Personal Values: a Cross Cultural Assessment of Self Values and Values Attributed to a Distant Cultural Stereotype

J. Michael Munson, University of Santa Clara
Shelby H. McIntyre, University of Santa Clara
ABSTRACT - This research investigates the potential of a particular value assessing instrument, the Rokeach Value Survey (RVS), to distinguish cultural differences in personal value structures. Results suggest personal values can be productively employed by international marketers to discriminate peoples of culturally diverse backgrounds. A discriminant analysis disclosed significant differences in the self values held by people from three cultures: Thailand, Mexico and the United States. Additionally, the stereotype held of an average American's personal values differed significantly across the three cultures. The discriminant functions derived from responses to the RVS were able to correctly classify 65% of the people according to their cultural background as compared to a chance level of 22%.
[ to cite ]:
J. Michael Munson and Shelby H. McIntyre (1978) ,"Personal Values: a Cross Cultural Assessment of Self Values and Values Attributed to a Distant Cultural Stereotype", in NA - Advances in Consumer Research Volume 05, eds. Kent Hunt, Ann Abor, MI : Association for Consumer Research, Pages: 160-166.

Advances in Consumer Research Volume 5, 1978      Pages 160-166

PERSONAL VALUES: A CROSS CULTURAL ASSESSMENT OF SELF VALUES AND VALUES ATTRIBUTED TO A DISTANT CULTURAL STEREOTYPE

J. Michael Munson, University of Santa Clara

Shelby H. McIntyre, University of Santa Clara

ABSTRACT -

This research investigates the potential of a particular value assessing instrument, the Rokeach Value Survey (RVS), to distinguish cultural differences in personal value structures. Results suggest personal values can be productively employed by international marketers to discriminate peoples of culturally diverse backgrounds. A discriminant analysis disclosed significant differences in the self values held by people from three cultures: Thailand, Mexico and the United States. Additionally, the stereotype held of an average American's personal values differed significantly across the three cultures. The discriminant functions derived from responses to the RVS were able to correctly classify 65% of the people according to their cultural background as compared to a chance level of 22%.

INTRODUCTION

The determination of cultural differences is an important element in formulating international marketing strategy. in view of the increasing participation of American corporations in the international market place, and the threat of growing competition from foreign multinational corporations, knowledge and identification of culturally related differences in buyer behavior is a critical concern for marketing research. A fundamental area where cultural differences should be identified is in the culturally generalizable aspects of personality, attitudes, and personal values. Until recently, there has been little empirical research devoted to the analysis of values and their relation to consumer behavior. Research which has been done has focused solely on the values held by people in the United States (Scott and Lamont, 1974; Vinson and Munson, 1976; Henry, 1976; Vinson, Scott and Lamont, 1977).

While it is a widely accepted notion that cross cultural differences in values do exist, there is little empirical evidence actually documenting such differences or suggesting the particular form or pattern they take. In order for international marketers to assess possible cultural differences in values, a methodology and instrumentation is required which permits systematic comparisons of sufficient sensitivity to disclose those differences which might exist. Although there are a variety of segmentation techniques available which the marketer might consider for initially investigating cultural differences, such as benefit segmentation (Haley, 1968), activities, interests and opinions (Wells and Tigert, 1971), life style (Plummer, 1971), motivation research (Pernica, 1974), or cultural assimilators (Fiedler, Mitchell, Triandis, 1971; Worchel and Mitchell, 1972), each has serious limitations. Although all these techniques, except cultural assimilators, have been frequently applied to domestic marketing problems, their application to cross-cultural research is limited. Their conspicuous absence in the sphere of international marketing research stems in large measure from either the vast amounts of data each requires or their difficulty to implement in terms of effort, time and money. For example, to develop a meaningful AIO battery, might require the researcher to generate as many as 300 initial items (Wells and Tigert, 1971). Although cultural assimilators, which are self-administered programmed culture training manuals, provide an apparently effective method for assisting members of one culture to interact and adjust successfully to members of another, they too require considerable start-up effort. The training manual is composed of a number of "critical incidents" which are elicited from either Americans or foreign nationals. Each incident describes some specific intercultural occurrence or event that made a major difference in each respondent's attitude or behavior toward members of the other culture. The time required to prepare a battery of 75-100 such incidents is approximately 800 man hours (Fiedler, Mitchell, Triandis, 1971).

On the other hand, an approach to segment international markets based on personal value analysis might present fewer limitations. If cultural or subcultural groups were initially found to be dissimilar in their basic value orientations, the marketer might then consider the follow-up use of more complex and costly techniques such as AIO or benefit segmentation. Based on the above, the objectives of this study are primarily four-fold:

a) to provide empirical evidence that cross-cultural differences in personal value structures do exist.

b) to investigate the utility of a particularly value assessing instrument, the RVS for disclosing differences in cultural values.

c) to investigate the extent to which stereotypes of Americans differ across cultures.

d) to determine the extent to which a person's own value system differs from his stereotype of a person from a different culture.

PERSONAL VALUES

Domestic Research

The notion that values are of central concern in understanding consumer behavior has been pointed out by a number of researchers (Rosenberg, 1956; Lessig, 1975; Vinson and Munson, 1976; Vinson, Scott and Lamont, 1977). Rosenberg suggests attitudes are derived from values: "the attitude object tends to facilitate the attainment of a number of important values." Similarly Lessig (1975) asserts that attitudes stem in part from the value subsystem and are the expression of values toward given situations and objects. Vinson and Munson (1976) emphasize the centrality of personal values: "they represent the most basic element of the consumer's cognitive world, they structure the individual's perception and understanding of himself, significant others and of the objects and behavior which constitute his psychological environment" (p. 316).

Recent research has shown that not only are values differently held by specific groups of Americans, but more importantly to marketers are findings which suggest values are linked through attitudes to product attributes (Vinson and Munson, 1976; Vinson, Scott and Lamont, 1977; Munson, 1977). Vinson et al (1977) suggest a linkage may exist between general or global values and both more domain specific values relating to specific product qualities as well as subsequent product attribute evaluations. Vinson and Munson (1976) found parents and their university age children exhibited significantly different preferences for automobile attributes which seemed to be related to their respective underlying value structures. For example, university students attributed significantly greater importance than their parents to the values of "an exciting life" and "pleasure'' while their parents placed significantly greater importance on "family security" and being "responsible" (values reflecting standards of social and individual responsibility). Parents emphasized those attributes which signified utilitarian or functional characteristics associated with automobile ownership (e.g., quality of warranty, service required, handling) while students were more concerned with aesthetic and socially observable features (e.g., styling, prestige, luxury interior).

If values serve to segment buyers and are also tied to product attribute evaluations among a rather homogeneous population (e.g., people from the same general culture) one might expect these values to offer even more potential for segmenting peoples of diverse cultural background. This statement however presupposes that the values really do vary significantly across cultures.

Cross Cultural Research

It is a cardinal rule of marketing that it is always easier to appeal to existing cultural wants or expectations than to try to change the culture or to create new needs (Kassarjian and Robertson, 1973, p. 454). If marketers hope to formulate more effective product and promotional strategies for international markets, they must become sensitive to the core values of any given country, and their role in consumption differences. Available research evidence would seem to suggest that success or failure for the international marketer might be related in at least two primary ways to personal values. The first way concerns the importance or centrality ascribed to particular values by another culture, while the second way concerns inappropriate cultural stereotyping of the marketer's culture by the target culture. A few examples will illustrate the types of situations in which a marketer could make the first kind of error. The highly fragmented retail structure in France, with separate shops for fish, meat, pastries, etc., seems to be influenced by traditional French values. The desire to retain family control over the store and maintain status within the community may be more important than showing much profitability. Within American society, the business behavior of many small retailers appears similarly motivated. Wittreich (1962) found owners of small proprietorships were often less concerned with profitability than achieving the feeling of status, prestige, or family security through owning their businesses.

An example of marketing failure related to inappropriate cultural stereotyping is provided by Kitano (1969) In writing about the Japanese-American sub-culture he comments (p. 100), "To many people, Japanese culture suggests mainly tea ceremonies and flower arranging. As often happens, the quaint, the unusual, and even the trivial become so firmly associated in people's minds that complex explanations of behavior are often built upon pursuits that may in fact represent only the interests of a select population." Cultural stereotypes and folklore can be deceiving and perhaps even harmful to the marketer. He should guard against both the possibility of his own inaccurate perceptions of other cultures, or mistakenly assuming the attribution by other cultures of a favorable or specific stereotype of Americans. Chinese visitors to this country are surprised to learn that Americans attribute chop suey to the Chinese, as are Italians upon learning we consider pizza to be an Italian food (Kassarjian and Robertson, 1973).

THE ROKEACH PARADIGM

One approach which might be particularly well suited for assessing values has been developed by Rokeach (1968-69), who views values as having to do with preferable modes of conduct and end-states of existence. The distinction between modes of behavior and end-states of existence implies a differentiation between means and ends or between what Rokeach calls "instrumental" and "terminal" values. Instrumental values relate to modes of conduct and represent a single belief which is personally and socially preferable in all situations with respect to all objects (e.g., ambition, independence and self control). Terminal values are a single belief that some end-state of existence is personally and socially worth striving for (e.g., leading an exciting life, family security and pleasure).

The ability of the Rokeach paradigm to differentiate the value orientations of various domestic groups has been very impressive (Rokeach, 1973). For example, various combinations of terminal and instrumental values have been found which significantly differentiate men from women, "hawks" from "doves," policemen from unemployed blacks (Rokeach, 1968-69). Additionally, some research has been done on the validity and reliability of the RVS itself. Both Rokeach (1973) and Robinson and Shaver (1971) reported relatively high test-retest reliability coefficients for the instrument over three week intervals. Vinson, Munson, and Nakanishi (1977) derived factor structures which clearly indicate the existence of two distinct value categories. A factor analysis of interval scaled importance ratings for each of the items in the RVS yielded 6 factors composed of terminal and 5 factors composed of instrumental values. With only one exception, there was no overlap of instrumental values loading on terminal factors or vice versa.

HYPOTHESES

It was felt that an adequate demonstration of the utility of the RVS in conducting cross cultural research could be achieved by testing its ability to discriminate between groups which were known to represent differing cultural backgrounds. An attempt was made to include both groups whose cultural backgrounds were assumed to be very different from that of a baseline American group and a second group whose heritage was assumed to be distinct from the American culture, but perhaps somewhat less so. These considerations and the availability of subjects resulted in the selection of Thais and Mexicans. Additionally, evaluations of a common object, the average American stereotype held within each culture, was felt to provide a further test of the sensitivity of the RVS to distinguish perhaps more subtle differences. The following null hypotheses were developed:

H1: Across cultures there will be no differences in the personal value systems of Thais, Mexicans or Americans.

H2: Across cultures there will be no difference in the personal value system each culture attributes to an average American (American Stereotype).

H3: Within a culture, there will be no difference between that cultures' own value system and the stereotype value system it attributes to an average American (American Stereotype).

While it was strongly anticipated that self value structures might differ markedly between cultures, it was less clear how evaluations of the personal values of the stereotype of an average American might vary. It was felt that cultural norms might serve to not only identify acceptable behaviors, but also to identify acceptable attitudes toward members of another culture. One might expect the pressure on an individual to be very great to conform to his own cultural's stereotype of another culture. Acting incorrectly, holding the "wrong" stereotype, could meet with social disapproval. One might also anticipate greater homogeneity of the perceptions found within a specific culture than between diverse cultures regarding perceptions of a common third object. Such perceptual homogeneity within culture and heterogeneity across culture could reflect, in part, a) the individual's acceptance of important norms within his own culture, b) cultural perceptual biases and/or c) differences in the information available to members of different cultural backgrounds.

METHOD

Sample and Instrument

Subjects were drawn from three separate populations. The Thai sample consisted of 32 people selected from a listing of 162 members of the Thai Students' Organization in Los Angeles. A sample of 63 Mexicans was obtained from the members of several Mexican-American organizations in and around the greater Los Angeles area. A sample of 75 Americans was obtained from graduate business students at a large Western university. An attempt to match the samples on major demographics was made. The three samples were very similar with respect to age (25-35 years) and education level (college graduates). A major difference occurred for religion, with the Thai group being 82% Buddhist, while the Mexican group was 85% Catholic, and the American group was primarily Catholic and Protestant.

Ail respondents completed a questionnaire containing the 36-item RVS and a separate page requesting traditional demographic data. The perceived importance of each value was measured using a seven point Likert-type scale. Each subject provided information on the importance he attributed to each of the 36 values (Self evaluation), as well as his perceptions of the importance placed on each value by an average American (Thai-American stereotype, Mexican-American stereotype and American-American stereotype).

RESULTS AND DISCUSSION

Given that one of the main objectives of this study was determination of the capability of personal values as measured by the RVS to reveal cultural differences, stepwise discriminant analysis was employed. The criteria used for controlling the stepwise entry of variables was both the minimum Wilks' lambda and the maximum Mahalanobis distance between groups. The Mahalanobis criterion seeks to maximize the Mahalanobis distance between the two closest groups. The Wilks' lambda criterion is the overall multivariate F ratio for the test of differences among group centroids. The variable which maximizes the F ratio also minimizes Wilks' lambda, a measure of group discrimination.

Ten of the original 36 values were selected using this stepwise procedure (Table 1). The ten values produced a very high degree of separation between the six classification groups as indicated by the final Wilks' lambda (.205), and the canonical correlations for the first five discriminant functions. The square of the canonical correlation indicates that the first discriminant function explains approximately 49%, the second, 30% and the third, 24% of the variance in the grouping variables. Thus the first three functions are highly correlated with the groups while the last two are moderately correlated.

The eigenvalue is a measure of the relative importance of each function. The sum of the eigenvalues is a measure of the total variance existing in the discriminating variables. Expressing each eigenvalue as a percentage of the total sum of eigenvalues indicates the relative importance of the associated function and is shown in the column designated Relative Percentage (Table 1). The Wilks' lambda associated with each function is an inverse measure of the discriminating power in the original variables which has not yet been removed by the discriminant functions. Wilks' lambda was .892 after the 5th discriminant function was derived, which approximates a chi-square of 37.74 (p<.001). A lambda of this magnitude or smaller has a p<.001 of occurring due to chances of sampling even if there was no further information to be accounted for by the fifth discriminant function.

The relative contributions of the 10 discriminating variables to the respective functions can be seen in Table 2. The first function would seem to primarily represent individual characteristics with an emphasis on the values "broadminded" and "sense of accomplishment'' which are over 1.5 times as important as "equality'' or "responsibility." F2 is largely composed of "obedience" and "salvation," both making negative contributions. F3 appears most related to "true friendship" and "responsibility," both making negative contributions. Several variables make approximately equal contributions to F4, with "national security" and "responsibility" being the largest negative and "ambitious'' the largest positive contribution. F5 is predominately related to "equality."

Discriminant Classification Table

After a discrimination procedure had been established it was of interest to determine whether the discriminant function was useful in classifying people using their scores on the personal values into their own cultural group. The classification matrix (Table 3) provides a convenient method of summarizing the number of correct and incorrect classifications made by the discriminant function. The diagonal elements denote the number of correct classifications (hits) and the off-diagonal elements denote the number of incorrect classifications (misses).

The classification power of the discriminant procedure is quite encouraging in that 65% of the cases are correctly identified as compared to an expected 22% if group assignment were to the most frequent category (i.e., the maximum chance criterion). Morrison (1969) argues in favor of a more lenient chance model which he calls the proportional chance criterion. Under this latter chance model, the chance criterion is adjusted to take into account the fact that some classifications are being made to the smaller groups. In our case, this proportional chance criterion would yield a 13% chance hit rate. However, we utilize the more conservative criterion of 22% in light of the fact that the Thai sample size was judged to be insufficient to permit the use of both an estimation and validation sample. To statistically test whether the 65% hit rate was better than the 22% maximum chance criterion a Q statistic was calculated (Press, 1972). The Q value exceeded a value of X2 = 10.82 (p<.001), hence the classification procedure does significantly better than chance.

Hypothesis 1, 2 and 3 regarding the differences in the self value and stereotype value structures between cultures were tested by comparing the centroids derived for each group in the five dimensional discriminant space. The centroids represent mean discriminant scores for each group on the respective discriminant functions.

TABLE 1

DISCRIMINANT ANALYSIS SUMMARY TABLE

TABLE 2

STANDARDIZED DISCRIMINANT FUNCTION COEFFICIENTS

The F ratios based on the Mahalanobis distances between all pairs of group centroids were significant beyond the p<.001 level. Therefore one can reject the hypotheses that the value profiles from any two groups are equal. The significance of these findings suggest at least the following conclusions:

1. Self value systems do differ markedly across the three cultures studied.

2. Conceptions of American stereotypes also differ markedly across cultures. Apparently there is not one universally held American stereotype.

3. Within cultures there are significant differences between self value systems and those values attributed to an American stereotype.

Given that the primary thrust of this research was on the ability of personal values to discriminate various cultural groups, and the applicability of the RVS to this task, it was not the intent to focus on differences in specific isolated value items. However, readers interested in making specific comparisons are referred to Table 4 which presents mean importance ratings on both self value and American stereotype value structures for each culture. The univariate F ratios indicate that at least one pair of the groups are significantly different on each value. Some interesting differences in the evaluations of both instrumental and terminal values are observable across the three groups. For example, Thais rate the terminal values of "a comfortable life" and an "exciting life" as significantly less important than either Americans or Mexicans. However they attach significantly greater importance than Americans to those terminal values related to social harmony -that is, "world at peace," "national security" and "family security." On the other hand, Thais attach significantly less importance than Americans to the instrumental values of "ambition" and "imagination."

CONCLUSION

Personal values represent the most fundamental element of the consumer's cognitive world. They enable the individual to structure his perception and understanding of himself, of significant others and of the objects embodied within his psychological reality. Inasmuch as it has been repeatedly observed that values are antecedent to many types of behavior, one might expect differential value orientations to manifest themselves in differential overt behavior. Knowledge of the value orientations held by a specific group may help in understanding, explaining, or perhaps predicting subsequent attitudes and behavior. However there is a limited empirical research, outside the current study, which focuses on either value measurement or their potential useful-nesses to those engaged in international marketing.

The findings of this study indicate that personal values can be productively employed to discriminate peoples with culturally diverse backgrounds. Significant differences in the self value orientations were observed between three cultural groups: Thais, Mexicans, and Americans. Significant differences were also found between the three groups in the stereotype each culture holds of an average American's personal value system. Additionally, within each culture significant differences were observed between a person's self values and those he attributed to an average American. Discriminant functions derived from responses to the Rokeach Value Survey were able to correctly classify 65% of the people according to their cultural background, significantly better than the 22% expected by chance.

TABLE 3

CLASSIFICATION MATRIX

Because of the robust ability of values to distinguish among peoples of different culture, value analysis may represent a convenient and economical technique to aid in market segmentation strategies. The RVS might offer marketers and researchers a valuable tool for identifying normatively based cross cultural differences. It is easily administered and could be employed to develop a profile of a general population or profiles of individual segments prior to the application of other more expensive and complex segmentation techniques. Value analysis should become an increasingly important area of research for the international marketer.

TABLE 4

SUMMARY OF MEAN IMPORTANCE VALUE RATINGS FOR THAIS, MEXICANS, AMERICANS AND AMERICAN STEREOTYPES WITH UNIVARIATE F-TESTS

REFERENCES

Fred E. Fiedler, Terence R. Mitchell and Harry C. Triandis, "The Culture Assimilator: An Approach to Cross Cultural Training," Journal of Applied Psychology, Vol. 55, No. 2 (1971), 95-102.

Russell J. Haley, "Benefit Segmentation: A Decision-Oriented Research Tool," Journal of Marketing, (July 1968), 30-35.

Waiter A. Henry, "Cultural Values Do Correlate with Consumer Discontent," Journal of Marketing Research, 13 (May 1976), 121-127.

Harold H. Kassarjian and Thomas S. Robertson, Perspectives in Consumer Behavior, (Glenview, Illinois: Scott Foresman and Co., 1973).

Harry H. L. Kitano, Japanese Americans, (Prentice-Hall, 1969).

P. V. Lessig, "A Measurement of Dependencies Between Values and Other Levels of the Consumer's Belief Space," Journal of Business Research, 3 (July 1975), 227-241.

Donald G. Morrison, "On the Interpretation of Discriminant Analysis," Journal of Marketing Research, 6 (May 1969), 156-163.

J. Michael Munson, "An Investigation of the Rokeach Value Survey for Applications to Cross Cultural Research," in J. Stolen and J. Conway, eds., Proceedings of the American Institute of Decision Sciences, 1977.

Joseph Pernica, "The Second Generation of Market Segmentation Studies: An Audit of Buying Motivation," in W. D. Wells, ed., Life Style and Psychographics, (Chicago: American Marketing Association, 1974), 277-313.

Joseph T. Plummer, "Life Style Patterns and Commercial Bank Credit Card Usage," Journal of Marketing, (April 1971), 35-41.

S. James Press, Applied Multivariate Analysis, (New York: Holt Rinehart and Winston, 1972).

J. P. Robinson and P. R. Shaver, eds., Measures of Social Psychological Attitudes, (Ann Arbor: Institute for Social Research, 1971), 463.

Milton J. Rokeach, "The Role of Values in Public Opinion Research," Public Opinion Quarterly, (winter 1968-69), 547-549.

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

Milton J. Rosenberg, "Cognitive Structure and Attitudinal Affect," Journal of Abnormal and Social Psychology, 53 (November 1956), 367-72.

J. E. Scott and L. H. Lamont, "Relating Consumer Values to Consumer Behavior: A Model and Method for Investigation,'' in Thomas V. Greer, ed., Increasing Marketing Productivity, (Chicago: American Marketing Association), 283-288.

Donald E. Vinson and J. Michael Munson, "Personal Values: An Approach to Market Segmentation," in K. L. Bernhardt, ed., Proceedings, Marketing: 1776-1976 and Beyond, (Chicago: American Marketing Association, 1976), 313-317.

Donald E. Vinson, J. Michael Munson and Masao Nakanishi, "An Investigation of the Rokeach Value Survey for Consumer Research Applications," in W. O. Perreault, ed., Proceedings, Association for Consumer Research, 4 (1977), 247-252.

Donald E. Vinson, Jerome E. Scott and Lawrence M. Lamont, "The Role of Personal Values in Marketing and Consumer Behavior," Journal of Marketing, 41 (April 1977), 44-50.

William D. Wells and Douglas J. Tigert, "Activities, Interests and Opinions," Journal of Advertising Research, (August 1971), 27-34.

Warren J. Wittreich, "Misunderstanding the Retailer," Harvard Business Review, 40 (May-June 1962), 147-159.

Steven Worchel and Terence R. Mitchell, "An Evaluation of the Effectiveness of the Culture Assimilator in Thailand and Greece," Journal of Applied Psychology, Vol. 56, No. 6, (1972), 472-479.

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