1% “-fi —-w_# >- ____-._'—.__~__...,~ fi.‘ . . kt, -w r...— —J\- 4:, ' -__..;—.‘,-- . IIIIII 3......x - _.._. _.:_:—'_"'".J“ M'— ":1”. 3:;- ”a _.~ H..— I II‘IIII' 5 III :.';:IIIIi III ‘I --*‘-.— m ‘ T I ~ ’w'..-”~ .’ ‘4»..- ~ __q III . III: ,1 IIIiII . 'IIIIIII III III I“ #. IIIIIIIII III'm ’ II o . " . Ir}? Izz‘IIIIIIII :I II ‘I; . . III” I III .1 II’II‘,’ . II “In? I?“ w I IIII.I IIII. IIIIIII IIIAI'I‘IIII‘I .1va HIIIII IIIIIII II II“ ._——— - . __.._. i.-> _ . d... _..——‘ .__'—‘ I ...__.‘M A .‘— I IIIIIII III'IIIIIII IIIII II‘I‘I'fiII'I IIIIIIII II” III I III: I I II" I IIIIII'IIIIIII I .0 -- _~ ___._. . . - _~"‘. . .2“... _ v - y -... __ _-Z_ V V ...> f -< "3‘4‘ V * - "fr-meat”? ,. ,_ 4.... - a. _ . .. ‘ . ._ - a— ‘ I. ~— ‘u. " ' ' u...— v _._-' ‘- -VI aha - -Am4 -u r .VMvv'“._,_.._ -L'ZLEY‘ - - “-3 - -4—5-’ ' ' ’ - *-.- W... III I II‘. .III III .II 3] III; III IIIIIII III' III LIII . III II‘ “in“ ”II, IIIIIIIIII.‘ I'I‘I‘ IIIIIIIIII III{I "* III 3483109 MICHlGAN STATE UNIVERSITY LIBRARIES a I “ll I III ||iil||;;;|‘;;i}|li1|i?ilii’.llii7ill 1| HII il 3 1293 00605 7271 LIBRARY Michigan State University This is to certify that the dissertation entitled An Investigation of the Indirect Relationship Between Psychographics and Buyer Behavior presented by Sammy D. Fullerton has been accepted towards fulfillment of the requirements for Ph. D. degree in Marketinq 545%”. Major professor August 28, 1987 Date MS U is an Affirmative Action/Equal Opportunity Institution 0in w «V' PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before due due. TKTW' ATE DUE DATE DUE DATE DUE AN INVESTIGATION OF THE INDIRECT RELATIONSHIP BETWEEN PSYCHOGRAPHICS AND BUYER BEHAVIOR BY Sammy D. Fullerton A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Marketing and Transportation 1987 ABSTRACT AN INVESTIGATION OF THE INDERECT RELATIONSHIP BETWEEN PSYCHOGRAPHICS AND BUYER BEHAVIOR BY Sammy D. Fullerton The role of psychographics in the task of explaining and predicting buyer behavior has yet to be resolved. This dissertation explores the possibility that a definitive relationship does exist; however, this relationship is presumed to be indirect. The influence of psychographics is mediated by two intermediate sets of variables: product usage situations and object benefits. Data were collected via a mail questionnaire. A net usable sample of 242 respondents resulted from the initial mailing of 400 questionnaires. The data were subjected to several measurement evaluation techniques prior to the testing of the hypothesized model. Once the structure of the model had been confirmed, the hypothesized relationships defined in the model were subjected to a causal (path) analysis. The causal analysis indicates that there is merit to the presumption that the relationship between psychographics and buyer behavior is indirect. There is a support for a causal relationship between psychographics and product usage situations. Similarly, the path analysis isolated a causal relationship between product usage situations and object benefits; this relationship was not as strong as was the aforementioned causal linkage. The strength of the observed relationships varied considerably from one psychographic dimension to another. The final linkage tested was the causal path between psychographics and object benefits. This relationship was very weak for all of the tested psychographic dimensions. The results are meaningful for the practitioner as well as the marketing theorist. It appears that psychographics does influence buyer behavior; however, that influence is indirect and filtered though the two aforementioned sets of intervening variables. iv ACKNOWLEDGMENTS The author appreciates the assistance and advice of the dissertation chairman Dr. Glenn S. Omura, who was instrumental in the development of the theoretical framework for this study. Also to be thanked are committee members, Dr. Forrest 8. Carter and Dr. Gilbert D. Harrell. Their assistance in developing the methodology and periodically reviewing this document was greatly appreciated. A final acknowledgment goes to fellow doctoral candidate David Hartman; his help in running the required computer programs was also appreciated. AN EXPLORATION OF THE INDIRECT NATURE OF THE RELATIONSHIP BETWEEN PSYCHOGRAPHICS AND BUYER BEHAVIOR TABLE OF CONTENTS INTRODUCTION AND OVERVIEW CHAPTER ONE C C O I O O O O O O O O O O O O O O O H INTRODUCTION . . . . . . . . . . . . . . . STATEMENT OF THE PROBLEM . . . . . . . . . . OBJECTIVES . . . . . . . . . . . . . . . . . SCOPE OF THE STUDY . . . . . . . . . . . CONTRIBUTION TO THE DISCIPLINE . . . . . ORGANIZATION OF THE FOLLOWING CHAPTERS . . . HOKDCDQN FJH LITERATURE REVIEW CHAPTER TWO 0 O O O O O O O O O O O O O O O O O 0 14 INTRODUCTION . . . . . . . . . . . . . . . 15 PSYCHOGRAPICS AS INPUT FOR STRATEGIC DECISIONS . . . . . . l7 PSYCHOGRAPHICS AND BEHAVIOR. AFFIRMATIONS . 21 PSYCHOGRAPHICS AND BEHAVIOR: DISAFFIRMATIONS 26 SUMMARY AND NEW DIRECTIONS . . . . . . 33 THE PSYCHOGRAPHICS- -PRODUCT USAGE SITUATION LINKAGE . . . . . . . . . . . 35 RESEARCH HYPOTHESIS ONE . . . . . . . 41 THE PRODUCT USAGE SITUATIONS- -OBJECT BENEFIT LINKAGE . . . . . . . . . . . . 42 RESEARCH HYPOTHESIS TWO . . . . . . 48 THE PSYCHOGRAPHICS- OBJECT BENEFITS LINKAGE . 49 RESEARCH HYPOTHESIS THREE . . . . . . . 52 OVERVIEW OF THE LITERATURE REVIEW . . . . . . 52 METHODOLOGY AND AN ILLUSTRATIVE EXAMPLE ‘ OF THE MODEL CHAPTER THREE 0 O O O O O O O O O O O O O O O O O 53 INTRODUCTION . . . . . . . . . . . . . . . . 54 METHODOLOGY . . . . . . . . . . . . . . . . 54 FOCUS GROUPS . . . . . . . . . . . . . . . . 54 QUESTIONNAIRE . . . . . . . . . . . . . . . . 56 PRETEST . . . . . . . . . . . . . . . . . 58 SAMPLING PROCEDURES . . . . . . . . . . . . . 58 DATA ANALYSIS . . . . . . . . . . . 59 THE MODEL: AN ILLUSTRATIVE EXAMPLE . . . . . 64 vi RESULTS CHAPTER FOUR C O O O O O O O O O O O 0 C 0 INTRODUCTION . . . . . . . . . . QUALITATIVE RESEARCH EFFORTS . . . . FOCUS GROUP STRUCTURE . . . . . . . FOCUS GROUP RESULTS . . . . . . . . . THE QUESTIONNAIRE . . . . . . . . . PRELIMINARY PROCEDURES . . . . . . . THE QUANTITATIVE RESEARCH PROCESS . . SAMPLE RESULTS . . . . . . . . . . . PSYCHOGRAPHIC SCALES . . . . . . . . EVALUATION OF THE MEASUREMENT MODEL . SIMILARITY COEFFICIENTS . . . . CONFIRMATORY FACTOR ANALYSIS . ACTUAL VERSUS PREDICTED CORRELATIONS THE INDIVIDUAL SCALE ITEMS . . . EVALUATION OF THE STRUCTURAL MODEL . AMONG TEST OF THE EQUIVALENCE OF THE COVARIANCE MATRICES O O O C O O O O O O I C PATH ANALYSIS ISSUES . . . . . . . . PATH ANALYSIS RESULTS . . . . . . . . CONCLUSIONS AND RECOMMENDATIONS CHAPTER FIVE O O O O O O O O O O O O O O I INTRODUCTION . . . . . . . . . . . . THE MEASUREMENT MODEL . . . . . . . . THE STRUCTURAL MODEL . . . . . . . . HYPOTHESIS ONE . . . . . . . . . HYPOTHESIS TWO . . . . . . . . . HYPOTHESIS THREE . . . . . . . . THE MODEL REVISITED . . . . . . . . . LIMITATIONS . . . . . . . . CONTRIBUTION TO THE DISCIPLINE . . . MANAGERIAL IMPLICATIONS . . . . . . . ADDITIONAL RESEARCH . . . . . . . . . SUMMARY . . . . . . . . . . . . . . . APPENDIX . . . . . . . . . . . . . . . . . STATISTICAL TABLES . . . . . . . . . FOCUS GROUP GUIDELINES . . . . . . . DATA COLLECTION INSTRUMENT . . . . . BIBLIOGRAPHY . . . . . . . . . . . . . . . vii 66 67 67 68 69 73 73 74 74 79 80 81 86 87 88 89 93 97 98 98 .102 .104 .105 .106 .106 .107 .108 .110 .111 .112 .113 .114 .118 .120 .128 LIST OF TABLES Table 2-1 AFFIRMATIVE RESULTS DISAFFIRMATIVE RESULTS DISAFFIRMATIVE RESULTS: MEASUREMENT ISSUES INITIAL PSYCHOGRAPHIC DIMENSIONS EXAMPLE OF THE MODEL USING MATERIALISM PSYCHOGRAPHIC SCALES FOR THE INITIAL ANALYSIS COEFFICIENT ALPHA FOR EACH RETAINED SCALE PRODUCT USAGE SITUATIONS AND OBJECT BENEFITS FACTOR LOADINGS FOR THE 15 ITEMS WHICH MEASURE THE PSYCHOGRAPHIC SCALES FACTOR LOADINGS FOR THE 30 ITEMS WHICH MEASURE PSYCHOGRAPHIC, PRODUCT USAGE SITUATIONS AND OBJECT BENEFITS FACTOR LOADINGS TO TEST THE MEASUREMENT MODEL CHI-SQUARED TEST RESULTS TEST OF EQUIVALENCE OF COVARIANCE MATRICES EVALUATION OF ALTERNATIVE MODELS PATH AND R-SQUARED COEFFICIENTS FOR THE MODEL LINKAGES ‘RESULTS OF THE RELIABILITY ANALYSIS FOR THOSE SCALES NOT USED IN THE TESTING OF THE PROPOSED MODEL SIMILARITY MATRIX Page 27 34 34 70 72 72 76 77 78 83 85 87 89 91 94 114 117 CHAPTER ONE INTRODUCTION AND OVERVIEW INTRODUCTION The last twenty years have witnessed a tremendous infusion of psychographics into the attempts to further explain buyer behavior. Academicians and practitioners have incorporated psychographic dimensions within market segmentation schema, sales forecasting models, and general models of consumer behavior. For example, in the task of segmenting a heterogeneous market, psychographics is viewed as a tool which may suggest ways in which each unique segment can be best served [Frye and Klein 1974]; furthermore, it provides management with the basis for the development of a more effective marketing mix [Hodock 1974; Ziff 1974]. Psychographics is also presumed by many to have a role in the task of environmental scanning. Marketers scrutinize the changes which are occurring in the marketplace and evaluate the impact that these changes will have on demand. The resultant forecast is predicated, in part, upon consumer psychographics. Such scanning has spawned the popularity of syndicated suppliers such as SRI, Simmons, and Yankelovich, Skelly and White. The popularity of such commercial suppliers indicates that practitioners are somewhat fond of this concept; however, marketing theorists do not share their zeal for this technique. Typical of this feeling is Wells's [1985] belief that the use of psychographics in environmental scanning is suspect, especially when the task encompasses long range projections. Conversely, marketing theorists have frequently sought to include life styles and psychographics as integral parts of models which attempt to eXplain general buyer behavior [Nelson l97l; Pernica 1974]. For psychographics to assume a legitimate, broadly accepted role in today's marketing environment, additional research is needed. The role of psychographics has not yet been defined to the point that academicians and practioners feel comfortable with the routine incorporation of psychographics into the task of developing marketing strategy. This dissatisfaction is based upon a number of disappointments in the marketplace despite the fact that there have been a number of unqualified successes. Perhaps the problem then is one of understanding the nature of the relationship between psychographics and consumer behavior. The task thus becomes one of developing a model which reflects the system of components as well as the relationships among these components as they relate to the linkage between psychographics and consumer behavior. Virtually every study in this area has sought to directly tie psychographics to buyer behavior. Unfortunately, previous studies have not been particularly fruitful. Inconsistent results have limited the abilityto generalize the relationship between psychographics and buyer behavior. Such inconsistencies must be explained; otherwise marketers will always be dubious of any attempt to incorporate psychographics as a source of information germane to the development of marketing strategy. Once this has been accomplished, marketers will be better able to make strategic decisions. Horton [1984] provides the basic framework which gives direction to the investigation of the relationship between psychographics and buyer behavior. He indicates thatpsychographic factors "tend to affect buyer behavior in broad non-specific ways." Pernica [1974] also explores the presumption that the linkage is indirect; his research indicates that psychographics seems to be filtered through some set of the consumers' values or choice criteria. Commercial suppliers such as Stanford Research Institute provide practical support for the belief that there is a linkage, albeit perhaps indirect, between psychographics and purchase behavior [Riche 1982; Nordberg 1981; Meyer 1983]. The key question to answer concerns the nature of the intermediate linkages between psychographics and behavior. Does some set of moderating variables exist, and if so, how might the system of relationships be described? One plausible relationship is that a product is purchased in order to accomplish some task deemed important to the buyer. (A task is basically a utility-providing function which is achieved via the purchase and consumption of a particular product.) The problem then becomes one of identifying factors which influence the purchase process and delineating those tasks which the consumer seeks to accomplish with the purchase of a product. This rationale supports the basic premise eXplored in this project. //1 The initial hypothesized linkage involves the benefits ,» which are derived from the purchase and consumption of a particular product; these benefits are presumably influenced jifby the particular situation with which the consumer is lconfronted. The presumption is that the psychographic profile of an individual will partially determine the set of product usage situations which mediate that individual's behavior in the marketplace. (A product usage situation is defined as an objective, unique to a particular situation, which an individual seeks to achieve from some specific mode of behavior.) Examples of product usage situations include livening up the party, conveying status, and facilitating interaction. The consumer's evoked set is developed based upon the aggregation of these product usage situations and a product's ability to provide the requisite benefits. I- How then does a product usage situation influence the purchase decision? The product being purchased must be 1 capable of resolving the needs of the individual purchaser; these needs are influenced by the product usage situation. This idea gives rise to the concept of an object benefit; an object benefit is more closely aligned with purchase .behavior and is an important criterion utilized in the development of a marketing strategy. (An object benefit is defined as a basic benefit, associated with a specific marketing mix component, which exists or accrues in the exchange or consumption process.) Examples of object benefits include speed for an automobile, softness for a sweater, and efficiency for a major appliance. The presumption is that the requisite object benefits are influenced by psychographics; more so, they are directly influenced by the product usage situation with which the consumer is faced. Finally, the object benefits sought by the consumer will influence the ultimate purchase decision; the product purchased will be the one which, in the consumer's mind, best coincides with the set of object benefits deemed salient by the purchaser. For example, one will not purchase a pair of designer jeans if the lone salient object benefit is a low price. In summary, the initial intervening component between psychographics and buyer behavior is hypothesized to be the product usage situation; this is followed by the object benefits and, ultimately, some specific mode of behavior in the marketplace. An example of this series of components is as follows. Consider the individual who is in the market for a new car. The psychographic variable, materialism, might lead that consumer to a product usage situation which reflects the need to convey status. This situation will mediate the consumer's evoked set in such a way that only automobiles which possess the required object benefits of being expensive and European-produced are included. In response to these object benefits, the consumer purchases a Mercedes—Benz 380 SL. Furthermore, it is expected that psychographics also influences the set of object benefits sought by the consumer. In this example, the presumption is that there is a relationship between materialism and the desire to purchase an expensive automobile. This basic system of relationships is subjected to scrutiny in this project; the model is depicted below. GENERAL PRODUCT OBJECT PURCHASE LIFE STYLES USAGE BENEFITS _ _9 BEHAVIOR AND SITUATION PSYCHOGRAPHICS STATEMENT OF THE PROBLEM The literature indicates that the examination of the underlying theories regarding psychographics is incomplete. Results have been varied at best, perhaps even disappointing. This research is directed toward the resolution of this deficiency. According to Kakkar and Lutz [1981], "the interaction of situational, individual, and object-related variables holds the key to scientific explanation—-as well as sound ')decision making by marketers--with respect to consumer I I I [ behavior in the marketplace." This "interaction" prov1des \\fhe foundation for the basic research problem. This research problem consists of two basic components. From a general perspective, the basic problem concerns the evaluation of the relationship between psychographics and buyer behavior. More specifically, the problem involves the evaluation of the two variable sets which are hypothesized to mediate this relationship. The following set of objectives provides specificity to the problem statement. OBJECTIVES There are several important objectives which are to be accomplished via this research. The literature review provides the vehicle for accomplishing the set of preliminary objectives, each of which relates to a specific section in chapter two. These preliminary objectives are: -to identify those research efforts which support and those which fail to support the use of psychographics in the analysis of buyer behavior; -to determine the underlying reasons why some efforts to use psychographics succeed and some fail; -to delineate support for the proposition that there is an indirect relationship between psychographics and buyer behavior; and -to develop a model which reflects the intervention of product usage Situations and object benefit variables between psychographics and buyer behavior. With each of these preliminary objectives accomplished, the focus will shift to the primary objective of this study. This objective will be accomplished via the collection and analysis of primary data. The objective is: -to determine the nature of the relationships among the variable sets identified in the literature review and delineated in the resultant model. This objective will be accomplished by empirically testing the model. The results should provide a valuable contribution to the marketing discipline. SCOPE OF THE STUDY This research is decidely theoretical in scope. As previously discussed, it involves the development of a model depicting a hypothesized, systematic relationship between psychographics and buyer behavior; the collection of primary data; and the evaluation of the relationships depicted in the aforementioned model. It is limited to an evaluation of those variables which are hypothesized to mediate buyer behavior; consequently, no attempt is made to explicitly tie any of these variables to the actual purchase process. The analysis is further constrained to a single product. However, it is the evaluation of the model which is paramount; the product selected simply facilitates that evaluation. CONTRIBUTION TO THE DISCIPLINE This project provides three basic contributions to the marketing discipline. First, it helps to clarify the role Of psychographics in marketing. As mentioned in the introduction, there is a great deal of controversy regarding that role. The model under scrutiny utilizes the concept of an indirect linkage. This departure from previous research initiatives will provide the opportunity to reassess existing models and aid in the development of future models of buyer behavior. Second, but more importantly, the research provides a basic theoretical framework which depicts and assesses a new approach to modeling the relationship between psychographics and consumer behavior. Finally, this project provides a basic framework by which the study may be replicated with those changes relevant to a marketer's particular research needs. For example, the list of products may be changed to be congruent with the offerings of a particular manufacturer or retailer. Ideally, the results will establish a theoretical and methodological framework for subsequent research which will aid in the development of marketing strategy. ORGANIZATION OF THE FOLLOWING CHAPTERS Chapter Two consists of an exhaustive literature review which pertains specifically to the problem at hand. The initial section of the literature review is an evaluation of the direct relationship between psychographics and buyer behavior. This section is intended to highlight the fact that there is no clear consensus of opinion, thereby indicating that further research is required in order to assess the dynamics of the presumed relationship. Chapter Two continues with a comprehensive review of the literature which addresses the concepts of the product usage Situation and the object benefits as well as their relationships with psychographics and buyer behavior; this section of the literature review relates Specifically to the model depicted earlier in this chapter. Research which addresses the relationship between psychographics and the product usage situations is presented first; this is followed by a review of that research which addresses the the existence of a linkage between product usage situations and object benefits. From this section of the review, the research hypotheses are extracted and formally stated. Chapter Three outlines the methodology utilized to conduct this study; it provides a detailed overview of the analytical procedures used. The chapter begins with a delineation of the techniques used in the qualitative research phase of the study and continues with a summary of the quantitative-based procedures. Also included is the rationale which supports the use of the selected techniques. Chapter Three also provides a detailed description of the proposed model. Chapter Four begins with a summary of the results of the qualitative research efforts and a discussion of the techniques used in the development of the data collection instrument. Preliminary issues related to the sampling process are also addressed in this chapter. Chapter Four continues with a comprehensive discussion of the results of the data collection and analysis procedures. The final set of psychographic variables is identified, and the items used to measure each scale are identified. Then, the results of the procedures used to test the proposed model are summarized. Information gleaned from Chapters Two and Four provide the basis for drawing conclusions and making recommendations; these tasks constitute the majority of Chapter Five. This chapter also contains an assessment of the research project. Conclusions are drawn regarding the usefulness of the results and the practicality of converting the results to management action. Additional research which is viewed as complementary is outlined. Then the results are evaluated as to the extent to which the project contributes to the advancement of the body of marketing knowledge. The final entries include the appendix and the bibliography. The appendix contains copies of the supporting documents. A facsimile of the questionnaire and copies of the basic documents used in directing the focus group interviews are also included. The bibliography is provided after the appendix. CHAPTER TWO LITERATURE REVIEW ”I A INTRODUCTION The future development and subsequent utilization of psychographics is dependent upon the development of a better theoretical linkage with buyer behavior [Plummer 1976]. Though Plummer's concerns were voiced ten years ago, such theoretical development has yet to result from subsequent research efforts. Typical of the outgrowth of Plummer's belief is the feeling that marketers will have to adapt their communications efforts to meet the changing life styles of the marketplace. Mass marketing will be a less viable approach [Meyer 1983]. Segmentation strategies based upon psychographics will be one of the keys to success. Marketers who monitor the changes in the individual segments will be better able to anticipate how such changes affect demand, and they will be able to respond accordingly [Hong 1979]. There is a modest level of support for the belief that the relationship, if indeed one does exist, is filtered through some set of mediating variables. Yet there is no consensus of opinion as to what Specific types of variables might intervene between psychographics and buyer behavior. For example, Belk [1975] and Kakkar and Lutz [1981] discuss the role of and the need for a situational perspective; these situational variables are viewed as potential mediators of buyer behavior. Pernica [1974] discusses "choice criteria" as a means of matching life style to purchase behavior. This apparent support for the existence of one or more intermediate linkages between psychographics and buyer behavior provides the basis for directing this review of the literature. This chapter consists of six distinct sections. Each section relates to a specific objective outlined in chapter one. The first section provides a review of that research which assesses the role of psychographics as input for marketing strategy decisions. The second section summarizes those studies which provide support for the belief that there is a direct relationship between psychographics and buyer behavior. This section is followed by a summary of those research efforts which appear to disaffirm the existence of such a relationship. The remaining sections concern the objectives associated with the development of the proposed model. Each section addresses one of the three hypotheses germane to the model. Section four examines the literature which reflects the nature of the relationship between psychographics and the consumers' product usage situations; this research provides the basis for stating the initial research hypothesis. Section five addresses the relationship between the product usage situations and the object benefits thereby providing the basis for the second research hypothesis. Finally, the relationship between psychographics and object benefits is reviewed. This review results in the development of the third research hypothesis. PSYCHOGRAPHICS AS INPUT FOR STRATEGIC DECISIONS A primary use of psychographics has been and continues to be the description of the general marketplace. These results are often applicable when formulating marketing decisions. This section addresses these applications. Though these articles tend to be more conceptual than theoretical, they provide a foundation of support for the premise that there is a relationship between psychographics and buyer behavior. A basic premise of some authors is the belief that psychographic analysis is beneficial to any firm which seeks to invoke the marketing concept [Reynolds and Wells 1978]. The most common areas of application involve the process of market segmentation, the development product strategy, and the development of the most appropriate communications strategy. As an indication of the presumed applicability of psychographics, Lazer [1981, 1983b], Jones [1982]; and Beck [1984] indicate that such issues are important in formulating marketing strategy. Another perspective on this relationship is provided by Davidson and Rogers [1981], who state that the key to success for marketers in the eighties will depend, in part, upon their ability to implement marketing strategies based upon psychographic considerations. It is viewed as a principle which allows management to better assess the needs of relevant market segments because demographic descriptors have proven inadequate for this task [Forest and Blumberg 1981]. Psychographics can be of tremendous value to marketing managers [Ahmed and Jackson 1979]; it facilitates the reduction of large, heterogeneous populations into a few basic groups [Rossen and Sweitzer 1981; Leonard-Barton 1981; Barnes and Sooklal 1983; McLaine 1983; Cutler 1983; Chilcutt 1983; Roseman 1978; Zarry 1979; Spiegel 1982; Baer 1982; Richards and Sturman 1977; Reynolds and Crask 1977; Nordberg 1981; and Maher 1983]. Psychographics is used to monitor changes in the pOpulation. It is believed that the recognition of these changes will alert perceptive marketers to new obstacles and opportunities in the marketplace [Blackwell 1980; Johnson 1979; Mitchell 1983; Tatzel 1983]. For example, attitudes toward marriage and cohabitation are reported to be factors which need constant monitoring by the life insurance industry [Price 1984]. Psychographic analysis adds character and multidimensionality to the consumer [Hoke 1982; Lehmkuhl 1983]. Such character and multidimensionality offer marketers a more viable approach for the development of marketing strategy in the long run, as well as in the short run. Product decisions are also influenced by psychographics. General applications of psychographics have been advocated for retailers [Blackwell and Talarzyk 1983]; the authors note that successful retailers have begun to implement a portfolio management approach which focuses on the need for product lines consistent with the psychographics-based needs of key target markets. Other product-related applications have been documented in the area of direct marketing [Schell 1982], the purchase of time—saving and time-using products [Blackwell 1980], housing and financial services [Thygerson and Parliament 1979; Koehn 1980; Lavidge 1979], self-protection devices [Wettenstein 1982], product sweeteners [Martin 1982], consumer electronics [Merahn 1981; Regan 1981], newspapers [Miller 1974; Heary 1984], and magazines [Lande 1982; LeRoux 1984]. Promotion decisions have not been immune from the influence of psychographics either [Aaker 1982]. Young and Rubicam credits the use of psychographic information for part of their success in developing effective advertising COpy and selecting appropriate media. Yet they note that such applications are in their infancy and that there is considerable room for improvement [1981]. Psychographic data also provide valuable insight into the task of creating an effective brand identity [Berry 1983; Feigin 1981]. Obvious psychographic differences are also in evidence when magazine readership is examined; these differences are important to publishers and advertiSers [Venkatesh 1980; Gordon 1981]. Other examples of promotional decisions influenced by psychographics include the repositioning of products to reflect an approach to relaxation [Langer 1982], convenience foods [Kruger 1980], and restaurants [Shapiro and Bohmbach 1978]. The preceding review has documented the wealth of practical support for the belief that there is a relationship between psychographics and buyer behavior. While it is important to note the presumed successes of psychographics, it is equally important to note that the existence of such practical support is inadequate for the task of developing or testing theoretical constructs. This review continues with the identification of empirical studies which either affirm or disaffirm the direct linkage and ends with the delineation of the research which addresses the nature of the intervening variables. It is this empirical research which begins to address the theoretical issues which represent the ultimate focus of this project. PSYCHOGRAPHICS AND BEHAVIOR: AFFIRMATIONS This section reviews those empirical studies which document and support the existence of a relationship between psychographics and buyer behavior. Included in this discussion is a tentative explanation for many of the affirmative results. Early research indicates that much of the emphasis revolved around the factor analysis of a large number of psychographic items and the subsequent evaluation of their relationship with some aspect of buyer behavior. Pessemier and Tigert [1966] report that factor analyzed scales seem capable of delineating homogeneous market segments. These results are echoed in a number of subsequent studies [Bass, Pessimier and Tigert 1967; Lessig and Tollefson 1971; Tigert 1969; Wells 1971; Wilkie 1970; and Wilson 1966]. Additional applications include segmentation studies by Bushman [1971]; Darden and Perreault [1976]; Goldberg [1971]; Hustad and Pessimier [1971a; 1971b]; Richards and Sturman [1977]; and Ziff [1971]. Despite the apparent successes documented here, the results are tempered by the fact that the resultant relationships are generally quite weak; the resultant clusters explain no more than ten percent of the total variance in buyer behavior. Recently, the application of psychographics has addressed specific modes of individual behavior. For instance, there is an indication that individuals who might be characterized as socially concerned tend to be more physically active and they exhibit a greater propensity to engage in philanthropic or altruistic activities [Belch 1982]. Those individuals who engage in consumerist activities tend to differ in life style from those individuals who do not generally engage in these types of activities [Bourgeois and Barnes 1979]. Kinnear, Taylor, and Ahmed [1972] evaluate ecologically concerned individuals; their research isolates a psychographics-based tendency to engage in conservation activities. Thus, it is evident that there is an apparent relationship between psychographics and the propensity to engage in a number of activities (physical, philanthropic, consumerist, and conservation). Numerous product-Specific examples are.in evidence in the literature. Meaningful relationships have been discovered or documented; they include automobiles [Bernstein 1978], convenience foods [Bernstein 1978], beverages [Huber et a1. 1982], artificial sweeteners [Martin 1982], and convenience food stores [Darden and Lumpkin 1984]. Other encouraging results document the relationship between psychographics and credit card use [Paksoy 1979], attendance at the performing arts [Andreasen and Belk 1980], the allocation of time to the mass media [Hornik and Schlinger 1981], newspapers [Bryant 1981], store patronage [Gutman and Mills 1982], vacation decisions [Schul 1983] and generic groceries [Plummer 1974; Anvik and Ashton 1979; Granzin 1981; Bellizzi et a1. 1981; McEnally and Hawes 1980]. _//E The relationship between psychographics and product 4 choice represents a major area of investigation. According to Cosmas [1982], there is a relationship between one's life style and one's total product assortment. The purchase of items pertaining to personal care, cooking and baking, self-indulgence, personal appearance, and children is related to the individual's psychographic profile. The relationship is best described as modest. The results of a study by Boote [1980] indicate that psychographic variables such as independence, novelty seeking, and traditional orientation exhibit a significant relationship with product choice. While citing the "overwhelming inconsistency" surrounding the use of psychographics, Kinnear and Taylor [1976] indicate that marketers of petroleum products can benefit from its use. Psychographic differences are in evidence for users of major brands, regional brands, and discount brands of gasoline; and the purchase of complementary products is also related to the brand segment to which one belongs. This study also documents significant psychographics-based differences between light and heavy beer drinkers. Psychographics is seemingly related to decisions to patronize selected financial institutions and to utilize certain services. Selection of a financial institution is directly tied to issues such as cosmopolitan orientation and traditionalism [Stanley, Danko and Hirshman 1980]. Differences in factors such as convenience orientation, innovativeness, and price sensitivity play a major role in determining one's propensity to use an automated teller machine [Stanley and Mochis 1983]. Results of another study indicate that there is a linkage between psychographic characteristics and credit card usage [Paksoy 1979]. Attitudes toward credit and fashion consciousness exhibit the strongest relationship with such use. The observed relationship can be attributed, in part, to the tautological nature of the variables; psychographic variables which relate specifically to credit are used to predict credit card usage. The purchase of food products has frequently been tied to psychographics. For example, Hawes and Kiser [1980] report that the generic brand-prone purchaser is likely to be an information seeker; Anvik and Ashton [1979] and Granzin [1981] document an apparent relationship between the propensity to purchase generics and innovativeness. To the extent that generics represent a new entry in the marketplace, perhaps this relationship should have been anticipated. Other affirmations of the existence of such a relationship between psychographics and behavior in the marketplace abound. For instance, the purchase of convenience foods is related to the purchaser's self-confidence, community orientation, financial optimism, and home orientation [Darden and Lumpkin 1984]. Semenik and Young [1980] indicate that there is a modest relationship between psychographics and attendance at cultural events. Fashion life styles can predict factors such as the number of apparel purchases and average price points for an individual consumer. This is viewed as important because fashion image is seen as a major criterion in one's decision to shop at a particular mall [Quarles 1982] or store [Bearden et a1. 1978]. Product and segmentation scenarios are not the only marketing issues to have been researched. A number of studies have looked at the relationship between psychographics and a variety of media-related issues. Mills [1983] delineated a relationship among buying motives, buying behavior, and media usage patterns. The relationship between life styles and media time use has also been examined. Hornik and Schlinger [1981] indicate that the selection of a particular medium is related to a number of psychographic dimensions. Similarly, the radio market in Perth, Australia, has several segments which exhibit distinguishable psychographic differences [Soutar and Clark 1983]. To the extent that different stations have different formats, the indication is that psychographics influences media exposure. Clearly, there is considerable support for the belief that psychographics can be used to describe or predict buyer behavior. Table 2-1 summarizes the research which reflects the direct, positive nature of this relationship; also included is an explanation of why each of the observed relationships exists. Not all results, however, have supported this relationship. The next section looks at these disaffirmative studies. PSYCHOGRAPHICS AND BEHAVIOR: DISAFFIRMATIONS The dilemma regarding the efficacy of psychographics as a predictor of buyer behavior is heightened by those studies which fail to indicate meaningful relationships. Such studies have found that either no relationship exists or that relationships which have been documented may be statistically significant but managerially meaningless, or that the results are too inconsistent to endorse psychographics as a determinant of marketplace behavior. In general, these failures can be attributed to a limited number of factors. Foremost among these factors are TABLE 2-1 AFFIRMATIVE RESULTS FIRST AUTHOR SUBJECT EXPLANATION Andreasen 1980 Performing arts Theoretical link Anvik 1979 Generics Theoretical link Bass 1967 Segmentation Modest relationship Belch 1982 Philanthropy Theoretical link Bellizi 1981 Generics Theoretical link Bernstein 1978 Automobiles Methodologically weak Boote 1980 Product choice Broad Bryant 1981 Newspapers Theoretical link Burgeois 1979 Consumer activity Tautology Bushman 1971 Segmentation Modest relationship Cagley 1980 Generics Theoretical link Carrey 1984 Castle 1981 Cosmas 1982 Crask 1978 Darden 1974 Darden 1976 Darden 1979 Darden 1984 Dickerson 1983 Goldberg 1971 Granzin 1981 Gutman 1982 Hawes 1980 Hornik 1981 Huber 1982 Hustad 1971 King 1971 Kinnear 1976 Kinnear 1976 Lessig 1971 Martin 1982 McEnally 1984 Michaels 1972 Mills 1983 Paksoy 1979 Pessimier 1966 Plummer 1971a Plummer 1971b Quarles 1982 Richards 1977 Roberts 1979 Roberts 1984 Schaninger 1981 Schul 1983 Semenik 1980 Soutar 1983 Stanley 1980 Solar energy Direct marketing Total assortment Magazines Store preference Segmentation Media usage Convenience foods Personal computer Segmentation Generics Store selection Generics Media time use Beverages Segmentation Media exposure Brand choice Conservation Segmentation Artificial sweet Generics Media exposure Shopping behavior Credit cards Segmentation Credit cards Advertising Apparel Segmentation Food shopping Coupons, brands Food/TV/clothes Vacation behavior Performing arts Radio Financial Inst. Spurious Very broad Modest relationship Reasonable Small sample Modest relationship Inferred relationship Small sample Theoretical link Modest relationship Theoretical link Tautology Tautology Theoretical link Theoretical link Modest relationship Modest relationship Reasonable Tautology Modest relationship Theoretical link Theoretical link Modest relationship Tautology Tautology Modest relationship Modest relationship Modest relationship Tautology Modest relationship Modest relationship Measurement deficiency Non-psychographics included Tautology Modest relationship Small sample Tautology TABLE 2-1 (Continued) FIRST AUTHOR SUBJECT EXPLANATION Stanley 1980 ATMs Non—psychographics included Teel 1979 TV viewing Broad Tigert 1969 Segmentation Modest relationship Wells 1971 Segmentation Modest relationship Wilkie 1970 Segmentation Modest relationship Wilson 1966 Segmentation Modest relationship Yaves 1980 Donor Broad Ziff 1971 Segmentation Modest relationship problems associated with measurement and methodological issues. Other basic problems found include sampling deficiencies and the lack of a theoretical basis to link psychographics and buyer behavior. Early discussions of psychographics provided a precursor of the limitations regarding its role. These studies delineate a presumed inherent weakness in psychographics; this weakness involves the imprecision of the measures. Demby [1977] states that psychographics provides only a measure of propensity; thus, it is seen as helpful but not a panacea for those involved in predicting buyer behavior. Blair [1976] indicates that psychographics does not provide an exact measure, but rather a reflection of general tendencies. Such statistical imprecision does not lend itself to accurate prediction and theory development. Wells [1975] states that psychographics may supplement demographics, but when used independently, ambiguity is likely to persist. Such ambiguity inhibits the successful use of psychographics in develOping marketing strategy. Yet another problem is that psychographics has been viewed as abstract and unable to predict forthcoming changes in buyer behavior. While some short-term relationships may be isolated, long-term trends are more difficult to project. Since consumers' psychographic profiles are dynamic rather I , \ ' L, “ J 1' than static, the efficacy of psychographics-based decision making will likely deteriorate over time [Wells 1985]. A sharp indictment against psychographics is provided by Adams [1982]. He delineates several areas which inhibit the capabilities of psychographics. For example, it is believed that life style segmentation results in redundancy. Researchers assign new labels to patterns of behavior previously documented. As a consequence, people who exhibit identical psychographic profiles might be assigned different labels. This problem is exacerbated by the contradictory approaches used by syndicated suppliers, and it further limits the applicability of psychographics in marketing. Concerning the usefulness of psychographic data from syndicated sources such as SRI's VALS, Yuspeh [1984] contends that such approaches to marketing are "too simple, too remote, and too unreliable"; and she recommends that their use be limited to descriptive applications rather than those applications which purport to predict behavior or effectively segment the market. Other common areas of complaint include the lack of standardized measures [Forest et a1. 1981], the misuse of psychographics by marketers and predictive associations which provide little actionable information [Boote 1980] and inconsistent results [Kinnear and Taylor 1976]. Kassarjian [1971] states that it is probably unreasonable to expect psychographic variables alone to explain a large percentage of the variability in the product purchase decision. Perhaps the most prevalent measurement problem is the inherent lack of reliability associated with many scales [Lastovicka 1982]. The following studies document this fact. Boote [1981] points out that much of the literature on :Fpsychographics fails to address the issue of reliability. 5'" ~. .This fact might easily explain the inconsistency of the "results emanating from the various studies, and is further . delineated in Lastovicka's [1982] subsequent study. This line of thinking has been echoed by others. In their test of the reliability of psychographic measures, Burns and Harrison [1979] imply that no set of responses to psychographic items will embody an acceptable level of precision for the prediction of individual behaviors. This does not bode well for those practitioners who are considering the use of psychographics and it can perhaps eXplain why so many studies have failed. ~ Wells [1976] likewise indicates that some of the failures to find meaningful relationships between psychographics and buyer behavior can be attributed to the instability of the measures themselves; his implication is that if these measurement problems could be overcome, then meaningful relationships might be documented. The lack of a theoretical linkage between psychographics and buyer behavior is also a fundamental reason for many previous failures. The following studies reflect this deficiency. Crask and Reynolds [1980] eXplore the nature of the market in terms of exposure to print as well as electronic media. While it is again apparent that clearly identifiable segments do exist, there is no theoretical support for the belief that psychographics reflects a viable means of adding to the distinctiveness of the observed segments. It is hypothesized that psychographics will outperform demographics in the task of identifying market segments as they relate to media usage. The results indicate that this is not the case; in this situation, demographics appears to be the more viable approach [Cannon and Merz 1982]. These results should not be surprising since there was no Specific theory which related psychographics to any specific behavioral manifestation. /// A broader perspective concerns the belief that each I / unique consumer life style has its own unique product I ) j)assortment [Cosmas 1982]. Psychographic dimensions provide / Exthe basis for the isolation of seven distinct groups. 1 fUnfortunately, they provide little valuable information when i f the task is switched to one involving the prediction of 1\ corresponding product assortment purchases. Since there is no theoretical linkage between broad-based psychographics and the purchase of a unique bundle of goods, this failure might also have been anticipated. Each study addressed in this section has evaluated psychographics from one of two perspectives. They have either highlighted a measurement or methodological deficiency or they have documented a failure to establish a reasonable relationship between psychographics and buyer behavior. Table 2-2 summarizes the results of these studies. None of the studies address the notion [Nelson 1971; Pernica 1974] that the relationship is indirect and, therefore, mediated by some set of intervening variables. Perhaps this shortcoming can explain the apparent lack of a meaningful linkage between psychographics and behavior. This possibility provides the impetus for the review of the literature which supports the existence of such intervening variable sets. These intervening variables reflect the next area for review. SUMMARY AND NEW DIRECTIONS The preceding sections have highlighted the dilemma regarding marketers' use of psychographics. The evidence indicates that there is considerable support for the concept yet it also highlights the basic implementation problems. The inconsistencies of these research efforts underscore the need to develop a theoretical basis which explains the FIRST AUTHOR TABLE 2-2 DISAFFIRMATIVE RESULTS Adams 1982 Bernstein 1978 Blair 1976 Cannon 1982 Cosmas 1982 Crask 1980 Demby 1977 Felson 1978 Kassarjian 1971 Kinnear 1976 Villani 1975 Wells 1975 Wells 1985 SUBJECT ___ RATIONALE SpeEIfic products Redundancy, validity Segmentation Non-exclusivity General Statistical imprecision Media choice No theoretical basis Product No Theoretical basis assortment Media exposure No Theoretical basis General Imprecision General Imprecision General Imprecision Gasoline No Theoretical basis TV viewing Sampling deficiency General Ambiguity, reliability General Short-term, abstract TABLE 2-3 DISAFFIRMATIVE RESULTS MEASUREMENT ISSUES FIRST AUTHOR SUBJECT RATIONALE Adams 1982 General Validity Boote 1981 General Reliability Burns 1979 General Reliability Burns 1982 General Reliability Forrest 1981 General Standardization, reliability Kassarjian 1971 General Validity Lastovicka 1982 General Reliability Martin 1982 General Reliability Moran 1974 General Reliability Yuspeh 1984 General Reliability nature of the influence that psychographics exerts on buyer behavior. The next three sections summarize those research efforts germane to the model delineated in chapter one. Each section provides a look at those efforts which provide support for one of the linkages proposed to exist between psychographics and buyer behavior. Consequently, support for various components of the model is outlined; this support leads to the development of the three research hypotheses. THE PSYCHOGRAPHICS-PRODUCT USAGE SITUATION LINKAGE AS discussed earlier, the basic premise of this research is that the relationship between psychographics and buyer behavior is mediated by two sets of intervening variables. The first of these intervening variable sets influenced by psychographics is the product usage situation. A product usage situation has been defined as an objective, unique to a particular situation, which an individual seeks to achieve from some Specific mode of behavior. Examples mentioned earlier include livening up the party and conveying status. The key question explored is: Does psychographics influence the set of product usage situations with which the consumer is confronted? '5: To answer this question, two broad areas are reviewed. First, psychographics oriented studies which imply that such a relationship exists are documented. Second, the situational literature is reviewed. These two areas provide the framework for verifying the existence of product usage situations and the influence psychographics exerted upon them. Among the limited set of product usage Situations which dominate the literature are economic issues, functional requirements, and social benefits. The need to address such product usage situations may be operationalized by consumers via the search for low cost alternatives [Hornik and Feldman 1982]; a product which helps an athlete cool down [Semenik and Young 1980]; or for an alternative which projects a positive image [Levy 1959; Levy 1963; Cosmas 1982]. The belief that the conveyance of a positive image or status may assume the role of a product usage situation implies that one prominent need is the perpetuation and demonstration of one's life style [Levy 1959; Levy 1963; Cosmas 1982]. The economic benefit is closely related to a product's ability to provide a consumer with the opportunity to reduce cash outlays. For instance, the do-it-yourself segment perceives the product usage situation as one of saving money [Hornik and Feldman 1982]. Other product usage situations relate specifically to the basic benefit associated with consumption. One purchases a pain reliever to address a specific product usage situation; the need to eliminate a headache. Additional support for the product usage situation concept is provided with the belief that there are variables which are "responsible for the selection and maintenance of the ends or goals toward which we strive and, at the same time, regulate the manner in which this striving takes place" [Gutman and Vinson 1979]. Such ends and goals may assume the form of product usage situations. Furthermore, Bushman [1982] discusses a taxonomy of life styles which culminates with a focus on specific buying behaviors. This taxonomy includes a component analagous to the product usage situation. The presence of mediating variables is further supported in the literature by the concept of Situational variables. The situation in which the consumer finds himself, in conjunction with his psychographic character, is important in the process of determining and prioritizing the set of product usage situations which the consumer faces. Thus any attempt to explain buyer behavior is best effected when the analysis transcends the basic evaluation of consumer characteristics [Belk 1975; Mausner 1963; Leigh and Martin 1981] as it is believed that such situations have a systematic, demonstrable impact on consumer decision making [Kakkar and Lutz 1975; Belk 1974]. The need for a situational perspective is summarized by the belief that consumer decisions depend on "stimulus situations" and that these decisions are "specific to the situation" [Mischel 1968]. Barker [1975] indicates that the situation in which a consumer finds himself coerces him to respond in accordance to that situation's unique set of circumstances. As an illustration of this type of influence, Belk [1975] highlights the situation where a person attending an auction turns out to be the high bidder for a cloCk which he would not have considered purchasing during an earlier visit to a department store. Leigh and Martin [1981] provide a taxonomy of situations; this taxonomy includes the category of psychological evaluation responses, a category predicated on the belief that situations which confront the individual are based on a set of subjective criteria, possibly a set which includes psychographics. Additional support for the concept of situational variables is provided by Lastovicka [1979]; he indicates that shopping for oneself creates a set of buying motives different from the set which exists when the shopping task is the purchase of a gift. People who are pressed for time are less likely to shop at department stores [Mattson 1982]. Tatzel [1982] discusses psychographics specific to shopping. She indicates that independent shoppers value the ability to shop and that these consumers need to be provided with a non-distracting, task-oriented shopping environment. There is also an indication that different activities promote a need for different products [Huber, Holbrook, and Schiffman 1982]. For example, a skier may have a need for a beverage which will warm him, while a tennis player may need a drink which will cool him down. There are also different product usage situations associated with the purchase of performing arts tickets; there are social reasons as well as business reasons [Semenik and Young 1980]. While it is apparent that product usage situations exist, there have been only a limited number of attempts to systematically relate such situational influences to psychographics; however, these studies have provided an element of support for the existence of an intermediate linkage between consumer characteristics and consumer behavior. There are a number of studies which support the notion that it is the product usage situation and not the product which is inextricably related to life styles. The individual who is concerned with health, nutrition and physical fitness is often faced with a product usage situation whereby the primary need involves the reduction of sodium intake [Fannin 1984]. Similarly, a disorganized individual or a Simplicity-seeker might desire product alternatives that make cooking easier [Greeno and Sommers 1977]. Current literature also provides the basis for the delineation of product usage situations as well as the opportunity to observe their relationship with psychographics. Calantoni and Sawyer [1978] indicate that "value seekers" are often faced with a product usage situation which reflects the desire to save money. Their broader overview indicates that different life styles appear to result in the presence of different product usage situations. While no inference regarding causation can be made, associations between life styles and product usage situations are documented. V Further support for this concept is provided with the discussion of the premise that a consumer's life style is reflected not only by the goods and services purchased, but also in the way in which these products are used [Levy if1963]; alternatively stated, life styles influence product EUsage situations. Similarly, Holbrook and Hirshman [1982] argue that consumption is directed toward the pursuit of happiness; that is, products are purchased based upon their ability to accommodate fantasy seeking and provide "positive '\ feelings." Auld [1982] documents a relationship between life style and product usage situations as well. In the purchase of a new home, the desired functional benefit may be shelter, a hedge against inflation, a tax shelter, or a means of conveying status; the individual's psychographic profile is likely to determine the product usage situation which the consumer deems important. Johnson [1979] indicates that discriminating buyers (psychographic) often seek products which provide a high level of value (product usage situation). In yet another example, it is noted that convenience-oriented consumers who are innovative in financial transactions tend to be more prOne to use automated teller machines [Stanley and Mochis 1983], perhaps because it conveys a modern image or because of the need for banking services at times when the bank is closed. This section of the literature review has emphasized two points. First, it is apparent that the concept of product usage situations is conceivable; second, there is a modest level of support for the belief that there is a causal linkage between psychographics and product usage situations. Consequently, the initial research hypothesis may now be stated: PSYCHOGRAPHICS INFLUENCES THE PRODUCT USAGE SITUATIONS ASSOCIATED WITH THE PURCHASE OF A PARTICULAR PRODUCT. THE PRODUCT USAGE SITUATIONS-OBJECT BENEFIT LINKAGE This section of the literature review is devoted to those studies which reflect the existence of a relationship between product usage situations and object benefits. An object benefit has been defined as a basic benefit, associated with Specific marketing mix components, which exists or accrues in the exchange or consumption process. Object benefits refer to those needs—satisfying components of the marketing mix; consequently, they may refer to characteristics of the product, the provider, or any other element of the marketing strategy. A product may have a low price; it may be a fashionable color; or it may have a sexy appearance. Object benefits are closely aligned with behavior in the marketplace in that the consumer will evaluate the various elements of the marketing mix and consummate the exchange so that the resultant level of satisfaction is maximized. In this study, it is presumed that those object benefits which ultimately influence the actual purchase decision are driven by the purchaser's psychographic profile as mediated by the relevant product usage situations. Intuitively, this linkage appears to be reasonable. It is generally accepted that the situation in which the consumer is placed will influence that consumer's decision making process [Belk 1975]. Much of the literature on Situational influences reflects the existence of a relationship between the situation and purchase behavior [Leigh and Martin 1981]. This line of logic is buoyed by the belief that situational variables may explain variation in buyer behavior better than do those variables which reflect characteristics of the buyer [Ward and Robinson 1971]. Even marketing theorists Engel, Kollat, and Blackwell [1969] indicate that both individual and situational factors should be considered viable influencers of buyer behavior. However, Lutz and Kakkar [1975] indicate that the "situation, in and of itself, is not a powerful predictor of buyer behavior."' They do indicate, however, that the individual consumer's psychological processes and overt behavior are systematically influenced by situational factors. The question then concerns the nature of that influence. In this project, it has been proposed that there is a systematic relationship between product usage situations and behavior in the marketplace. It is proposed, however, that this relationship is mediated by a set of object benefits germane to both. If one accepts the presumption that product choice is influenced by the benefits derived from the purchase and use of that product, then the task becomes one of supporting the presumption that object benefits are influenced by product usage situations. This section addresses the key question which must be answered: "Is it the physical object that relates to life style or is it the benefits associated with that object?" [Cosmas 1982]. For example, an idividual compelled to convey status has a house guest; he might purchase a national brand product instead of his usual private brand because it conveys a more prestigeous image to that guest [Thorndike 1947]. The product usage situation has influenced the sought object benefit and the ultimate purchase decision. If the role of the object benefit can be determined, then the answer to Cosmas' question will follow. One perspective of that role has been provided by Pernica [1974]. In noting an apparent deficiency in earlier research, he indicates that there has been a lack of consideration for "primary needs and product benefits." This concept closely mirrors the product usage situation-object benefits linkage. A person who suffers from diarrhea (product usage situation) will seek a product that provides the (object) benefits germane to his problem; a person with an ulcer will seek a product which offers a different set of basic benefits. For a particular pharmaceutical product, benefits such as effectiveness, safety, ease of swallowing, strength, and price are viewed as sought benefits dependent on the product usage situation. Subsequent to Pernica's earlier criticisms, several recent research efforts have systematically tied product usage situations to object benefits. Slacks being purchased for a formal affair are less likely to be rejected because of their expense [Monroe, Della Bitta, and Downey 1979]. The athlete performing in cold weather might seek object benefits which provide warmth; conversely, when that same athlete plays tennis in hot weather, the sought drink will likely be one which affords the opportunity to cool off [Huber, Holbrook, and Schiffman 1982]. In other product usage situations, efficiency was a sought object benefit for hair driers; Similarly, menus written in French were viewed more favorably than a comparable English language menu in certain situations [Hanson 1972]. Berkowitz, Ginter, and Talarzyk [1977] evaluated the interaction between specific situations and automobile attributes; many of these attributes could be classified as object benefits. Though a causal relationship was not hypothesized to exist, the two model components were shown to be correlated to each other as well as to product choice. Similar conclusions were reached in two studies which evaluated the dynamics of the factors which influence the demand for beer [Bearden and Woodside 1977; Woodside, Bearden, and Closkey 1977]. A comparable scenario is presented by Johnson [1979]. As an example of the linkage between product usage situations and object benefits, the author refers to the desire for immediate satisfaction and its relationship with the preference for material goods. Immediate satisfaction represents a plausible product usage situation and the materialistic nature of the purchased product qualifies as a desirable object benefit; a causal relationship also seems reasonable in this example. Campbell has utilized the product usage situation- object benefits concept in marketing soup. The object benefit of reduced sodium is predicated upon the product usage situation reflecting a need for a healthier product [Fannin 1984]. The insurance industry caters to those who seek to satisfy the product usage situation of providing financial security; however, it is evident that there are several object benefits for the consumer to consider (eyg4, term, whole life, accident) and that the desired object benefit may be a function of life style. It is also likely that the sought benefit will be mediated by the relevant product usage situations (security for self, spouse, children, personal liability, etc.). Similarly, the need to make cooking easier is a product usage situation which results in the demand for food products which are quickly and easily prepared [Greeno and Sommers 1977]. In each of these studies, the sought object benefits are a function of the product usage situation. Gift giving constitutes yet another area where the situation seems to interject its influence. The gift giving situation might result in an effort to find a product which minimizes the inherent social risk while a similar product purchased for one's own consumption often results in the desire for a product which minimizes financial risk [Vincent and Zikmund 1976]. In analyzing recent changes in American life styles, it is apparent that new product usage situations and new priorities are emerging. Self protection is a good example [Wettenstein 1982]; a weapon intended for self protection will probably require a set of object benefits much different from the weapon used for hunting or target shooting. Strategic elements other than the product variable have implications in this linkage as well. Price is a very important consideration for some consumers. The discriminating buyer (psychographics) will seek a product which provides a high level of value (product usage Situation); this consumer might ultimately purchase a low priced alternative (object benefit) [Johnson 1979]. Of course there are many other object benefits which might be Satisfactory in this particular situation. The need to reduce risk may also result in the preference for value-oriented object benefits. A low price may give way to free merchandise, a premium, or some other benefit deemed important to the consumer [Raju 1980]. The product usage situation-object benefit dyad has also influenced the retailing environment. The amount of time available to the consumer influences the attributes deemed important and sought by the consumer [Blackwell 1980]. Besides influencing their product assortments, retailers are forced to make operational decisions based upon this phenomenon. Object benefits associated with the retailer may be given more consideration than the benefits associated with the needed products. For example, direct marketers have the opportunity, through electronic catalogs, to give consumers who prefer to spend more time at home an easier mode of shopping. This section has provided a summary of those research efforts which denote that a relationship between product usage situations and object benefits exists. This fact provides the basis for the statement of the second research hypothesis: PRODUCT USAGE SITUATIONS DETERMINE THE OBJECT BENEFITS SOUGHT BY THE POTENTIAL BUYER. THE PSYCHOGRAPHICS-OBJECT BENEFITS LINKAGE The following research efforts document the existence of the relationship between psychographics and object benefits. It differs from those research efforts which SUpport the presence of the product usage situations, thereby reflecting the second unique path in the proposed model. This section provides basic support for the third and final research hypothesis. The basis for this linkage is the belief that the consumer's evoked set will consist of those products which fulfill the benefits deemed most desirable by the consumer. This belief is consistent with the presumptions germane to hypothesis two; it differs in that these object benefits are presumed to be directly influenced by the consumer's psychographic make-up. The concept of object benefits is illustrated by potential purchasers of men's cologne. Specific attributes such as a sweet smell or a refreshing feeling may be attractive to the consumer. It is evident that object benefits do exist; the problem revolves around the determination of the factors which influence the sought Object benefit set. Specifically, the question of whether Object benefits can be tied to psychographics must be addressed. One theoretical basis for this linkage is predicated upon the self-image concept. An individual's perception of himself influences the goals which that individual will seek to accomplish [Kassarjian 1971; Landon 1974]. Thus, the person who views himself as materialistic is likely to exhibit a preference for products which reflect materialism. It is the congruence between one's self-image and the goals one seeks to achieve which presumably impacts consumer decisions [Grubb and Stern 1971; Dolich 1969; Hamm and Cundiff 1969; Grubb and Hupp 1968]. It is reasonable to believe that these goals will specifically relate to sought object benefits; it is also reasonable to believe that one's self-image is partially based upon psychographic considerations [Snygg and Coombs 1949]. Within this paradigm, it has been noted that a consumer's self-image, which is influenced by psychographics will coincide with the needs-satisfying components of the most preferred product or brand [Dolich 1969]. In essence, a product could be seen as possessing "product psychographics." Thus, this linkage is typified by the need to match "people psychographics" to "product psychographics" [Demby 1976]. An example of this is the conservative consumer who seeks products which possess conservative object benefits. This concept has been operationalized by Kodak through their development of "brand personalities.“ Such a personality is established and nurtured by incorporating object benefits within a product's promotional theme with different themes being associated with different products. These products are then targeted toward specific market segments which have been delineated on the basis of psychographic variables [Emotions 1985]. The strategies used to market cameras and radios are consistent with Levy's [1959] assertion that a product will be used and enjoyed when it coincides with the consumer's self-perception. The individual who seeks to attract attention will often seek products with attention-attracting object benefits. This overt action may be predicated upon the consumer's perception of his or her own psychographic profile. The individual who sees himself as moving up the social ladder may want to demonstrate this belief. To accomplish this task, he seeks products which possess object benefits which signify that "he has arrived" [Segal and Sosa 1983]. Psychographics has also been used to define segments in the radio market. The sought object benefits differed from segment to segment, thus providing practical support for the proposed linkage [Sands and Warwick 1981]. These research efforts have provided support for the existence of object benefits and the belief that the benefits sought by the consumer are influenced by psychographics. They provide support for the statement of the third research hypothesis: PSYCHOGRAPHICS DIRECTLY INFLUENCES THE OBJECT BENEFITS SOUGHT BY THE CONSUMER. OVERVIEW OF THE LITERATURE REVIEW As mentioned earlier, the review of the literature delineated a number of endorsements and caveats for the use of psychographics in marketing. The approach appears to be conceptually appealing but difficult to operationalize. This difficulty underscores the need to seek new directions in the field of psychographic research. The direction taken here begins with the belief that psychographics does have a role in influencing buyer behavior. It departs from tradition, however, by presuming that that role is filtered through some set of intervening variables. The literature review endorses the presumption that product usage situations and object benefits mediate product choice. As a result, three research hypotheses have been developed. The focus now turns to the testing of those three hypotheses. C H A P T E R T H R E E METHODOLOGY AND AN ILLUSTRATIVE EXAMPLE OF THE MODEL INTRODUCTION This chapter consists of an enumeration of the methodology used in this study and an illustrative example of the proposed model. The methodology is divided into three distinct phases; it began with qualitative research, continued with a second, more directed review of the literature, and concluded with quantitative analyses. The model under scrutiny has been developed as a consequence of the literature review; the research supporting this model was summarized in chapter two. METHODOLOGY FOCUS GROUPS The first phase of research required the use of focus groups in an effort to ascertain the psychographic dimensions which seemed to impart some influence on behavior specific to photographic goods and services. In an attempt to assure that behavior indigenous to any certain geographic sector of the domestic marketplace was not overlooked, focus group interviews were undertaken in Atlanta, Georgia; Los Angeles, California; and Minneapolis, Minnesota. A total of six group interviews, two in each city, were performed. In each city, participants were enlisted via the employment of a local marketing research organization. Each respondent was asked to bring to the interview the pictures which they had most recently had developed. These pictures provided the basic framework for reviewing behavioral issues as well as a common bond which facilitated interaction among the group members. Participants were asked to select a picture that they had enjoyed taking and one which they enjoyed viewing. They were then directed to pass the pictures to one of the other group members who was in turn asked to project why the colleague had enjoyed taking or viewing the selected photographs. This procedure allowed for the evaluation of the influence of general psychographics on consumer behavior. The product usage Situations were also discussed. Participants were asked to indicate the reasons for taking and viewing photographs. What was the photographer seeking to accomplish? Then the object benefits associated with photography were delineated; of interest were those issues related Specifically to photography-oriented goods and services. Copies of the focus group question guidelines are included in the appendix. The transcripts of the interviews were content- analyzed. These analyses allowed for the determination of those psychographic dimensions which appear to exert some influence on behavior specific to the market for photography—oriented products. Once these dimensions had been delineated, the second phase of the research was initiated. QUESTIONNAIRE For each of the psychographic scales identified, there was an attempt to locate a study which had previously measured that psychographic variable. It was felt that this approach would prove superior to the process of developing "homemade" items to measure the Specific psychographic variable. For those variables where the review yielded insufficient or otherwise unsatisfactory items for measurement purposes, specific items were developed. These procedures allowed for the construction of the first section of the questionnaire; this section contained a battery of general psychographic statements. Upon completion of this section, the individual items were randomized. The second major section of the questionnaire consisted of the product usage situation statements. These statements provided the foundation for the initial causal link. Statements reflecting attitudes toward taking and viewing pictures were developed. These statements reflect more specific behavioral factors and were derived from the initial set of psychographic scales as well as from the focus group interviews. The underlying presumption here is that, for instance, an individual who is characterized as materialistic will also exhibit materialistic tendencies when viewing and taking pictures. The third section further narrowed the focus through the development of object benefits statements. These statements were also based upon the psychographics foundation established in section one. The scope of the study in this section involved more specific issues regarding the selection of cameras and camera equipment. As such, it established the basis for evaluating the second causal link; one reflected by the belief that the object benefits are a function of the product usage situations. For example, the rationale is that individuals who experience materialistic tendencies in reference to the relevant set of product usage situations will manifest these tendencies when deciding on a specific camera. Sections one and three provide the basis for the evaluation of the third research hypothesis. This hypothesis reflects the belief that there is a causal linkage between psychographics and the object benefits. Thus, it ignores the possible influence of the product usage situation. Each statement in the first three sections was evaluated by using a five-point Likert-type scale. This format produced intervally scaled data which facilitated subsequent data analyses. Also, it is believed that the use of a five-point scale provided adequate discrimination while retaining a relatively high degree of reliability [Boote 1981]. Reliability was further enhanced by labeling all points rather than only labeling the extreme points [Boote 1981]. The final section elicited information on a number of consumer demographics. Included were education, marital status, number of children, age, and income. PRETEST The completed questionnaire was then pretested. A convenience sample of 20 college students was selected for this task. Each student responded to approximately half of the questions on the tentative data collection instrument. Of concern was the elimination of non-discriminating items and the clarification of ambiguous terms and instructions. The resultant document was then readied for distribution via the mail. SAMPLING PROCEDURES The questionnaire was sent to a sample of 400 residents of Grand Rapids, Mighigan. Grand Rapids was chosen because of its relatively close approximation of national demographics [Szorek 1982]. The probability sample was drawn from the telephone directory. Prior to mailing the questionnaire, each potential respondent was phoned and agreement to participate was solicited. To further enhance the response rate and to speed the return of the completed questionnaire, $100 was allocated for a lottery; each respondent who returned the completed questionnaire by the stated deadline was eligible for the drawing. A postage-paid envelop was included for return mail. It was expected that each of these tactics would have a positive impact on the response rate. DATA ANALYS I S The initial analytical step involved an assessment of the reliability of the variable set. Cronbach's alpha is considered by many to be the most appropriate measure of reliability [Churchill 1979]. Since coefficient alpha indicates how well the sample items capture the essence of the construct being measured, a high value was desirable. Each scale was evaluated independently and those scales which exhibited a low level of reliability were treated in one of two ways. Items within individual scales were dropped if their deletion resulted in a sufficient increase in the value of alpha; when such modification proved unproductive, the entire scale was deleted from consideration in the subsequent data analysis procedures. The next step required the evaluation of the relationships among the variables within the three primary variable sets. Since there were multiple variables in each set, a confirmatory factor analysis was used to generate the correlation coefficients among the latent variables within the measurement model. Hunter's [1982] multiple groups algorithm was used to estimate these parameters. The initial analysis was performed to supplement the reliability analysis. The confirmatory factor analysis technique was used to evaluate those psychographic scales which met the standards of reliability and qualified the variable for the subsequent data analysis techniques. The procedure was duplicated, but all items relevant to a specific psychographic variable were included. The items used to measure the psychographic variables as well as their attendant product usage situations and object benefits were included. The final confirmatory factor analyses were used to generate the coefficients for the latent variables used to assess the structure of the model as well as the hypothesized causal relationships. The resultant correlation matrices facilitated the analysis of the covariance of the model components; an approach recommended by Box [1949] was used. To test the equality of the psychographic, product usage situation, and object benefit covariance matrices, a principle components analysis using varimax rotation was performed on each of the three resultant correlation matrices. This procedure was replicated on a fourth matrix which reflected the pooled estimate Of the common covariance matrix. This allowed the modified general likelihood ratio to be calculated. The basic premise under scrutiny in this analysis was the belief that each model component was measuring a different phenomenon but that the relationships among the variable dimensions germane to each component were equivalent. The efficacy of the model was further evaluated via the manipulation of the results of the confirmatory factor analyses. As mentioned earlier, one iteration involved the enumeration of three "groups." Each group represented a model component for one of the retained psychographic variables. Three groups were Specified for each retained variable. The second iteration aggregated all items across all three model components for each retained psychographic variable. Using the factor loadings for items assigned to each latent group and the correlation coefficients between the latent groups, the correlation matrix for all retained items was recreated. The actual correlations were compared to these predicted correlations. The a priori expectation was that the analysis resulting from the less aggregated model would yield a closer fit. This fit was tested by calculating the sum of squared differences between the actual and predicted correlation coefficients for each pair of retained items. ‘5- u b To measure the extent of the relationships between the variables within and among the three variable sets, a similarity matrix was calculated. This technique addresses the extent to which two variables are correlated, not with each other, but with the same variables [Anderson and Gerbing 1982]. This technique, in conjunction with the final confirmatory factor analysis, provided the basis for the assessment of the measurement model [Anderson and Gerbing 1982]. Next, the issue of causation was addressed. The hypothesized causal relationships were more difficult to measure due to the presence of multiple variables in both the dependent and independent variable sets. This problem was overcome with the application of path analysis. Path analysis allowed for the specification of causal influence indicators which correspond to the hypothesized causal relationships implied by the model [Harrell et a1. 1980]. The path analysis algorithm, PACKAGE [Hunter e£_§l. 1982], was used. Path analysis was used to explore each link in the path consisting of psychographics, product usage situations, and the object benefits as well as the linkage between psychographics and the object benefits. Since there is a hypothesized causal relationship for each of the three fl» linkages, the path coefficients were instrumental in the task of assessing the strength of the Observed series of relationships. The statistical significance of the path coefficients was evaluated by developing single-value measures for each multi-item scale. Then simple regression analysis was applied to each pair of scales specified by the model and subjected to a path analysis. The resultant coefficient of determination provided the basis for calculating t-statistics and measures of significance. Two additional, plausible model configurations were also tested. The first configuration ignored the linkage between psychographics and object benefits; the second presumed that there was no linkage between psychographics and product usage situations. Each of these analyses was performed on each of the retained psychographic dimensions; thus, the initial focus addressed only one dimension at a time. The best model was determined by again comparing the actual and predicted correlations among the three model components. With the best mode1(s) identified, the focus shifted. A comprehensive model which was comprised of all retained psychographic dimensions was subjected to a path analysis. The question addressed concerned the efficacy of a one- dimensional model versus a multi-dimensional model. THE MODEL: AN ILLUSTRATIVE EXAMPLE The initial path of the proposed model reflects hypotheses one and two. It is predicated upon the belief that psychographics determines the product usage situations with which the consumer is confronted. It continues with the presumption that the sought object benefits are similarly determined by the existing set of product usage situations. The second path reflects the third research hypothesis. This hypothesis (path) reflects the belief that psychographics also influence object benefits; it ignores the product usage situations. The model is depicted below. PRODUCT OBJECT BUYER l PSYCHOGRAPHICS H1 USAGE H2 BENEFITs_ _ _ flBEHAVIOR SITUATIONS H3 For the purpose of eXplaining the model, a hypothetical example has been developed. One psychographic variable which might influence buyer behavior is provincialism. Provincial individuals might find themselves in a product usage situation which reflects the desire to exhibit patriotism and love of country. Consequently, these consumers might seek products with object benefits which connote patriotism. Product-oriented object benefits such as being produced domestically or painted in one's national colors might be sought. This scenario is consistent with hypotheses one and two. The second path, reflecting hypothesis three, implies that provincialism directly influences the set of object benefits deemed salient by the consumer. In both paths, the ultimate purchase decision is influenced by the extent to which the competing product Offerings provide the requisite object benefits. This Specific scenario is depicted below. DESIRE TO DOMESTId PRODUCT PROVINCIALISM H1 EXHIBIT H2 PRODUCT CHOICE PATRIOTISM H3 C H A P T E R F O U R RESULTS INTRODUCTION Chapter Four provides a summary of the results of the data analyses. It begins with a comprehensive overview of the qualitative research efforts. More specifically, the initial sections of this chapter address the focus group studies, the development of the measurement scales, and the development of the data collection instrument. The latter sections summarize the results of the quantitative research procedures. This summary includes the assessment of reliability of the measurement scales, the assessment of the measurement mOdel, the evaluation of the structural model, and the analysis of other plausible model configurations. QUALITATIVE RESEARCH EFFORTS The initial thrust for this research project involved the identification of those psychographic variables which impart some influence on the consumer's purchase decision. This task was achieved by analyzing the results of the six focus group interviews. For each psychographic dimension which was identified, corresponding sets of product usage situations and object benefits were developed. FOCUS GROUP STRUCTURE As mentioned earlier, a total of six focus group interviews were undertaken. Each participant was asked to bring the set of photographs which they had most recently had developed. The intent was to question each participant so as to determine the psychographic dimensions which impart some influence upon the individual's decision to take a picture, look at a picture, and purchase photography equipment and supplies. In the first two focus groups, each participant was asked to eXplain his or her motivations for taking a particular picture. It became apparent that some participants were unwilling to express their true feelings. This lack of candor necessitated a change in format for the four subsequent focus group interviews. For the final four groups, each participant was asked to pass a favorite photograph to the individual to their immediate left. That person was then asked to project the picture-taker's motivations. The questions "why do you think he (or she) took this picture" and "why do you think that he (or she) enjoys looking at it" were asked. It is believed that this approach provided more valid results. Respondents were more willing to express ideas about others than they were about themselves. For example, men would readily indicate that one of their male counterparts took a picture because of love, yet they had generally refrained from making that statement about themselves. This projective technique allowed the extraction of sensitive information without giving the appearance of being overly personal. As a consequence of the need for specific information, the procedures were somewhat more structured than is typically the case. The free-form, open forum approach was modified. Each respondent was directly asked for input on each issue of interest. Furthermore, no attempt was made to disguise the purpose of the research. FOCUS GROUP RESULTS The focus groups proved instrumental in the task of determining those psychographic dimensions which influence behavior in the photography marketplace. Table 4-1 provides a listing of the psychographic dimensions which were identified. Once the focus group studies had been completed, the next task was the development of the mail questionnaire. THE QUESTIONNAIRE The initial task involved the development of the scales which would facilitate the measurement of each of the seventeen psychographic scales identified in the six focus TABLE 4-1 INITIAL PSYCHOGRAPHIC DIMENSIONS -CREATIVITY [Derived from Yankelovich 1977] -SENSUOUSNESS [Derived from Yankelovich 1977] ~INFORMATION-ORIENTATION [New, based on Naisbitt 1982] -HIGH TECH/HIGH TOUCH [New, based on Naisbitt 1982] -PERSONAL CONTROL [Gurin, et a1. 1969] -SOCIABILITY [Douglas and Urban 1977] -PLANNING [New] —NOVELTY/CHANGE [Derived from Yankelovich 1977] —MULTIPLE OPTION [New, based on Naisbitt 1982] -TECHIE JUNKIE [New] -MATERIALISM [Campbell 1973] -TRADITIONALISM [New] -NATURALISM/REALISM/HONESTY [New] -SIMPLICITY/CONVENIENCE [NeW] -EGO/SELF-ESTEEM [New] -ECONOMIC RATIONALIZATION [New] ,:SELF-HELP [New, based on Naisbitt 1982] flu. . 01. V. Q‘s .. . «a nu. «Au «3‘ Q ’n ml group interviews. Where possible, the scale items were gleaned from previous research efforts. When the review of the literature failed to provide a reasonable set of items for the measurement of one of the identified psychographic variables, a new set was developed. The next task was the development of those items which reflect the product usage situations. Each product usage situation statement was developed with one of the aforementioned psychographic scales in mind; two to four product usage situations were developed for each psychographic variable. This line of logic continued to serve as the basis for the development of the object benefits section of the questionnaire. Consequently, each object benefit statement was similarly tied to one of the psychographic variables. Again, two to four object benefit statements were developed for each psychographic variable. These three sets of variables provided the basis for the evaluation.of the proposed model. Table 4-2 illustrates the Specific items for each of the three primary model components for one of the identified psychographic variables: materialism. These three model components are presumed to reflect the same psychographic dimension; however, they are also presumed to measure a unique component of that psychographic dimension. In other words, the components are not TABLE 4-2 EXAMPLE OF THE MODEL USING MATERIALISM PSYCHOGRAPHIC ITEMS a) The only people who say money can't buy happiness are those people who never had a chance to try. b) If I'm lucky enough to buy the good things in life, I'm going to enjoy them. c) Material possessions are superficial when compared to the real joys of life. PRODUCT USAGE SITUATIONS a) I deserve the best camera money can buy. b) I like to take pictures of luxurious settings. OBJECT BENEFITS a) I always buy the least expensive picture frames. b) If I was buying a camera, I would buy it at the store with the best services regardless of price. c) I like expensive looking prints best. TABLE 4-3 PSYCHOGRAPHIC SCALES FOR INITIAL ANALYSIS CREATIVITY* DECENTRALIZATION ECONOMIC RATIONALITY EGO/ESTEEM* EMOTIONAL HIGH TECH/HIGH TOUCH INFORMATION ORIENTED MATERIALISM MULTIPLE OPTION NOVELTY/CHANGE PLANNING PERSONAL CONTROL SELF-HELP SENSUOUSNESS SIMPLICITY* SOCIABILITY* TECHIE-JUNKIE* '7’) redundant; each component provides unique information. This presumption is addressed and tested in the quantitative phase of the research process. A similar set of items was developed for each of the psychographic variables used to test the model. The questionnaire also contained questions on market behavior and consumer demographics though these variables did not fall within the domain of the proposed model. PRELIMINARY PROCEDURES The initial questionnaire was pretested by a sample of twenty undergraduate students. Non-discriminating items were deleted, and questions which were identified as having biased or unclear wording were modified. The revised instrument was then readied for dissemination to a sample of 400 potential respondents. THE QUANTITATIVE RESEARCH PROCESS The following sections provide an overview of the quantitative research process. They outline the sampling results as well as the analytical procedures applied to the resultant data set. Consequently, these sections provide the basis for the evaluation of the three research hypotheses; these results will allow the evaluation of the proposed model. The results are summarized in this chapter; conclusions are drawn and presented in chapter five. SAMPLE RESULTS Upon completion of the pretest and minor modifications to the data collection instrument, it was disseminated via first class mail to the selected sample of 400 potential respondents. As mentioned earlier, the questionnaire was lengthy (320 total questions), and specific actions were undertaken to ensure that an acceptable response rate would be achieved. Advance agreement to participate was attained from each sample member, and each respondent was given a chance to win $100. Of the 400 questionnaires mailed, 242 usable returns were received by the stated deadline. The response rate of 60.5% greatly exceeded industry standards [Mautz and Neumann 1970]; it also reduced the potential impact of nonresponse error. Age was used as a validation variable as a means of evaluating the representativeness of the sample. A chi- squared test indicated that the difference between the respondent group and the general population was not statistically significant. PSYCHOGRAPHIC SCALES The inherent problem associated with the lack of reliability of psychographic scales was evident in this study. Of the 19 scales initially considered for further analysis, only five exhibited a level of reliability high enough to merit their inclusion in the testing of the three hypotheses. Measurements of reliability were accomplished by using Cronbach's alpha; the alpha coefficients for the original 19 scales ranged from .037 to .780. The modest results were somewhat surprising since most of the scales had been used in previous research studies. The initial scales considered for further analysis are shown in Table 4-3. The scales with asterisks met the standards established for coefficient alpha for inclusion in the remaining data analysis procedures. The results of the Cronbach's alpha test of reliability for each of the retained scales are detailed in Table 4-4. For those scales which failed to meet minimal standards for reliability, a comparable presentation of the results is presented in the appendix. For eXploratory research, an alpha value of .7 is deemed acceptable [Nunnally 1978]. More modest levels of reliability are viewed as an acceptable tradeoff for the strenuous methodological requirements needed to attain very high levels of reliability [ibid.]. Each scale was analyzed; item deletion scores were used to help determine the final composition of each scale as specific items were dropped when significant gains in reliability resulted from their deletion. Only those scales which exhibited an alpha coefficient of at least .5 were retained for further analysis. COEFFICIENT ALPHA FOR EACH RETAINED SCALE SCALE ALPHA IF ITEM DELETED CREATIVITY -A person doesn't have to be an artist to be creative. .52 -I feel the need to be creative. .20 -Hobbies are an expression of creativity. .12 COEFFICIENT ALPHA FOR ENTIRE SET: .37 DECISION: INCLUDE USING ONLY ITEMS 2 AND 3. EGO/ESTEEM -I feel I'm a person of worth, at least on an equal basis with others. .18 -I am able to do things as well as most other people. .05 -I feel I have few qualities. .56 COEFFICIENT ALPHA FOR ENTIRE SET: .24 DECISION: INCLUDE USING ONLY ITEMS 1 AND 2. SIMPLICITY -I try to find shortcuts to everything I do. .54 -I always try to do things in a way that saves time. .48 -I prefer many of my household products to be as compact as possible. .52 —I often use modern conveniences to reduce the frustrations of many daily tasks. .57 COEFFICIENT ALPHA FOR ENTIRE SET: .60 DECISION: INCLUDE USING ENTIRE SET. SOCIABILITY —I am always glad to join a large gathering. .58 -I find it easy to mingle among people at a social gathering. .57 -I consider myself to be a sociable, outgoing person. .55 COEFFICIENT ALPHA FOR ENTIRE SET: .78 DECISION: INCLUDE USING ENTIRE SET. TECHIE-JUNKIE -I have more electronic products than the average person. .62 -New electronic products are fascinating. .59 -I like to buy products with the newest technology. .57 -I need to own the newest gadgets. .64 COEFFICIENT ALPHA FOR ENTIRE SET: .67 DECISION: INCLUDE USING ENTIRE SET. TABLE 4-5 PRODUCT USAGE SITUATIONS AND OBJECT BENEFITS SCALE/ STATEMENT* CREATIVITY -I always try to add a creative touch when taking a picture. -I try to think of creative ways to display my pictures. -I would pay more for a camera that improves my creativity in picture taking. EGO/ESTEEM -It boosts my ego when I take a good picture. -I only keep pictures that I will be proud to show other people. -Give me a few minutes to study any camera and I can use it as well as anyone. SIMPLICITY -Taking pictures should be made simple. -I would look at my pictures more often if it wasn't such a bother to get them out. -The simpler a camera is, the better. SOCIABILITY -I like to take pictures at parties. -It is fun looking at pictures with old friends. -I like cameras which can be passed around at parties so everyone can take a picture. TECHIE-JUNKIE -I'm constantly seeking to buy photographic accessories with the newest technology to help me take better pictures. -The ability to Show still photos on TV is a super idea. -Although my old camera worked well, I've bought a new one because its technology was much more advanced. *For each scale, the first two items are product usage situations and the third is an object benefit. TABLE 4-6 FACTOR LOADINGS FOR THE 15 ITEMS WHICH MEASURE THE PSYCHOGRAPHIC SCALES FACTOR ITEM 1 2 3 4 5 3 53* 17 15 13 20 4 66* 32 31 21 9 8 40 41* 16 26 8 9 26 86* 20 25 24 7 14 16 50* 3 27 10 15 29 63* 15 22 13 2 21 54* 8 25 15 37 31 43* 23 25 14 13 21 18 57* 14 16 32 46 18 86* 25 17 28 40 16 78* 19 2 3 5 35 13 56* 5 41 19 31 19 60* 6 18 23 34 12 68* 11 -9 15 36 16 50* * correctly loaded .wa a: 355 Table 4-5 provides a detailed look at the product usage situation statements and the object benefits statements associated with each of the five retained psychographic scales. EVALUATION OF THE MEASUREMENT MODEL The coefficient alpha measure of reliability was used to evaluate the internal consistency of the basic psychographic dimensions. Such measures, however, are inadequate in the task of evaluating the overall dimensionality of the model. To accomplish this task, similarity coefficients for each retained item were computed. This technique facilitated the measurement of the latent variables prerequisite to the analysis of the causal relationships [Anderson and Gerbing 1982]. Analysis of the measurement model was also facilitated by the application of the results from a series of confirmatory factor analyses. These analyses provided input which allowed chi-squared statistics to be calculated. These statistics reflected the ability to recreate (predict) the original correlation matrix from the resultant factor loadings of the retained items and the factor correlations. SIMILARITY COEFFICIENTS Similarity coefficients provided the basic input for the evaluation of the measurement model. The similarity matrix was reordered to facilitate visual inspection; this procedure allowed the mesurement model to be specified by grouping items which share high similarity coefficients [Anderson and Gerbing 1982]. Ideally, the resultant groupings would approximate the a priori groupings specified in the original model development process. Once the variables were shown to "reasonably" fit the proposed model, the hypothesized causal relationships could be evaluated [Anderson and Gerbing 1982]. Within the framework of the reordered similarity coefficient matrix, the dimensionality of each psychographic scale was confirmed. The only non-conforming item was one used to measure simplicity. For the other four psychographic scales, all items within each scale exhibited the expected relationships. Because of the relatively high level of reliability of each scale, this result was anticipated. There was a general tendency for product usage situation items and object benefit items to be grouped together. The lone deviation from this tendency was within the sociability dimension; the two product usage situation items had a similarity coefficient of .28. Also, the object a .I ,A\ [1C] N'f It's- 4 nu It] n— \A‘0 .t-.! “.n 55' ll., ti‘; :a‘ ‘11 (I) benefit item failed to exhibit parallelism with any of the sociability-oriented items. This evidence supports the exclusion of sociability from the subsequent path analyses; however, the decision was made to subject it to the causal analysis. It was expected that the subsequent multiple groups and path analyses on the sociability construct would yield marginal results. The other four dimensions reflected a more viable means of evaluating the hypothesized causal structure of the model. The complete similarity matrix is included in the appendix. CONFIRMATORY FACTOR ANALYS I S The multiple groups algorithm in Hunter's [1982] PACKAGE was used to evaluate the five psychographic dimensions retained upon completion of the assessment of reliability for the original psychographic variable set. This evaluation consisted of three separate confirmatory factor analyses. The initial analysis included only the fifteen items which comprised the five basic psychographic scales. The second analysis included all thirty items; these items measured the psychographics, product usage situations, and object benefits. In each case, the analysis was performed along a single dimension; thus, five factors were specified. A third confirmatory factor analysis was Performed with fifteen factors specified. Each factor represented one of the three basic model components for one of the five psychographic dimensions. The results of this analysis provided the basic input required for further evaluation of the measurement model and the subsequent path analyses. The initial analysis reaffirmed the dimensionality of the five basic psychographic scales. Since each scale exhibited a relatively high level of reliability, it was eXpected that each item would load on the factor representing the psychographic construct that it was designed to measure. [All fifteen items loaded as expected on their respective psychographic "group." Table 4-6 shows the factor loadings for each of the fifteen items. The results provided support for the inclusion of all five scales in the next step of the data analysis process. The second confirmatory factor analysis also specified the presence of five factors. This time, however, the analysis was expanded to include all psychographic, product usage situation, and object benefit items germane to the five retained psychographic scales. This analysis provided a more wholistic perspective of the model. Ideally, each item was expected to load on its respective psychographic factor; however, some deterioration of the model was expected. Table 4-7 shows the factor loadings for each of these items. TABLE 4-7 FACTOR LOADINGS FOR THE 30 ITEMS WHICH MEASURE PSYCHOGRAPHIC, PRODUCT USAGE SITUATIONS, AND OBJECT BENEFITS FACTOR ITEM 1 2 3 4 5 3 49* 33 17 15 27 4 35* 20 33 18 12 21 53* 33 26 30 36 27 52* 32 29 36 40 31 45* 27 22 26 41 8 26 44* 12 25 6 9 19 69* 22 27 19 19 56i 22 23 21 12 25 16 26* -3 21 12 32 17 41* 8 23 28 7 16 17 53* 13 36 10 14 19 79* 31 20 13 12 16 50* 18 22 15 40 26 38 33 471 18 -7 5 26* 24 -1 26 10 10 18* 18 -15 29 34i 17 -18 6 17 14 19 21 28 65* 11 16 28 40 26 78* 23 17 24 30 25 59* 19 20 35 28 27 32 44i 24 28 31 36i 20 23 30 17 23i 14 15 20 2 5 13 34 23 47* 5 41 28 32 35 52* 6 34 30 32 22 68* ll 12 18 41 34 49* 22 47i 34 27 22 39 23 30 10 14 18 45* 28 36 15 33 21 36* * properly loaded i incorrectly loaded As a part of the multiple groups algorithm, coefficient alpha was calculated for each factor. The reliability measures for each of the five "psychographic" factors were .58, .48, .47, .58, and .68. This compares to initial values of .52, .56, .60, .78, and .67 for the psychographic scale items only. For three of the five factors, the level Of reliability dropped. For the remaining two, slight increases were observed. Of the thirty items under scrutiny, twenty—four loaded on the appropriate factor. Of the six improperly loaded items, three occurred on a single factor- sociability. It was the sociability dimension for which the earlier problems regarding dimensionality were documented. In general, the results support the belief that the product usage situations and object benefits are related to psychographics in some way. The third confirmatory factor analysis provided the basis for the subsequent analytical procedures. Each of the three basic model components was evaluated within the context of its relevant psychographic dimension. Since initially there were five psychographic scales under scrutiny, fifteengfactors (3x5) were Specified for the analysis. The factor loadings are shown in Table 4-8. However, the Objective of this procedure was the procurement of measures of correlation among the model components as Specified by the fifteen factors. TABLE 4-8 FACTOR LOADINGS FOR 15 GROUP, 30 VARIABLE SCENARIO FACTOR ITEM 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 3 53 35 l3 l7 58i 8 15 14 5 13 10 l 20 27 14 4 66* 19 5 32 4 -3 31 14 6 20 11 -7 9 l7 3 21 22 62* 38 7 57 23 18 -4 27 9 48 ll 21 52 21 27 31 62* 25 18 62 2 20 22 17 18 48 6 21 54 29 31 21 51 100* 1 69 10 14 —6 27 8 31 20 26 49 29 8 40 ll -4 41* 33 16 16 -2 -3 26 12 -2 8 5 -2 9 26 12 -l 86* 34 25 20 12 6 25 ll 6 24 9 -1 19 38 48 30 16 32 l3 l6 16 15 5 10 8 26 491 29 25 7 13 14 18 32* 15 -9 l3 5 10 8 26 8 15 5 32 6 20 10 30 44 100* 8 -15 12 5 33 17 23 32 13 7 14 88 ll 16 12 5 50* 46 -4 3 l9 8 27 32 26 10 15 12 -l 29 l O 63 90i—12 15 40 8 22 5 8 l3 2 16 6 21 -4 9 54* 36 -7 8 25 5 25 7 10 15 37 29 14 31 13 3 43* 17 -2 23 39 -l 9 23 18 18 3 -8 —10 19 l -16 40a 30 -31 18 ll 14 2 -7 0 26 -2 17 7 -3 27 7 14 30* 6 20 2 2 13 18 3 29 13 19 6 21 20 8 18 -41 100* l 15 -l 8 21 l7 14 13 19 6 21 20 2 18 51 l 57* 24 15 14 l 5 16 32 16 7 46 20 8 18 39 3 86* 22 9 25 6 12 17 28 14 5 40 11 1 16 50 -2 78* 12 0 19 12 7 20 2 52 21 12 37 19 30 -l 4 14 52* 21 37 43 19 24 17 29 11 20 35 16 30 23 12 13 52* -5 14 35 9 30 4 14 20 3 43 17 10 26 -l 11 15 100* 23 9 5 2 3 5 1 5 5 17 35 9 6 13 24 7 56* 25 12 5 45 24 17 19 28 16 31 ll 7 19 33 17 60* 29 14 6 18 29 24 23 37 9 34 5 3 12 19 12 68* 41 28 ll -9 20 18 15 9 11 36 34 4 16 37 18 50* 26 23 22 11 50 40 7 9 24 20 7 20 1 27 l 27 51* 29 23 14 37 10 1 17 9 13 12 1 8 22 8 26 51* 25 28 7 40 29 5 29 7 30 4 17 ll 26 5 33 52 100* * correctly loaded a incorrectlly loaded within correct psychographic dimension i incorrectly loaded ACTUAL VERSUS PREDICTED CORRELATIONS AMONG THE INDIVIDUAL SCALE ITEMS Before proceeding with the tests of the structural model and the three research hypotheses, one final evaluation of the measurement model was undertaken. By using the factor correlations and the factor loadings of those variables associated with each model component, a predicted correlation coefficient for each pair of scale items could be developed. ' The aggregated five group scenario was compared to the fifteen group scenario. The a priori assumption was that the fifteen group scenario would yield better predictions because of the unique character of the three basic model components. In reality, there was virtually no difference in the final results. The predicted correlation matrix was compared to the actual correlation matrix by evaluating the sum of squared differences between each entry in the 30x30 matrix. The sum of squares in the five group scenario was 5.37 while the corresponding statistic for the fifteen group scenario was 6.76. The two paradigms were compared by evaluating the difference in the chi-squared test statistics. These results are summarized in Table 4—9. TABLE 4-9 CHI-SQUARED TEST RESULTS AREA OF EVALUATION CHI-SQUARE DF SIG FIVE GROUP SCENARIO 5.37 358 n.5. FIFTEEN GROUP SCENARIO 6.76 411 n.s. COMPARISON OF SCENARIOS 1.39 53 n.s. The results for each of the three analyses were not statistically significant. These results imply two things. First, each of the two scenarios proved capable of generating a reasonable approximation of the actual correlation matrix. Second, neither scenario proved superior to the other in this task. While these results refute the initial premise that the fifteen group scenario would prove superior to the five group scenario, they are not particularly alarming. In essence, they lend some credibility to the measurement model as well as to a potential paradigm which evaluates behavior without distinguishing among the three model components. EVALUATION OF THE STRUCTURAL MODEL The structural model involves the delineation of the causal relationships which exist among the theoretical constructs [Gerbing 1979]. Before drawing conclusions about the three research hypotheses, however, the basic structure of the model had to be evaluated. These procedures are summarized in the following sections. TEST OF THE EQUIVALENCE OF THE COVARIANCE MATRICES For each of the three model components, a 5x5 correlation matrix was derived from the fifteen group confirmatory factor analysis. It was expected that the three matrices would Share a common structure. Alternatively stated, the relationships among the psychographic variables should approximate the system of relationships which exist among the product usage situation scales as well as the object benefit scales. To test the equivalence of the three covariance matrices, Box's modified generalized likelihood ratio statistic was used. This procedure required a principal components analysis to be run on each of the matrices as well as the matrix representing the pooled estimate of the covariance. The test statistic is a generalization of the Bartlett test of homogeneity of variances where the determinants assume the role of generalized variances. The incorporation of this statistic with Box's scale factor results in a quantity which is approximately distributed as a chi-squared variant. These results are presented in Table 4-10. TABLE 4-10 TEST OF EQUIVALENCE OF COVARIANCE MATRICES MC-l 3769 d.f. 30 significance .000 Clearly, the hypothesis of equivalence should be rejected. While this rejection is disappointing, it is not .a death knell for the model. These analyses have indicated that the matrices do have similar dimensional structures, but that the latent variables interact in different ways. That is to say that the product usage situation variables interact differently than do the psychographic and object benefits variables despite their apparent similar dimensional structure. It is the latter condition which supports the continued analysis of the proposed model. PATH ANALYSIS ISSUES Prior to proceeding with the evaluation of the prooosed model, two plausible alternative models were designated. One model ignored the hypothesized linkage between psychographics and product usage situations while the other ignored the hypothesized linkage between psychographics and object benefits. The evaluation continued with the application of path analysis for each of the three potential models. The initial point of interest was not the path coefficients, however, it was the ability to predict the level of correlation between the components of the model which was under scrutiny. As before, the sum of squared differences between the actual and predicted correlation coefficients served as the test statistic. A lower sum of squares would tend to reflect a model's enhanced viability in the task of delineating a system of causal relationships. Path analysis was performed on each model in each of the five retained psychographic dimensions. These results are presented in Table 4-11. Clearly model B, the configuration which does not designate a relationship between psychographics and product usage situations is not a viable alternative. In only one case out of five was the path analysis algorithm able to generate estimated path coefficients. In that one case, however, the attendant sum of squared deviations was very high (4.29). These facts lead to the decision to drop configuration B from further consideration. Model C does not include the causal linkage between psychographics and object benefits. This model appears to be a plausible one. An evaluation of the test statistics shown in Table 4-11 indicates that slightly lower values TABLE 4-11 EVALUATION OF ALTERNATIVE MODELS DIMENSION MODEL* SUM OF SQUARES Creativity A .01 B 4.29 C .01 Ego A .01 B _ C .07 Simplicity A .01 B _ C .04 Sociability A .00 B _ C .01 Techie-Junkie A .00 B _ C .00 *A is originally proposed model B is model which ignores PSY-PUS linkage C is model which ignores PSY-OB linkage - path analysis algorithm would not estimate parameters due poor fit tend to be associated with the original model. However, these differences are small and not statistically significant. One additional analysis was was performed as a means of comparing the two models. Path analysis was used to evaluate each model from a different perspective. Instead of looking at a single psychographic dimension, all five dimensions were incorporated concurrently. This analysis of the “supermodels" yielded results comparable to the earlier path analyses. For model A, the sum of squares was 11.40; for model C, the value was 9.84. This result would seem to support model C over model A: however, the support was not overwhelming. By testing model A, conclusions could be drawn about both potential models as each potential linkage was evaluated. The results of the evaluation of the supermodels' structure did indicate that the comprehensive model is not as viable as the model which scrutinizes a single psychographic dimension. This fact is documented by the comparatively high sum of squares associated with the two models. Consequently, the discussion presented in Chapter Five will primarily focus on the one-dimensional model. PATH ANALYS I S RESULTS Path analysis was used to evaluate the hypothesized causal relationships specified in the model. These analyses are based upon the correlation coefficients generated by the confirmatory factor analyses. A surrogate indicator of statistical significance was used to evaluate the relationship between components which are linked in the proposed model. Single index values were generated for each scale and simple regression was performed on the indices in accordance with the direction specified in the research hypotheSes.. The resultant coefficients of determination provided measures which reflect the significance of the observed relationships. The path coefficients and the coefficients of determination are presented in Table 4-12. Clearly, the highest path coefficients occur in the linkage between psychographics and product usage situations. Equally clear is the fact that there is considerable variation in the path coefficients from one psychographic dimension to another. For example, for the simplicity dimension the path coefficient is .94; for the sociability dimension the path coefficient is .28. The linkage between product usage situations and object benefits tends to exhibit the next highest path coefficient. In each of the five iterations, this linkage results in a coefficient which exceeds the corresponding psychographics- TABLE 4-12 PATH AND R-SQUARED COEFFICIENTS FOR THE MODEL LINKAGES COEFFICIENTS DIMENSION LINKAGEfl) PATH R-SQUARED SIG. CREATIVITY P-PUS .56 .19 .01 CREATIVITY P-OB -.10 .04 .01 CREATIVITY PUS-OB .50 .26 .01 EGO/ESTEEM P-PUS .63 .30 .01 EGO/ESTEEM P-OB .08 .09 .01 EGO/ESTEEM PUS-OB .39 .20 .01 SIMPLICITY P-PUS .94 .81 .01 SIMPLICITY P-OB .13 .01 .07 SIMPLICITY PUS-OB .59 .18 .01 SOCIABILITY P-PUS .28 .07 .01 SOCIABILITY P-OB .08 .01 .10 SOCIABILITY PUS-OB .13 .02 .03 TECHIE-JUNKIE P-PUS .62 .27 .01 TECHIE-JUNKIE P-OB .08 .ll .01 TECHIE-JUNKIE PUS-OB .48 .27 .01 l P Psychographic variable PUS Product usage situation OB Object benefit L) (J' \__.J- object benefits linkage. In every case but one, the path coefficient for the psychographics-object benefits linkage is less than .10. The results of the regression analyses support the results of the path analyses. The coefficients of determination exhibit the same tendencies as the path coefficients; the highest coefficients are found between the psychographic and product usage situation constructs; the lowest are found between the psychographic and object benefit constructs. Each of the fifteen coefficients was statistically significant at the .10 level with twelve of the fifteen significant at the .01 level. These results also help explain the results of the earlier evaluation of the alternative model configurations. The comparatively small values which summarize the linkage between psychographics and object benefits in conjunction with the relatively large values for those coefficients which summarize the psychographics-product usage situations linkage indicate that any model of relationships should 'retain the latter relationship. The smallest path coefficients occurred within the sociability dimension. It was this dimension which presented the most significant problems in the confirmatory factor analysis. This fact is notewothy as it indicates that the model is likely to deteriorate if all items relecting a single psychographic dimension are not systematically and theoretically linked together. With the path analyses now completed, conclusions regarding the structural model, implications, and recommendations can be discussed. This discussion is the focus of chapter five. C H A P T E R F I V E CONCLUSIONS AND RECOMMENDATIONS : . his INTRODUCTION The primary purpose of this chapter is to assess the results of the research project. Conclusions will be drawn regarding the theoretical constructs; each hypothesis will be discussed. Managerial implications will also be discussed. The extent to which this project contributes to the advancement of the marketing discipline is delineated. Limitations of the study are outlined and additional research is recommended. THE MEASUREMENT MODEL The well documented problems associated with the lack of reliability for psychographic scales were a major obstacle in this study. Of the nineteen scales originally considered, only five exhibited levels of reliability which were considered sufficient for inclusion in the subsequent data analysis procedures. Of those five, items had to be deleted in three of the scales in order to achieve the required values for coefficient alpha. This lack of reliability is certainly one reason that many previous research projects have fallen short in their efforts to document a relationship between psychographics and buyer behavior. The incorporation of the intervening sets of product usage situations and object benefits variables generally resulted in somewhat lower levels of reliability within a psychographic dimension. This fact implies that all variables within a given dimension should be developed in such a way that the aggregation of the three variable sets would result in a reasonable level of reliability. The dimensionality of the measurement model was evaluated via the analysis of similarity coefficients. This technique confirmed the basic design of the model constructs. The content of each component was generally substantiated by the ordered similarity matrix. The psychographic items for each of the five scales reflected unidimensionality. The model clearly grouped the items which measured the basic psychographic variables. The product usage situation items within each psychographic dimension tended to be associated with the same inferred construct; the lone problem area was within the sociability dimension. Since each object benefit construct consisted of a single item, there was no question of Similarity. It is noteworthy, however, that the object benefits items tended to exhibit a high degree Of Similarity with the relevant product usage situation items. The lone exception was within the sociability dimension. In the confirmatory factor analysis, problems were experienced in the loading of the elements within the sociability dimension. In general, there was a greater tendency for product usage Situation items and object benefits items to load incorrectly. Of the thirty items, six misloaded. Of those six, five were either product usage situations or object benefits. This problem was ameliorated when the confirmatory factor anaysis was expanded to reflect a fifteen factor domain. Utilizing this paradigm, the number of misloaded variables was reduced to three. Thus, the belief that the constructs were unidimensional was supported. The results of the confirmatory factor analyses and the analysis of similarity coefficients provided general support for the measurement model. Although these procedures suggested that the model not be tested on the sociability dimension, the decision not to delete it from the subsequent analyses was made. It was eXpected that path analysis on this dimension would yield the least significant results. The last procedure which was designed to evaluate the measurement model was the attempt to recreate the original correlation matrix from the information provided by the confirmatory factor analyses. As mentioned earlier, the initial analysis combined the variables used to measure each of the components within a psychographic dimension; thus, a five group analysis was undertaken. The second iteration was based upon the delineation of fifteen groups; there was one group for each of the three model components associated with each of the five retained psychographic dimensions. The a priori assumption that the fifteen group analysis would be better able to recreate the original input matrix was incorrect. The five group scenario resulted in a slightly smaller chi-squared test statistic, but the difference between the two scenarios was not statistically significant. These results are meaningful. First, they provide support for the basic model and its three components. The results of the multiple groups amgorithm utilizing the fifteen group scenario proved capable of providing the basis for the recreation of a close approximation of the original correlation matrix. The results also indicate that there is intragroup homogeneity among the scale items and intergroup heterogeneity among the scales. Each of the fifteen groups is unique in that each measures a different construct; however, the groups tend to exhibit high levels of internal consistency. Thus the indication is that the measurement model has merit and should be tested further. The fact that the five group scenario also satisfactorily recreated the original input matrix should also be addressed. Despite the earlier support for the presence of the unique model components, the aggregation of the items used to measure all three components is not devoid of merit. While it ignores the hypothesized causal linkages, it provides an approach for the marketing professional to incorporate the components into a more pragmatic view of buyer behavior. This combined approach could be especially effective for tasks such as the delineation of market segments. These results do not refute the viability of the proposed model; therefore, the analysis continued with the assessment of the structural model. THE STRUCTURAL MODEL Evaluation of the structural models required the specification of the causal relationships. Thus, the testing of the hypothesis was undertaken. Before doing this, however, there was an attempt to determine whether the proposed model configuration was the most appropriate to use. This determination began with an assessment of the structure of the covariance matrices for the three model components; it continued with the evaluation of alternative model configurations. The evaluation of the structure of the three covariance matrices was accomplished by applying Box's generalized likelihood-ratio statistic. The results clearly indicated that the three matrices were not equivalent. This outcome represented the most disappointing conclusion arising from the analysis of the data. The results imply that the correlations among the five psychographic scales are structurally different from the 102 correlations among the five psychographics—based product usage situations and object benefits. Ideally, the individual components should reflect unique measures; the earlier confirmatory factor analyses indicated that this is the case because the correlations between the latent variables were not exceedingly high. It was also expected that the interaction among the psychographics-based components would be comparable for each of the three basic components. Box's test statistic indicated that this is not the case. However, the results of the principal components analyses indicated that a similar dimensional structure seemed to exist for the three matrices. While the rejection of equivalence presents a limitation to this project, the belief that a similar dimensional structure pervades the three components provides the rationale for the continuation of the project. It also demonstrates the need to devote the same energy to the development of product usage situation and object benefit scales as one would devote to the development of the psychographic scales. Thus, the pretest of the questionnaire should be expanded to include an evaluation of the structural equivalence of the model components. The initial path analysis procedures were instrumental in the determination of several key issues germane to the proposed model. They delineated a weakness in the model which ignored the linkage between psychographics and product usage situations. Furthermore, they indicated that the model which ignored the hypothesized psychographics-object benefits linkage reflects a plausible set of relationships among the model components. Yet another fact determined by the initial evaluation of the path analysis routines was the fact that the model was much more efficient when a Single dimension was used to assess the system of relationships. The high sum of squared deviations associated with the comprehensive supermodels underscored this deficiency. .Each of these findings provided support for the continued evaluation of the proposed model. While there was some question regarding the hypothesized linkage between psychographics and object benefits, it was presumed that the planned analytical procedures would identify any weakness in this regard. Therefore, the focus remained on the proposed model, and the research hypotheses were evaluated. These results are summarized in the following sections. Hypothesis One: The first hypothesis stated that there is a causal relationship between psychographics and product usage situations. This hypothesis was confirmed. The five path coefficients reflecting this linkage range from a high of .94 to a low of .28. Each of these coefficients was statistically significant thereby providing support for the initial hypothesis. The .28 coefficient was associated with the sociability dimension. The lowest path coefficient for the remaining three dimensions was .51. Clearly, the aforementioned problems within the sociability dimension resulted in the anticipated reduction in the efficacy of the model. These results not withstanding, there appears to be a causal relationship between psychographics and product usage situations. Therefore, hypothesis one was accepted. Hypothesis Two: The second hypothesis stated that there is a causal relationship between product usage situations and object benefits. This hypothesis was also confirmed although the relationship was not quite as strong as the observed relationship between psychographics and product usage situations. The path coefficients ranged from a high of .59 to a low of .13. If the sociability dimension is discounted, the lowest remaining path coefficient is .39. While these values are somewhat lower than those resulting in the test of the first hypothesis, they confirm the belief that a causal relationship exists. The significance of the coefficients of determination used to assess the linkage between product usage situations and object benefits 1n: provides additional support for the existence of the hypothesized relationship. Therefore, hypothesis two was accepted. Hypothesis Three: The third hypothesis stated that there is a causal relationship between psychographics and object benefits. There is virtually no support for this hypothesis. The five path coefficients ranged from a high of .13 to a low of .08. While some of the resultant path coefficients were statistically significant, they were SO small that any documented relationship would be of no real value to the marketer. Therefore, hypothesis three was rejected. THE MODEL REVIS ITED The decisions regarding the three hypotheses imply that the initial specification of the model was in error. The linkage between psychographics and object benefits should be deleted. The final model reflecting the observed relationships is as depicted below. PRODUCT OBJECT PSYCHOGRAPHICS USAGE BENEFITS SITUATIONSI 106 This model was supported by the low path coefficients associated with the third linkage and the earlier analysis which supported the efficacy of the alternative model depicted above. LIMITATIONS In addition to the limitations which were discussed in Chapter Three, some problem areas surfaced during the implementation phase of this study. These problem areas are discussed below. Because of the problems encountered in attaining acceptable levels of reliability for the psychographic scales, the scope of the study was reduced from a potential of nineteen scales to five. Also problematic was the fact that the five retained scales did not exhibit the levels of reliability that one would strive to achieve. An alpha coefficient of .5 qualified a scale for continued analysis. A coefficient of at least .7 is desirable; unfortunately, only one scale exhibited this level of reliability. This fact reduces the generalizability of the model. There was no a priori testing of the reliability of the product usage situation statements. A pretest of the questionnaire should include the testing of the reliability of these items. Such an evaluation would lead to even better dimensionality of the underlying measurement constructs. The lack of equivalence of the covariance matrices reflects another limitation which inhibits the generalizability of this research. It is conceivable that a more comprehensive analysis of the reliability of the various model component scales would help overcome this problem. In the task of recreating the original input correlation matrix from the results of the five and fifteen group confirmatory factor analyses, the five group scenario proved as efficient as the fifteen group scenario. This is a major limitation as it implies that the model components were not unique; there was redundancy among the three, constructs. There was only one product—oriented object benefit statement associated with each psychographic dimension. While this is acceptable for the task of theory development, it does not reflect the conditions under which the consumers make product decisions. A more exhaustive set of objective benefits Should be used, especially if the task involves managerial applications. CONTRIBUTION TO THE DISCIPLINE This research has delineated a systematic series of relationships between psychographics and buyer behavior. It has addressed the fact that previous research on psychographics has often provided inconclusive results. 108 The results present a comprehensive theory of how psychographics influences buyer behavior. It not only addresses the belief that the influence is indirect, but identifies and defines two relevant sets of intervening variables. While product usage situations and object benefits are not new to the process of describing buyer behavior, this model reflects a newly modelled system of relationships. Theory development provides the basis by which earlier research and theories can be reevaluated. Previously developed models of buyer behavior routinely incorporate psychographics. Practitioners often speak of the problems inherent to such a model. This research provides the framework for the potential resolution of this conflict. This meshing of theory and practice reflects a significant contribution to the marketing discipline. Subsequent models of buyer behavior should be influenced by this research as well. Any comprehensive model would undoubtedly include many variables other than psychographics. The extent to which other variables, such as demographics, influence product usage situations and object benefits should be explored. If such a model proved capable of better explaining the variability in consumer decision making, then this project has resulted in a major contribution to all marketers. MANAGERIAL IMPLICAT IONS This research lends additional credibility for the use of psychographics in the development of marketing strategy. Particularly important is the idea that psychographics- based product usage situations can be an integral input in the task of strategy formulation. While psychographic variables reflect characteristics of the consumers and object benefits reflect product characteristics, product usage situations reflect a syntheSis of the two. People who possess certain psychographic characteristics will find themselves in related product usage situations; this will result in the desire for products which possess specific object benefits. The product which will ultimately be purchased will be the one that comes closest to matching the set of desired object benefits. Product usage situations may require a more extensive look into the development of a promotional strategy. Commercials such as the Miller Brewing Company's current "Miller Time" series should find increasing acceptance among marketing practitioners. Market segmentation strategies have long been predicated upon psychographics and product benefits. Product usage Situations also appear to represent a viable approach to segmenting the market. Practitioners could develOp their own battery of items which reflect their market segments and the needs of those segments. This process could lead to greater efficiency in predicting behavior in the marketplace and thus reduce the risks associated with managerial decision making. ADDITIONAL RESEARCH Two basic research needs are apparent; both would complement this study. A more comprehensive set of object benefits statements needs to be developed for the product under Scrutiny. This process would allow the testing of the model where the object benefit construct is more diverse. It would be important to determine whether the causal relationship between product usage situations and object benefits would diminish under these conditions. Also, there is a need to devote more energy to the evaluation of the complete model. The presumed linkage between object benefits and buyer behavior should be evaluated in conjunction with the linkages documented in this study. By subjecting the entire model to empirical testing, additional contributions to the discipline would be achieved. SUMMARY This research provides a contribution to marketing theory. It demonstrates that there is a relationship between psychographics and buyer behavior; however, it shows that the relationship is indirect. It could conceivably lead to better models of buyer behavior. With better models of buyer behavior come better strategic marketing decisions. Thus, not only does this research contribute to marketing theory, but it also has the potential to help reduce the risk associated with the development of an organization's marketing mix. The academician and the practitioner can both benefit from a understanding of the results of this study. APPENDIX 113 TABLE A-l RESULTS OF RELIABILITY ANALYSIS FOR THOSE SCALES NOT USED IN THE TESTING OF THE PROPOSED MODEL PSYCHOGRAPHIC ALPHA IF SCALE ITEM DELETED DECENTRALIZATION -My state government is doing more for me than the U. S. government. .17 -Every region should save its resources for its own people. .12 -If a community has a problem, it is up to its residents to solve it. .08 COEFFICIENT ALPHA FOR ENTIRE SET: .17 ECONOMIC RATIONALITY -People who pay list prices aren't getting their money's worth. .04 -I enjoy negotiating lower prices for things I buy. .37 -People should buy more when prices are lower. .11 -I always pay list price for things I buy. .02 COEFFICIENT ALPHA FOR ENTIRE SET: .13 EMOTIONAL —In making important decisions, the head should rule the heart. .18 —I always give my relatives a big hug when they visit. .10 -I make it a point to Show people how much I care about them. .03 -I often cry at sad movies. .06 COEFFICIENT ALPHA FOR ENTIRE SET: .24 HIGH TECH/HIGH TOUCH -I don't.like to leave messages on telephone answering machines when calling friends. .39 -I prefer to write my own checks rather than have the bank autimatically pay my bills. .21 -If it was possible, I would prefer to have meetings through telephone communications rather than face-to-face. .18 COEFFICIENT ALPHA FOR ENTIRE SET: .14 14 0 TABLE A-1(Continued) INFORMATION ORIENTED -It would be nice if I could determine what stories would be included and excluded from my copy of the daily newspaper. .32 -Knowing how to get the information I need to make a major decision is more important than having the information. .01 -While I may not use it, I want all relevant information available to me when I make a major decision. .20 COEFFICIENT ALPHA FOR ENTIRE SET: .05 MATERIALISM -The only people who say money can't buy happiness are those who never had a chance to try. .12 -If I'm lucky enough to buy the good things in life, I'm going to enjoy them. .19 -Material possessions are superficial when compared to the real joys of life. .25 COEFFICIENT ALPHA FOR ENTIRE SET: .10 MULTIPLE OPTION -There are always several good solutions to a problem. .27 —No one style of fashion is necessarily "correct." .20 -When facing a decision, I'm comfortable with a limited number of options. .04 COEFFICIENT ALPHA FOR ENTIRE SET: .24 NOVELTY/CHANGE SEEKING -New experiences enrich one's life. .11 -I would like to try something new every day. .12 -I have a daily routine. .39 COEFFICIENT ALPHA FOR ENTIRE SET: .27 PLANNING -Before I begin a project, I always make sure I have everything I need to finish it. .04 -To avoid unpleasant surprises, I usually plan exactly what I'm going to do each day. .03 -Too much time spent organizing results in little time to finish the job. .27 COEFFICIENT ALPHA FOR ENTIRE SET: .14 TABLE A-1(Continued) PERSONAL CONTROL -When I make plans, I am almost certain I can make them work. -What happens to me is my own doing. COEFFICIENT ALPHA FOR ENTIRE SET: .27 REALISM —Dressing in far-out clothes is a good way of self-expression. -People who act naturally rather than try to be different are going to be successful in life. -People should be themselves. COEFFICIENT ALPHA FOR ENTIRE SET: .13 SELF-HELP -People should solve their own problems rather than expect someone else to solve it for them. -If parents are competent, they should be allowed to teach their children at home rather than send them to school. —Exercising and eating right is a better way to good health than medical insurance. -I often work on do-it-yourself projects at home. COEFFICIENT ALPHA FOR ENTIRE SET: .12 SENSUOUSNESS -Sme11ing, touching, and feeling are more important than seeing. -It's important to hold a product before I buy it. COEFFICIENT ALPHA FOR ENTIRE SET: .07 TRADITIONALISM -I rarely celebrate holidays and birthdays. -I like having traditions which Specify how certain things should be done. -People should try to do things their parents did. COEFFICIENT ALPHA FOR ENTIRE SET: .16 TR NA NA .25 .04 .04 .05 .21 .04 .08 NA NA .30 .11 .12 SIMILARITY COEFFICIENT MATRIX TABLE A-2 ON ON ON cn O« cN cN Nc «o «N n nN Nn Nm am Nm mm mc an ac «n oc On on an nc NN ON ~N «N OO« h« O« nN « NN NN nn mN m« m« On ON O« Nn «n nn rN NO ON n« «N ON n ON nn an On On cm ~« OO« «m on 5N on NI N NN o« « NI O« ON On ON On an mN ON nr O« m mN cm NN :N c« mN mm m« «m OO« an n NN NN ON ON h NN cN O« 5c ~N nn cN «c O« O« nu NN ON On On cc mn w« mN ON nN 0n on OO« no «m ON mN nn hm «NI n« Nc Oc an cm mm Nc Nc «n «« ON nc mc «c on h« h« ON on « ON h nn OO« «5 ON O« N« N O« ON ON «N «n O« s« O mN an on cc m« On nn O an mm em h NN On NN «m «h OO« «n mN ON m« «« Nn Nn ON cn 5N hN :« Nn On «c sc oc mc kc on on Or cm a NN NI NN ON ON «n OO« On mn ml oc Om cc mm mm On ON cN nN ON Nn Om cc hc «c mc ¢N mN hN «« nn N ON MN O« mN Om OO« @m m cm we «c On mm mm mc on pm On ON mm ac mc Oc. «m NN ON nN m mN NN ON nn N« ON mn Om OO« oc N «n oc am no NO mo we Oc Oc nN mm cc mm mm mm ON ON Nn ON O« c« h On N m« ml n oc OO« Nmr ~«I on ON Nc an «m an cc ON n« «« h« nN nN h« m n «« ON m« « NN «NI O« «« Oc cn N le OO« co ¢« N« w ¢«I n«r m«l cu m NN sN m n« n c« On Nn mm m« on NI cN n« 0N Nn O0 co «m >«I 00 OO« Om Nc cn «N nN c« cN cN nc Om mc «c Nn oc pm on ma O hN O« O« Nc ON Nn cc «c Oc On m« Om OO« hm Nm nm No Nc cm on oc hc hm mc «n mm mm On on Nn o« ON hc cc «N ON On On an ON N« Nc hm OO« O~ me no Ow Oc nn h« no cc hm om Ow On ON mm 0« Nn an hN mm «m cm mm mm hO Nc 0 cm Nm Oh OO« ms co mo no hm «c on mm mm mm cc On On «c mN «n ON nn cm O« 5N On mm No on o«l «N nm we ms OO« ss O» OO cm ON Cc ON Om mm mc pN NN ON O« an On cN mm ~« ON ON mc OO «0 n«r nN No no OO up OO« Oh we we mN Om Nc am On hc @« :« nN ON ON mm «c Nc O c« cN on no an m«0 c« Nc Oo me Op 35 OO« ow mc «N mc ON Om om mc cN ON mN ON Nn n O« Nc mN Nn nN On Oc cc cl cN mm Oc hm O0 O0 ow OO« Om nn Nc «n cc «c ON c« cN mN ~« ON ON O« «n on On ON on Oc 0N m cN on nn hm cm Oc mc on OO« co Oc 3N mm cm mN On cn an N n« m: n1 «« mm «c Nn ON nN n« NN nc Oc h« «c ON mN «N on cw OO« Om ON Nm «n h ON hc 5c n «N O« NN ON cc hc Om mm mm «« ON Om 5c no om mc Om mc Nc Oc om OO« Om O0 ch N0 Nc oc «m n« ON m ON nc m« Oc oc cc cc h« o mc hm Oc mn ON Nc ON «n ON ON Om OO« Os am we nN Nn On « nN MN On mc On mc sc mc Nm nN n« «c mc hm nm On O0 O0 cc mm Nm Oo Oh OO« «a On cm nc Oc c ON cm mm «c nn hc «c Oc mw nN n Nn «n On mm mm mm mm «c cm «n c» OO «O OO« Nh On cn «c 0 mm NN cc on O On mc «m On On O« @c n so cc mc kc Cc ON mN N No we Om Nb OO« Nc ON mm O« on oN mm ~« On on 0N NN ON n On On Nn On on 5N o« cN N« on ON Nc mN cn On Nc OO« Oh «m N« on c« m« 5« mm Om mN ON ON n Nn On On ON On NN a« ON cN cn sc wc Nn mc cn ON sh OO« cO m« on MN nN ON om on ON nN Nn «« mN an mm mm «c ON nN mN nN On pc «m On Oc «c mm «o cO OO« c« cN nN ON on O a «« o mN ON 0« O NN O« mN O« ON ON h« N n n« « c m O« N« m« c« x«¢» .Hamomc cmq 10> mm emu mm cows: 030 on» .nmc3pow0 L30> Form 30> can: coHpm3pHm on» 03000 xcwna 03.non amass» mebo 000:» 00:» 30> 0>Hw mcfixmp 0L300H0 0000 can: acmmsmcu own» on soc kn: wmcflcp 05mm 020 cmfiaasooom 00 00 0H300 30> mmao mcficpmeom mcocu 0H wmmc300H0 mxmu 30> >23 m>mzam was» 0H wmc30030 was» memo 20> can >c3 m>zxg mHHmmzmm ¥mmzam 30> on wwemgs 0m 30> 00 00:3 wmcHzpmeom pcmmmcqmc 0H moon wmflcu mm: mums: seems» >mcs \30> 0L0: >23 mos: «memsso mmoc300H0 wcwxm» 30:3 0coH< ”mama: .zmmz .0=:H wm>mza< womam 00 mcmnpo xcwss 30> om mums» mxfla Hmmu m>mza< wpcmomc magnumsom was» ma L0 .>m: 0H2» Homo m>mza< mmmc3uowa on» mcfixmu 0L0: 30> 00:3 mmcfiammm H>ng mmHHH>HHu< A “.0“ mcfipoosm mHHL: 0005 on» 00>0ncm .moc300H0 L30> soc» 030 onm "0000000000; oev mmeHon 02Hu