CONSUMER ACTIVITY AND ATTITUDE MEASURES

TABLE 3

CLASSIFICATION OF THE CONCURRENT VALIDITY OF SELECTED ACTIVITY AND ATTITUDE MEASURES

A sum score variable called New-product Proneness was created from the responses to these questions.

In another part of the questionnaire a large number of recently introduced, innovative products were listed and respondents were asked to indicate ownership of the products. The ownership responses were summed by product types and for all products. These scores became the criterion for testing the concurrent validity of the New-product Proneness Index. Recall, the research started with a theoretical construct, new-product proneness. It was developed, measured and stated as a single index (for greater reliability and content validity). Next, it was tested by examining the degree to which it was correlated with relevant reported behavior. The results appear in Table 4 along with comparisons with other related variables of interest.

TABLE 4

CORRELATION OF NEW-PRODUCT PRONENESS INDEX WITH THE NEW-PRODUCT USE INDEX, BRAND AND SLOGAN RECOGNITION INDEX AND SELECTED DEMOGRAPHIC VARIABLES. (912 FEMALE SUBJECTS, 1970 PURDUE CONSUMER BEHAVIOR RESEARCH PROJECT)

It is reassuring to find that the New-product Proneness Index is a significant specific predictor of the associated behavior, new-product ownership. Of additional interest are the findings that basic demographic variables are not closely related to the New-product Proneness Index, but that the proneness index is associated with brand and slogan recognition. Although demographic and communication variables are significantly associated with some aspects of new product ownership, the proneness index makes a useful contribution to the challenge of identifying consumers who will tend to buy new and innovative products.

A second set of data illustrates a somewhat different situation. On an exploratory basis, self-designated occupation skill (SOS) variables were introduced in the 1966 PCBRP. Factor analysis of these variables yielded results which closely approximated the prior hypothesized relationships about common groups of self-designated occupation skills. No a priori hypotheses were conceptualized to identify potential relationships between either the raw variables or the factor (or sum) scores and various behavior criteria such as media exposure or product-class usage rates. On an exploratory basis, however, SOS variables were used as independent variables in various analyses. Subsequently, these variables were revised and additional hypotheses developed before a new SOS variable set was included in the 1969 Chicago Lifestyle Study. The data from this study were factor analyzed, yielding results similar to those hypothesized. Also, the raw SOS variables were correlated with the readership of print media and the usage of various product classes. Again the findings were encouraging, motivating the use of a second revised set of occupation questions in the 1970 PCBRP survey. Factor analysis of these variables yielded satisfactory results and the raw variables have been used as predictors in several specific investigations (DeBruicker, in progress; Hustad, in progress). In addition, the occupational variables have been correlated with several large sets of potential criteria variables. Since it is impractical to present all of the relevant findings, here we will consider only the simple correlations between the raw self-designated occupation skill variables (as distinguished from occupational interest which was also studied) and four sets of criterion variables. The results of this test of concurrent validity are presented in Table 5.

TABLE 5

CORRELATIONS OF 50 SELF-DESIGNATED OCCUPATION SKILL VARIABLES WITH 66 PRODUCT USAGE, 40 PRODUCT OWNERSHIP, 57 MAGAZINE READERSHIP AND 77 PERSONAL AND SOCIAL ACTIVITIES VARIABLES. (912 FEMALE SUBJECTS, 1970 PURDUE CONSUMER BEHAVIOR RESEARCH PROJECT)

Clearly, the number of significant correlations is more than 10 times the level one would expect by chance, but the SOS skill variables are more closely associated with participation in various activities than the directly market-related behavior such as usage, ownership, and readership. Of course, if the sets of usage and readership variables had been reduced such that the results were expected to be significant on the basis of the 1969 Chicago study findings, the performance of these measures could be further improved. In any event, these variables do describe a dimension of the consumer's character not easily tapped by other measures.

Although the relationships are relatively weak, SOS variables may help shape product, copy and media strategy. The reader is left to decide for himself whether evidence of the type cited above is adequate to judge the general validity of the variable set for scholarly and applied purposes. Recall, however, that even for scholarly efforts in our field of interest validity must ultimately be considered in terms of explicit market phenomenon. Clearly, the set of interesting criterion variables potentially related to SOS variables has not been exhausted. Finally, judgments about useful levels of validity cannot be made in the absence of a specific research question (and associated cost and efficiency issues).

A third type of validity issue can be illustrated by direct application of the SOS variables to a study of consumer attitudes toward marketing institutions and practices (Hustad & Pessemier, in press). Consumers were divided into pro, anti and uncommitted groups by their attitudes about business and marketing practices. Having identified the groups, the investigators were interested in the degree to which members of the three groups differed in terms of their self-designated occupation skills (one of many classes of descriptor variables used to profile the three previously defined groups). The results for occupation variables (SOS) appear in Table 6.

TABLE 6

50 SELF-DESIGNATED OCCUPATION SKILL (SOS) VARIABLES: VERY GOOD (1) - VERY POOR (6 NORMALIZED WITHIN INDIVIDUAL, MEAN = 0, S. D. = 1.0 (912 FEMALE SUBJECTS, 1970 PURDUE CONSUMER BEHAVIOR RESEARCH PROJECT)

Twenty out of fifty SOS variables were significant, six at the .001 level and six at the .01 level. Furthermore, when these variables were combined with a large set of additional independent variables in a discriminant analysis designed to predict group membership, the discriminant function correctly reclassified 53% of the holdout sample compared to about 33% expected by random assignment (Hustad & Pessemier, in press).

The principle point to note about this illustration is the problem of the criterion, group membership. The definition of pro, anti and uncommitted groups presents a validity problem in its own right. However, even if imperfectly defined, it is not unreasonable to expect personality traits associated with self-designated occupation skills to be associated with the differences in attitudes about business practices. Therefore, the concurrent test of validity is not devoid of theoretical support in terms of the two constructs, self-designated occupation skills and attitudes toward business.

The three examples presented in the preceding discussion illustrate some of the problems encountered in examining simple concurrent validity. The balance of this paper is devoted to reviewing the reported findings we have classified in Table 3 and have summarized in Table 7.

SYNTHESIS OF FINDINGS

In the first study which we examine, Tigert utilized 37 factor scores derived from the responses of 344 individuals to 300 activity, interest, and opinion (AIO) measures. In this phase of his research, Tigert hypothesized the existence of dimensions that would be relatively independent of demographic measures. Thus, the validation criteria used by Tigert were applied in a negative sense--the lower the correlations, the more powerful the exclusive descriptive capacity of the activity and attitude measures.

Tigert found that many of the factor scores were, in fact, relatively independent of the demographic variables (Tigert, 1969).

The second study represented an attempt by Pessemier, Burger, and Tigert to discriminate among buyers of a new product utilizing numerous variables, including activity and attitude descriptors (Pessemier, Burger, & Tigert, 1967). In this study a buyer of new detergent was anyone who bought the product at least once in a seven-month period of diary keeping. An early buyer was one who purchased the product in the first 70 days after introduction. All remaining subjects who bought were classified as late buyers. The validation methodology employed was cross-classification analysis. Specifically, responses to a battery of measures were cross-classified against the early, late, and nonbuyers category.

The authors found that the significant variables were in four distinct categories: socioeconomic, trial-proneness, product-related, and informational. The following comments summarize the author's findings:

1) The opinion leadership question, the AIO factor scores on information seeking and risk avoidance, and the media factor scores on information seeking and risk avoidance, and the media factor scores were conspicuous for their inability to distinguish the kind of buyer.

2) Usage rate for the product class did not differentiate among early, late, and nonbuyer.

3) Nonbuyers had the least preference for the new brand, and the early buyer the greatest.

4) Early buyers were significantly less confident about their past brand purchases than late buyers, and late buyers were less confident than nonbuyers.

TABLE 7

SUMMARY DESCRIPTIONS OF THE CONCURRENT VALIDATION EFFORTS EXAMINED IN THIS STUDY

5) Early and late buyers indicated a greater willingness than nonbuyers to try new brands.

6) Early buyers see themselves as experimenters to a significantly greater degree than late or nonbuyers. However, early buyers did not perceive themselves as innovators.

7) Early buyers exhibit a higher degree of transmission of product information.

8) Late buyers, compared with early and nonbuyers, had a significantly higher educational level.

The third study was an extensive effort by Summers to identify women's clothing fashion transmitters (Summers, 1969). The focus was on the identification and profiling of the opinion leaders on interpersonal communication in the context of women's clothing fashions. The validation methodology employed was the cross-classification of various activity and attitude measures against fashion transmitter variables. Summers' conclusions are of interest:

"The results were consistent with previous research findings in the area of opinion leadership, and with few exceptions were also consistent with a priori judgments concerning women's clothing fashion transmitters. These findings demonstrate that real differences do exist between transmitters and nontransmitters, and that these differences can be measured.

"However, in the aggregate, these differences were not large. This indicates that, although concentrations of women's clothing fashion transmitters do exist and can be identified, opinion leadership in the area of women's clothing fashion is a wide spread trait" (Summers, 1968).

A study conducted by Ness focused on isolating, identifying and profiling key change agents in the architectural diffusion process. Ness concluded that none of his measures very effectively differentiated between innovators and noninnovators (Ness, 1968).

In a taxonomic study based on magazine readership, Bass, Pessemier, and Tigert utilized attitude and activity measures to describe readership segments (Bass, Pessemier & Tigert, 1969). The four AIO factor scores significant at the .05 level were named by the authors as follows: compulsive, orderly housekeeper; careless, irresponsible; negative attitudes toward advertising's value; and outdoor, casual activist. In addition, the authors found that subjects in the three clusters exhibited significant differences in demographic, average usage rates on selected grocery products, and average scores on brand recognition.

Another study by Tigert was undertaken to show how advertising copy can be made to be more compatible with consumer life styles. (Tigert, 1969). In this analysis, 300 AIO characteristics were correlated with the reading levels of 53 individual magazines. Most of the correlations reported by Tigert fall in range of .10 - .19. The author attributes a partial explanation to the large number of responses which piled up in the nonreadership category.

In an article dealing with the relationship between activity and attitude measures and media exposure, King and Summers present empirical data relating personality and attitudes to the level of media exposure within six broad media classes and to exposure across the media classes. (King & Summers, 1971). Included in the six media classes was television viewing, as well as five magazine groupings: general interest mass appeal, news, home, women's general interest, and women's fashion. The measurements of personality and attitudes and values were based on a standardized personality inventory developed by the authors. Cross classification analysis was undertaken to determine the relationship existing between exposure to each of the six media groups and the activity and attitude measures.

1) The set of activity and attitude variables used were ineffective in identifying respondents highly exposed to television.

2) As a group, the activity and attitude measures used were effective in predicting magazine readership for all five magazine groups analyzed.

3) Some personality and attitude and values measures were more relevant to particular magazine groupings.

4) Most of the activity and attitude measures reflected low intercorrelations with age, education, and income.

More recently Baumgarten utilized activity and attitude measures to aid him in describing fashion adoption among male college students (Baumgarten, 1971). Baumgarten's conclusion are as follows:

1) In general the results indicate that opinion leaders are most significantly different from their followers along social activities, sociological, psychological and media exposure characteristics.

2) Demographics were found to be relatively poor indicators of both opinion leadership and innovativeness.

Bruno utilized activity and attitude measures to determine the relative usefulness of a persuasibility descriptor in a media evaluation model (Bruno, 1971).

In this research the author hypothesized the existence of characteristics which would be relatively independent of demographic measures and which would be linked to television exposure rates. Bruno found few strong relationships. Only one relationship, the correlation between self-esteem and age, was above .2. Only two correlations, the correlations between social activity and morning viewing and social activity and afternoon viewing were above .1.

A rudimentary, but illuminating summary of the studies cited in Table 7 appears below:

TABLE

SIGNIFICANT RELATIONSHIPS IDENTIFIED BY CORRELATING OR CROSS-CLASSIFYING ACTIVITY AND ATTITUDE VARIABLES FROM PURDUE STUDIES WITH CONCURRENT VALIDITY CRITERION VARIABLES

It is interesting to note that, for the Purdue studies considered in this paper, the proportion of significant product and brand specific direct and mediating variable relationships is higher than the proportion of significant general and institutional relationships identified. Moreover, of the individual consumer variables, preference bears the highest number of relationships to activity and attitude variables. The reader is cautioned to avoid seeking generalizable conclusions; however, because the studies tabulated have their origins in the same research tradition, because a number of the studies were multiply classified, and because Ness' rather negative findings dominate the totals. Nevertheless, this type of summary, when considered with similar evidence from other research traditions, might provide useful insights into the applicability of certain types of activity and attitude measures.

CONCLUSIONS AND IMPLICATIONS FOR FUTURE RESEARCH

The preceding discussion provides an empirical backdrop against which we can cast some generalized conclusions and also identify areas requiring additional research.

First, it is clear that the activity and attitude research reviewed in this paper seldom incorporated explicit concurrent validity goals into the research design and subsequent analysis. Prior, general questions about validity appeared to be of secondary importance to the central focus of the author's(s') investigation. In many cases the expected relationships were not formally hypothesized. When explicit hypotheses were postulated, seldom was the extent of the expected relationship clearly described. This deficiency is probably due, in part, to the exploratory and empirical nature of the studies considered here. However, as empiricists we must be assured of the validity and reliability of measures which we utilized in describing consumer behavior. Hopefully, subsequent activity and attitude research will employ more extensively substantive a priori hypotheses and will document more clearly the extent of various predictive relationships indicated by empirical findings.

Second, the nature of the external criteria utilized for validation was often a source of concern. Occasionally, the criteria selected were arbitrary. Frequently the validity of the external criteria themselves was in doubt. On the whole, we conclude that the selection and analysis of validation criteria could be substantially improved.

Third, strong relationships were not identified in the studies reviewed here (most authors reported rather ambiguous results). This condition may be due to the fact that strong relationships between activity and attitude measures and certain external criteria do not exist. However, often it appeared that the criteria problem may have contributed to the dearth of strong identified relationships. Also, although many of the reported relationships were compatible with one or more theories of consumer behavior, there is little reason to believe that the links are strong and direct, especially for many interesting forms of low-salience classes of market behavior. We suggest that several justifiable forms of external criteria be considered in all future activity and attitude research and that the expected upper bounds of predictive power always be stated a priori. By considering alternative forms of the external criteria in this light, the validity and consequently, the potential usefulness of activity and attitude measures can be more conclusively verified.

We have stated elsewhere that standardizing some activity and attitude measures is a prerequisite to the development of efficient general purpose instruments. An equally important step should be a consolidation of the variety of approaches from diffuse exploratory studies. Efforts should be made to categorize the myriad of available measures as we have done here on a very modest scale. Our obvious goal is to build a set of activity and attitude measures which has wider application and acceptance due to its demonstrated linkage to various patterns of consumer behavior.

We are equally certain, however, that for many applied problems, activity and attitude measures that are specific to a product class are required. These variables take advantage of the specificity of the contexts, benefits, activities and attitudes that are typical of the brand strategy questions confronting marketing managers.

At this point, we will take an ambivalent position regarding the ultimate marketing usefulness of activity and attitude measures. After having considered the reliability of a selected set of them elsewhere (Bruno, 1971) and after examine their concurrent validity here, we are encouraged. However, much work remains to be done. We believe that it would be particularly helpful to draw generalizations from the results of large scale proprietary studies undertaken during the past five years. These empirical studies represent much larger and more comprehensive data bases than those conducted under the more modest auspices of the academic community.

REFERENCES

Bass, Frank M., Edgar A. Pessemier, & Douglas J. Tigert. A taxonomy of Magazine Readership Applied to Problems in Marketing Strategy and Media Selection. Journal of Business, July 1969, 42.

Baumgarten, Stephen. A Study of Fashion Adoption Among Male College Students. Ph.D. Thesis, Purdue University, June 1971.

Bruno, Albert V. An Explicit Model for the Evaluation of Television Audiences. Ph.D. Thesis, Purdue University, August 1971.

Bruno, Albert V. & Pessemier, Edgar A. Pessemier. An Empirical Investigation of Reliability and Stability of Selected Attitude and Activity Measures. Proceedings of 2nd Annual Conference of the Association for Consumer Research, University of Maryland, September 1971.

DeBruicker, F. Steward. Project Overview and Design Details. In Pessemier, DeBruicker and Hustad, The 1970 Purdue Consumer Behavior Research Project. Krannert Graduate School, Purdue University, June 1971.

DeBruicker, F. Stewart. The Use of Product Class Specific, Situational, Personality, and Behavioral Measures in Market Segmentation. Ph.D. Thesis, Purdue University, (in progress).

Guttman, Louis. Measurement as a Structural Theory. Psychometrika, December 1971, 329-347.

Hustad, Thomas P. Information Handling Behavior for Consumer Types. Ph.D. Thesis, Purdue University, (in progress).

Hustad, Thomas P. & Pessemier, Edgar A. Segmenting Consumers Market with Activity and Attitude Measures. Institute Paper No. 298, Krannert Graduate School of Industrial Management, Purdue University, March 1971.

Hustad, Thomas P. & Pessemier, Edgar A. Will the Real Consumer-Activist Please Stand Up, Marketing Science Institute, in press.

Kassarjian, Harold H. Personality and Consumer Behavior: A Review. Journal of Marketing Research, November, 1971, 409-419.

Kerlinger, Fred N. Foundations of Behavioral Research. Holt, Rinehart and Winston. 1965.

King, Charles W. & Summers, John O. Attitudes and Media Exposure. Journal of Business, 1971, 42.

Ness, Thomas E. Change Agents in the Architectural Diffusion Process. Ph.D. Thesis, Purdue University, August 1968.

Pessemier, Edgar A. & Tigert, Douglas J. Marketing Applications of Self-designated Occupation Skill Variables. Institute Paper No. 274, Krannert Graduate School of Industrial Administration. April 1970.

Pessemier, Edgar A., Burger, Phillip, & Tigert, Douglas J. Can New Product Buyers Be Identified. Journal of Marketing Research, 1967, 4.

Pessemier, Edgar A., Burger, Phillip, Teach, Richard & Tigert, Douglas. Using Laboratory Brand Preference Scales to Predict Consumer Brand Purchases. Management Science, February 1971, B371-B385.

Summers, John 0. The Identity of the Women's Clothing Fashion Transmitter. Ph.D. Thesis, Purdue University, June 1968.

Tigert, Douglas J. A Psychographic Profile of Magazine Audiences: An Investigation of a Media's Climate. Working Paper, University of Chicago, August 1969.

Tigert, Douglas J. Consumer Typologies and Market Behavior. Ph.D. Thesis, Purdue University, 1966.

----------------------------------------

Authors

Albert V. Bruno, University of Santa Clara
Edgar A. Pessemier, Purdue University



Volume

SV - Proceedings of the Third Annual Conference of the Association for Consumer Research | 1972



Share Proceeding

Featured papers

See More

Featured

G6. Brightness Increases More Positive Views of Humanity and Prosocial Behavior of People Low in Moral Identity Internalization

Jun Yan, University of Manitoba, Canada
Luke Zhu, University of Manitoba, Canada
Fang Wan, University of Manitoba, Canada

Read More

Featured

Deprivation of Control and the Phonetic Appeal of Brand Names

Jamel Khenfer, Zayed University
Caroline Cuny, Grenoble Ecole de Management

Read More

Featured

Walking the Thin Edge: The Dark Side of Brand Communities and Collecting

Emily Chung, RMIT University
Marcia Christina Ferreira, Brunel University
daiane scaraboto, Pontificia Universidad Católica de Chile

Read More

Engage with Us

Becoming an Association for Consumer Research member is simple. Membership in ACR is relatively inexpensive, but brings significant benefits to its members.