Shortening the Rokeach Value Survey For Use in Consumer Research

J. Michael Munson, Santa Clara University
Edward F. McQuarrie, Santa Clara University
ABSTRACT - Although the Rokeach Value Survey (RVS) has been widely used in marketing studies, its high degree of generality comes at some cost in consumer research applications. Intuitively, not all of its 36 value items would seem equally relevant to consumption. This research describes efforts to shorten the RVS to reflect a set of more consumption relevant values. Results from five samples identify a subset of 24 value items as maximally relevant to product consumption. A measure of value instrumentality exhibiting satisfactory psychometric properties is constructed based upon this value subset. Implications are discussed for domestic and Gross cultural marketing.
[ to cite ]:
J. Michael Munson and Edward F. McQuarrie (1988) ,"Shortening the Rokeach Value Survey For Use in Consumer Research", in NA - Advances in Consumer Research Volume 15, eds. Micheal J. Houston, Provo, UT : Association for Consumer Research, Pages: 381-386.

Advances in Consumer Research Volume 15, 1988      Pages 381-386

SHORTENING THE ROKEACH VALUE SURVEY FOR USE IN CONSUMER RESEARCH

J. Michael Munson, Santa Clara University

Edward F. McQuarrie, Santa Clara University

ABSTRACT -

Although the Rokeach Value Survey (RVS) has been widely used in marketing studies, its high degree of generality comes at some cost in consumer research applications. Intuitively, not all of its 36 value items would seem equally relevant to consumption. This research describes efforts to shorten the RVS to reflect a set of more consumption relevant values. Results from five samples identify a subset of 24 value items as maximally relevant to product consumption. A measure of value instrumentality exhibiting satisfactory psychometric properties is constructed based upon this value subset. Implications are discussed for domestic and Gross cultural marketing.

INTRODUCTION

Philosophers, social scientists, and industrial leaders have long recognized the potential importance of values as influencers of human behavior. The disciplines of psychology, sociology, and cultural anthropology have suggested that values may underlie a variety of individual and collective behaviors. So too have marketers long realized the potential inherent in values and their central role in motivating and explaining consumption. However, they have been rather slow to embrace empirical research assessing specific relationships between values and consumption behaviors. Most such research has evolved only over the past decade. One of the earliest and perhaps most noteworthy studies in this vein was that of Vinson, Scott and Lamont (1977), who hypothesized a distinction between more global, underlying Rokeach-type values, and more superficial, domain-specific values. Their study served as a spring board for the rejuvenation of interest in values among many marketing researchers. That such interest continues to build is evidenced by several signs: the special conference on values and consumer behavior at the University of Mississippi in Summer 1984, and the subsequent publication of Personal Values and Consumer Psychology (Pitts and Woodside 1984); a special issue of Psychology and Marketing (Kahle 1985) devoted exclusively to values and consumer behavior; and the publication of values articles in other marketing journals (e.g., Kahle, Beatty and Homer 1986, Henry 1976; Gutman 1982; Munson and McIntyre 1979).

VALUES IN CONSUMER RESEARCH: TWO TRADITIONS

A review of the marketing literature suggests that values have been used in consumer research in two primary ways: in Value Hierarchies (VH) and/or Value Instrumentality (VI) assessments. The first tradition, VH, has used values to profile consumers using some preexisting inventory or list of general human values. Several such inventories have been used (e.g., Kahle 1983; Kluckhohn and Strodtbeck 1961; Rokeach 1973). The methodological significance of such a values hierarchy is that it enables researchers to describe, in a quantitative fashion, the values of virtually any group and to compare and contrast these values with those of any other group.

Marketing researchers have used value hierarchies as a means to better understand differences between consumer groups. The most widely used value inventory in consumer research has been the Rokeach (1973) Value Survey (RVS) (see Munson 1984 for a review). Studies in this vein are numerous and would include, for example, those profiling value differences across cultural groups and the stereotypic consumer within cultures (Munson and McIntyre 1978); subcultural (Powell and Valencia 1984), and/or social class groupings (Ness and Stith 1984,); studies using values to distinguish market segments (Vinson and Munson 1976); and studies on the relation of antecedent life cycle variables to values (Crosby, Gill and Lee 1984). Other research has focused on alternative ways to measure the value hierarchy, and compared rank ordering, scaling, and paired comparisons of individual Rokeach value items (Munson and McIntrye 1979; Reynolds and Jolly 1980).

Within the Rokeach Value Survey two sets of values are distinguished, one composed of 18 Terminal values, or desired end states of existence (e.g., an exciting life, national security); the other composed of 18 Instrumental values, or preferable modes of behavior (e.g., being ambitious, independent). A value is defined as an enduring prescriptive or proscriptive belief that a specific end state of existence or specific mode of conduct is preferred to an opposite end state or mode of conduct. The values within each respective hierarchy are rank ordered according to their importance as guiding principles in the individual's life. The output of this ranking procedure is a values hierarchy.

The second tradition, the Values Instrumentality approach, has focused on the means-end chain linking values and behavior. Values are viewed as desired goals or ends of consumption, and products and/or product attributes are perceived to be the means to realizing those values. The likelihood of consumption is felt to increase when the values associated with the product match those which are central to the consumer. The VI approach is evidenced, for example, in the conceptual arguments of Howard (1977) and Howard and Woodside (1984), as well as in the empirical works of several others (e.g., Gutman 1982; Reynolds and Gutman 1984; Prakash and Munson 1985; Vinson, Scott and Lamont 1977).

There are some important distinctions between the VH and VI traditions. Unlike the VH approach, VI research is not necessarily concerned with indexing the consumer's value profile ( i.e., values hierarchy) on an inventory of underlying values general to all human decision making (such as the RVS). Rather, the focus of VI is on identifying those "values" which can be linked to product attributes. Such values are not necessarily global human values, but can be less general, less abstract, and perhaps more consumption oriented outcomes. A product is a means to achieve some end, and features of the product (attributes) cause consumers to relate certain products to certain values.

The desirability of a measure of values instrumentality becomes all the more apparent when one realizes that this tradition is quite consistent with two other research orientations: Ostrom and Brock's (1968) involvement model, and Rosenberg's (1956) model of attitude structure. Both of these formulations suggest that some outcome, behavior or state is based upon an expectancy x value relationship. Whereas Ostrom and Brock contend that involvement is the product of value instrumentality weighted by value centrality, Rosenberg contends that one's attitude toward an object can be defined as the product of the object's value importance weighted by its perceived instrumentality.

Each of these two values research traditions is not without its problems. Researchers using the VH approach have noted that many of the individual value items in the RVS seem a priori to be largely irrelevant to consumption behavior (Beatty et. al. 1985; Vinson and Munson 1976; Vinson, Scott and Lamont 1977; Prakash and Munson 1985; McQuarrie and Langmeyer 1985). Indeed, findings from several studies confirm that many of the individual value items are unrelated to the marketing behavior of interest. Moreover, when sample sizes are expected to be small, valuable degrees of freedom may be lost when using such techniques as regression or discriminant analyses to determine possible differences in the value hierarchies of various groups. Furthermore, irrespective of sample size considerations, values perceived to be unrelated to consumption may increase the likelihood of respondent fatigue and error, and obscure these true relationships which are present between values and consumption (Beatty et. al 1985).

Despite these problems with the original 36 item RVS, we do not believe they would constitute sufficient incentive to shorten it, if the primary research goal were to measure traditional value hierarchies for specific populations. In such instances, the full inventory will almost always be more appropriate. Rather, the necessity and motivation to shorten the full RVS will be greatest when the research goal is to determine means-end or value instrumentality relationships, particularly where the number of product categories and attributes to be evaluated by each respondent is large.

In view of the desirability of constructing such VI measures we set out to shorten the list of Rokeach values. We chose to begin with this value inventory rather than some other, because of its widespread use in consumer research and in view of its claim to be an exhaustive set of human values. Our stance is that although the Rokeach values may serve as guides or standards for human life in general, they will not all be perceived as ends of consumption. Hence, the primary objective of this research is to identify the best set of consumption relevant values in the RVS and exclude those which are extraneous to product consumption. A shortened set of consumption-relevant, Rokeach value items could be used to facilitate several types of research applications: (1) values instrumentality studies; (2) constructing Rosenberg-type attitude measures in which the perceived instrumentality component was indexed to global values of the RVS type, as in linking attributes to values for a dynamically discontinuous innovation (McQuarrie and Langmeyer 1985); and (3) in product positioning and advertising planning concerned with identifying motivational appeals relevant to values (e.g. Pollay 1984; Tyebjee 1978).

The identification of a subset of consumption-relevant values within the broader set of human values also has intriguing implications for cross cultural research. It stands to reason that different values will be viewed as achievable through consumption activities in different societies. Although our planned value reduction procedure is a strategy for studying value instrumentality in the United States, it is of heuristic benefit when it comes to studying cross cultural consumption. For instance, it may be possible to show that two societies differ in terms of which values they believe can be served or reached through purchase and consumption of goods and services. Hence this research may offer us an additional insight into how cultures differ and into particular marketing problems which may be faced in certain cultures.

SAMPLE AND PROCEDURE

Data were collected from three samples in an attempt to reduce the size of the 36 item RVS (Rokeach 1973). In Sample 1, eighty undergraduates were asked to identify the 12 Rokeach values "most irrelevant" to the consumption of goods and services. Two dozen specific products were listed as examples. At a later point, these same undergraduates- rated nine products on their instrumentality with respect to each value in the RVS ("Sample 2"). In sample 3, a new group of 64 undergraduates was asked to indicate which Rokeach values were "unobtainable" or could not be reached through consumption of common products and services (eight product categories with examples; e.g. electronics (television, vcr...)). A subset of Rokeach values was used in this task, eliminating both those values shown by Samples 1 and 2 to be certainly irrelevant to consumption (e.g. "forgiving") or certainly relevant (e.g. "an exciting life"). It was hoped that the shortened list would aid subjects in making more discriminating judgements about the relevance of the remaining values to consumption.

Subsequently data were collected from two large samples to determine the psychometric properties of a value instrumentality inventory that utilized a reduced version of the RVS. Sample 4 contained 714 consumers and Sample 5 contained 333 consumers. The combined sample was 42% male and 58% female, and 36% students, 64% adults. Degree of instrumentality was measured in Sample 4 by a three point scale (not related, weakly related, strongly related), and in Sample 5 by a five point scale (no, weak, some, definite, or strong relation).

RESULTS

Identification of Consumption Relevant Values

Table 1 displays the results of the three attempts to reduce the size of the Rokeach Value Survey. For Sample 1, the quantity shown represents the number of students who indicated that the value was among the 12 least relevant to consumption. In Sample 2, the quantity shown represents the number of times that a product was judged to be not instrumental to that value (based on 80 students X 9 products). In sample 3, the quantity shown represents the number of times a value was judged to be unobtainable through consumption of any of a variety of products and services. The three samples, and the three different tasks, provide convergent evidence as to which values are relevant to ordinary consumption activities and which are not.

TABLE 1

ROKEACH VALUES JUDGED TO BE IRRELEVANT, UNOBTAINABLE BY OR NOT INSTRUMENTAL TO THE CONSUMPTION OF GOODS AND SERVICES

Table 2 provides a summary of these findings, in terms of the values most often judged not relevant to consumption activities across the three samples. The values viewed as irrelevant to consumption fall into three broad groups: 1) those that involve religion or spirituality (e.g., "salvation"); 2) those that concern collectivities of people more than individuals (e.g., "national security"; "world at peace") and 3) those that concern traits viewed as evidence of a virtuous character (e.g., "honest"). Taken as an ensemble, these values provide a portrait of the kinds of goals that these students viewed as unreachable through materialism. For these respondents, consumption does not lead to salvation, does not make one a better person, and cannot save the world. Summarized thus, the reduction procedure appears to have led to plausible results.

TABLE 2

SUMMARY OF ROKEACH VALUES JUDGED NOT RELATED

The remaining task was to fix the boundaries of the subset of consumption relevant values. This was necessary so that a value instrumentality inventory could be constructed. We did not feel that the results of the reduction procedure were sufficiently clear as to be applied mechanically. Rather, we exercised our judgement concerning values that appeared to lie on the boundary of relevance. For aesthetic reasons, we decided to choose exactly 24 Rokeach values for inclusion in the value instrumentality scale. The elimination of 12 values appeared to us to strike a balance between the goals of reducing the RVS to a manageable size, and retaining as comprehensive a picture of human values as possible.

Table 3 displays the 24 values we deemed most relevant to consumption, and the 12 that were set aside. Note that although "equality" and "wisdom" were generally viewed as irrelevant, in sample 1, 2, and 3, an earlier scale by McQuarrie and Langmeyer (1985) had included both. Hence, to fill out the 24 value set these two values were retained, as were the Instrumental values "broadminded," "logical," and"self-controlled," as possibly relevant to consumption.

Psychometric Properties of the Value Instrumentality Inventory.

In Samples 4 and 5, respondents rated the degree to which specific products were perceived as instrumental to obtaining the 24 consumption relevant values. These ratings were summed to form a Value Instrumentality Inventory (VII). Such an inventory has the potential to measure degree of involvement with products (Ostrom and Brock 1968). The 24 item sum shows a high degree of internal consistency in both samples, with coefficient alpha equal to .94 and .95, respectively. Of course, this consistency results in part from the fact that most products are viewed as not instrumental to virtually all the 24 values.

TABLE 3

VALUES JUDGED RELEVANT AND IRRELEVANT TO CONSUMPTION

In Sample 5 principal components factor analysis with varimax rotation revealed three factors with eigenvalues greater than 1.00. As shown in Table 4, this factor structure is readily interpretable. The first factor identifies products that help to fulfill adult responsibilities (e.g., "responsible," "self-controlled"); the second those that fulfill lifestyle goals (e.g., "exciting life," "comfortable life"); and the third those that are perceived to remove tension (e.g., "inner harmony," "cheerful"). It appears that the Value Instrumentality Inventory can help describe the motivations that govern purchase of specific products.

In the third test, the ranking of the products in terms of the degree to which they were viewed as instrumental to valued goals was compared across the two samples. As Table 5 shows, the ranking of the products conforms to intuitive notions about how involving these products would be. The size of the correlation between the scores accorded to the 12 products in the two samples adds to our confidence that the VII yields stable scores (r =.98, p<.001).

DISCUSSION AND IMPLICATIONS

The primary objective of this study was to shorten the Rokeach Value Survey to reflect a set of more consumption relevant values. Such an inventory would be particularly germane to researchers attempting to measure value instrumentality, as opposed to value hierarchies. Toward this end, three alternative reduction procedures were used. The result of these convergent reduction procedures is a subset of 24 value items identified as maximally relevant to consumption. As anticipated, many values in the original RVS, both Terminal and Instrumental, were found to be largely irrelevant to, or unobtainable via consumption behavior (e.g., world at peace, mature love, salvation, courageous, forgiving, honest, etc.). Using this 24 item subset, we were able to construct a Values Instrumentality Inventory (Table 4) which demonstrated satisfactory psychometric properties with respect to its internal consistency, stability over two independent samples, and factor structure. Furthermore, the rank ordering of the 12 products investigated based on their mean value instrumentality scores derived from the VII was consistent with expectations. For example, as shown in Table 5, automobiles show the highest value instrumentality, while for toothpaste it is moderate, and potato chips it is low.

TABLE 4

FACTOR ANALYSIS OF THE 24 ITEM VALUE INSTRUMENTALITY INVENTORY

TABLE 5

TWELVE PRODUCTS RANKED ACCORDING TO DEGREE OF VALUE INSTRUMENTALITY ON THE VII

The availability of this shortened, more consumption oriented VII suggests important practical implications. It should be of interest not only to those studying means-end value linkages using the more traditional approach (e.g., Gutman 1982), but also to researchers working in other areas. Value instrumentality is a central concept in Rosenberg's (1956) attitude model and Ostrom and Brock's (1968) model of involvement. Moreover, values instrumentality data can be highly useful in formulating product positioning and advertising strategies (e.g., Pollay 1984), with the objective being to link product attributes to consumption benefits or consequences, and ultimately to the consumer's underlying values.

Although we performed this value reduction procedure for the pragmatic goal of developing a more efficient and direct way of measuring value instrumentality, we suggest that the reduction procedure, when treated as an end in itself, could be a useful approach to studying cross cultural differences and the role of purchases and consumption in human life. For example, the Balinese in Indonesia, or the Chinese in Singapore, probably do not look to consumption of goods for the same values as consumers in the United States. In fact, subcultures within the United States may also exhibit major differences in product-value instrumentality relationships. Would Hispanics in San Antonio view the same values as relevant to consumption as the Hispanics in Miami, or the Pennsylvania Dutch? Would consumers in the Midwest or Southeast hold the same values to be relevant to consumption as those found for the West Coast consumers in this study? Would all three of the latter groups perceive the same product as instrumental to the same value or value set?

Questions such as the above point up a major - limitation of this study. Its results may not be generalizable to populations and product classes beyond those investigated here. It is axiomatic that values are culturally derived. Therefore, both value hierarchies and subsequent value instrumentality inventories will in all probability exhibit heterogeneity across different cultures and subcultures. This reality raises important cautions for values researchers and international marketers. It advises against the wholesale transportation of the original RVS, or any shortened derivative assessing values instrumentality developed in the United States, to another country or nation state. Although Rokeach (1973) suggests that the RVS is applicable to other "Western" countries, this proposition remains relatively untested. Hence, the approaches used here for values reduction, as well as the shortened Values Instrumentality Inventory, need to be tested for their generalizability on a country by country basis.

The 24 item measure presented here (Table 4) is not intended to constitute "the" values instrumentality inventory for all product classes or areas of consumption. We can well imagine that a value generally regarded as irrelevant to consumption and omitted from our inventory could assume prepotency for some products in some situations. For example, "salvation" may be assumed to be the most consumption relevant value in a book publisher's efforts to market the Bible. Such examples notwithstanding, this research and the values reported in Table 4 probably represent a "reasonable first cut" at reducing the original 36 item RVS to a more consumption relevant list.

Future research might extend the present study by investigating: (1) whether the value reduction procedures used here are applicable across other populations, both cross culturally and subculturally; (2) whether the values identified here as maximally relevant (or irrelevant) to consumption are generalizable to other cultures and subcultures; and (3) whether terminal and instrumental values exhibit differential patterns, across cultures or subcultures, in their respective relationships to value instrumentality judgements.

In conclusion, the values reduction methods investigated here can be used to purge the RVS of values which are largely irrelevant or tangential to most consumption behavior. The shorter, more consumption oriented VII may be particularly useful to researchers concerned with measuring value instrumentality relationships.

REFERENCES

Beatty, Sharon E., Lynn R. Kahle, Pamela Homer and Shekhar Misra (1985), "Alternative Measurement Approaches to Consumer Values: The List of Values and the Rokeach Value Survey," Psychology and Marketing, 2 (Fall), 181-200.

Crosby, Lawrence A., James D. Gill and Robert E. Lee (1984), "Life Status and Age as Predictors of Value Orientation," in Personal Values and Consumer Psychology, eds. R.E. Pitts and A.G. Woodside,Lexington, MA: Lexington Books, 201-218.

Gutman, J. (1982), "A Means-End Model Based on Consumer Categorization Processes," Journal of Marketing, 46 (Spring), 65-72.

Henry, W.A. (1976), "Cultural Values do Correlate with Consumer Discontent," Journal of Marketing Research, 13, 121-132.

Howard, J.A. (1977), Consumer Behavior: Application of Theory. New York: McGraw Hill.

Howard, J.A. and Arch G. Woodside (1984), "Personal Values Affecting Consumer Psychology," in Personal Values and Consumer Psychology, eds. R.E. Pitts and A.G. Woodside, Lexington, MA: Lexington Books, 3-12.

Kahle, Lynn R. (Ed.) (1985), Psychology and Marketing, 2, (Winter), 229-310.

Kahle, Lynn R. (1983), Social Values an Social Change: Adaptation of Life in America. New York: Praeger

Kahle, Lynn R., Sharon E. Beatty and Pamela Homer (1986), "Alternative Measurement Approaches to Consumer Values: The List of Values (LOV) and Values and Life Styles (VALS)," Journal of Consumer Research, (December), 405-409.

Kluckhohn, F.R., and R.L. Strodtbeck (1961), Variations in Value Orientations. Evanston, Ill.: Row, Peterson and Co.

McQuarrie, Edward F. and Daniel Langmeyer (1985), "Using Values to Measure Attitudes Toward Discontinuous Innovations," Psychology & Marketing, 2 (Winter), 239-252.

Munson, J. Michael (1984), "Personal Values: Considerations on Their Measurement and Application to Five Areas of Research Inquiry," in Personal Values and Consumer Psychology, eds. R.E. Pitts and A.G. Woodside, 13-33.

Munson, J. Michael and S.H. McIntyre (1978), "Personal Values: A Cross-Cultural Assessment of Self Values Attributed to a Distant Cultural Stereotype," in Contributions to Consumer Research, ed. H. Keith -Hunt, Chicago: Association for Consumer Research, (Proceedings), 165-166.

Munson, J. Michael and S.H. McIntyre (1979), "Developing Practical Procedures for the Measurement of Personal Values in Cross-Cultural Marketing," Journal of Marketing Research, 16, 55 - 60.

Ness, Thomas E. And Melvin T. Stith (1984), "Middle Class Values in Blacks and Whites," in Personal Values and Consumer Psychology, eds. R.E. Pitts and A.G. Woodside, Lexington, MA Lexington Books, 231-237.

Ostrom, Thomas M. and Timothy C. Brock (1968), "A -Cognitive Model of Attitudinal Involvement," in Theories of Cognitive Consistency, eds. R. Abelson et. al., New York: Rand McNally.

Pollay, Richard W. (1984), 'The Identification and Distribution of Values Manifest in Print Advertising, 1900 -- 1980," in Personal Values and Consumer Psychology, eds. R.E. Pitts and A.G. Woodside, Lexington, MA: Lexington Books, 111-135.

Powell, Terry E. and Humberto Valencia (1984), "An Examination of Hispanic Subcultural and Regional Value Orientations," in Personal Values and Consumer Psychology, eds. R.E. Pitts and A.G. Woodside, Lexington, MA: Lexington Books, 219-230.

Prakash, Ved and J. Michael Munson (1985), "Values, Expectations From the Marketing System and Product Expectations," Psychology & Marketing, 2 (Winter), 279-296.

Reynolds, T.J. and J.P. Jolly (1980), "Measuring Personal Values: An Evaluation of Alternative Methods," Journal of Marketing Research, 17, 531-536.

Reynolds, T.J.and Jonathan Gutman, (1984), "Laddering: Extending the Repertory Grid Methodology to Construct Attribute - Consequence - Value Hierarchies," in Personal Values and Consumer Psychology, eds. R.E. Pitts and A.G. Woodside, Lexington, MA: Lexington Books, 155-167.

Rokeach, M. (1973), The Nature of Human Values. New York: Free Press.

Rosenberg, M.J. (1956), "Cognitive Structure and Attitudinal Affect," Journal of Abnormal Social Psychology, 53, 367-372.

Tyebjee, Tyzoon T. (1978), "Refinement of the Involvement Concept: An Advertising Point of View," in Attitude Research Plays for High Stakes, eds. J. Maloney and B. Silverman Chicago, Ill: American Marketing Assoc., 94-111.

Vinson, D.E. and J.M. Munson (1976), "Personal Values: An Approach to Market Segmentation in Marketing: 1776-1976 and Beyond," ed. K.L. Bernhardt, American Marketing Association, (Proceedings), Chicago, 313-318.

Vinson, D.E., J.D. Scott, and L.M. Lamont (1977), "The Role of Personal Values in Marketing and Consumer Behavior," Journal of Marketing, 41, 44-50.

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