Fictional Materialism

William D. Wells, University of Minnesota
Cheri L. Anderson, University of Minnesota
ABSTRACT - Richins(1994) found that materialism, as measured by Richins and Dawson's (1992) Materialism scale, mediates the meanings of consumers' prized possessions. The present study tested, confirmed and extended that finding by observing and analyzing the possessions and behaviors of 71 fictional characters in well-known TV comedies and dramas. This outcome shows that television stories, and presumably other popular narratives, can supplement and complement more standard research data.
[ to cite ]:
William D. Wells and Cheri L. Anderson (1996) ,"Fictional Materialism", in NA - Advances in Consumer Research Volume 23, eds. Kim P. Corfman and John G. Lynch Jr., Provo, UT : Association for Consumer Research, Pages: 120-126.

Advances in Consumer Research Volume 23, 1996      Pages 120-126

FICTIONAL MATERIALISM

William D. Wells, University of Minnesota

Cheri L. Anderson, University of Minnesota

ABSTRACT -

Richins(1994) found that materialism, as measured by Richins and Dawson's (1992) Materialism scale, mediates the meanings of consumers' prized possessions. The present study tested, confirmed and extended that finding by observing and analyzing the possessions and behaviors of 71 fictional characters in well-known TV comedies and dramas. This outcome shows that television stories, and presumably other popular narratives, can supplement and complement more standard research data.

INTRODUCTION

In "Special Possessions and the Expression of Material Values," Marsha L Richins (1994) administered the Richins and Dawson (1992) Materialism scale to a large sample of real consumers. She divided her respondents into upper and lower quartiles, and analyzed the public and private meanings of high materialists' and low materialists' most prized possessions.

The present study followed this pattern with three notable exceptions: (1) The "consumers" were fictional characters in prime time television comedies and dramas. (2) The Materialism scale was administered by proxy. (3) Judges identified the "consumers"' prized possessions.

METHOD

Richins (1994) mailed questionnaires to "random samples of 400 households in a small northeastern city and 300 households in a more rural area of the northeast." She received replies from "144 urban and 119 rural respondents, yielding response rates of 36.0 percent and 39.6 percent, respectively" (Richins 1994, 524). In an early part of her questionnaire she asked, "Many people have a few possessions that they care a lot about or that are especially important to them. In the spaces below list your most important possessions and explain why each is important to you"(104). A later part of the questionnaire included the Richins and Dawson (1992) Materialism scale.

Our procedure was somewhat different. We started with the 74 ongoing characters in the 19 highest Nielsen-rated prime time television comedies and dramas. For each character, we asked panels of at least four judges who were blind to the purpose of the investigation, "What are this person's most valued possessions?" and "What are the reasons he or she values them?" We asked different panels of at least four judges who also were blind to the purpose of the investigation to complete the Richins and Dawson (1992) Materialism scale as they thought each character would answer it. When we checked interjudge reliability, we found that, for three of the characters, agreement was not high enough to yield a stable estimate. We dropped those three characters, leaving 71 for this analysis.

Richins (1994) asked judges to sort her respondents' most valued possessions along a series of dimensions: social visibility, estimated market value, type of possession and private meaning. She also asked judges to decide, for each possession, whether people who own that possession are "likely to be materialistic or not materialistic" (530). On the basis of their Materialism scale scores, she divided her respondents into quartiles. She then compared the most valued possessions of the "high materialists"-the top quartile-with the most valued possessions of the "low materialists'~--the bottom quartile-on all the rated and judged dimensions. We followed those procedures as closely as possible.

FINDINGS

Materialism Scale Statistics

Richins did not report the mean and standard deviation of the Materialism scores from her (1994) respondents. However,Richins and Dawson (1992) reported means ranging from 46 to 48 and standard deviations ranging from eight to ten from their scale-validation samples. In the present study, the 71 TV characters averaged 54 on the Materialism scale, with a standard deviation of 14. Thus, the TV characters appear to be slightly more materialistic, and more varied in materialism, than Richins and Dawson's (1992) samples of real persons. We will comment on those statistics later.

Richins (1994) reported Materialism scale alphas of .84 and .86 from her two samples. Richins and Dawson (1992) reported alphas of .80 to .88. In the present study, the Materialism scale's alpha was .98-significantly higher. We will comment on that statistic later.

Hypotheses

From previous research, Richins (1994) developed, tested and confirmed four sets of hypotheses concerning differences between high materialists and low materialists:

H1: Consumers with different levels of materialism will value different types of possessions. Compared to the possessions valued by consumers low in materialism, possessions valued by high-materialism consumers (a) tend to be publicly (rather than privately) consumed, (b) are more expensive, and (c) are less likely to be associated with important others.

H2: The private meanings of possessions valued by high-materialism consumers (as compared to those valued by low materialism consumers) are (a) less likely to concern the possession's role in representing or facilitating interpersonal ties and (b) more likely to relate to the financial worth of the possession.

H3: The public meanings of possessions value by high materialism consumers (as compared to those valued by low-materialism consumers) are more likely to refer to success or prestige.

H4: The possessions valued by those low in materialism and those high in materialism will be consistent with socially constructed stereotypes of nonmaterialistic and materialistic consumers, respectively (Richins 1994, 523-524).

Using our fictional respondents and their "most valued" possessions (see Table 1), we retested these relationships.

Social Visibility

In the original investigation, judges sorted possessions along a "social visibility" continuum that ran from (1) used in private, through (2) displayed in the home and (3) displayed on the person, to (4) used in a public place. As expected, she found that high materialists' most valued possessions were significantly (p.<05) more visible than low materialists' most valued possessions. Our data showed exactly the same pattern (p.<.01, Table 2). If anything, the relationship in the TV data is a little stronger.

TABLE 1

SOME OF THE POSSESSIONS VALUED BY LOW AND HIGH MATERIALISTS IN THE TV STORIES

TABLE 2

SOCIAL VISIBILITY OF LOW MATERIALISTS' AND HIGH MATERIALISTS' MOST VALUED POSSESSIONS

Estimated Market Value

In the original investigation, judges estimated the possessions' market value. The high materialists' most valued possessions were significantly (p.<01) more expensive than the low materialists' most valued possessions. Again, the fictional data replicated the pattern (p.<01, Table 3).

Type of Possession

In the original investigation, judges classified possessions into types: sentimental objects, practical objects, aesthetic objects, financial assets, extensions of the self, and possessions used for: transportation, recreation, and personal appearance (Richins 1994, Table 2).

In Richins' study, high materialists were significantly less likely to value sentimental and recreational possessions, and significantly more likely to value transportation, personal appearance and financial possessions. Expected differences for extensions of the self, practical objects and aesthetic objects were not statistically significant.

In the fictional replication (Table 4) high materialists were significantly more likely to value transportation (p.<01) and personal appearance (p.<00) possessions. As in the original study, high materialists did not differ significantly from low materialists in valuing extensions of the self. The remaining categories contained too few entries to permit firm conclusions. Thus, the fictional replication confirmed most, but not all, of Richins' findings about high materialists' and low materialists' valuation of various types of possessions. Failures to confirm were traceable to low numbers of observations. They were not statistically significant contradictions.

TABLE 3

ESTIMATED MARKET VALUE OF LOW MATERIALISTS' AND HIGH MATERIALISTS' MOST VALUED POSSESSIONS

TABLE 4

TYPES OF POSSESSIONS VALUED BY LOW MATERIALISTS AND HIGH MATERIALISTS

Private Meanings

Richins(1994) asked respondents to explain why their possessions were important. We asked judges who were blind to the TV characters' imputed Materialism scores and blind to the purpose of the experiment, "What are this person's most valued possessions?" and "What are the reasons he or she values them?"

Two of Richins' findings reemerged: In the replication, high materialists were significantly more likely to value possessions for financial and appearance- related reasons (both at p.<01, Table 5). Two of Richins' findings did not reemerge. We did not find significant differences in valuing possessions for their personal connotations, or for their "utilitarian" functions. One of Richins' predictions, supported but not significant in the original, was supported and significant (p.<001) in the replication. High materialists were less likely than low materialists to value possessions for identity reasons (Table 5).

Thus, even though imagining the motives of fictional characters is at best uncertain, the fictional replication confirmed several of Richins' (1994) predictions about "private meanings."

TABLE 5

PRIVATE MEANINGS OF LOW MATERIALISTS' AND HIGH MATERIALISTS' MOST VALUED POSSESSIONS

TABLE 6

JUDGED MATERIALISM OF POSSESSIONS AND MEASURED MATERIALISM OF POSSESSIONS' OWNERS

"Materialistic" and "Nonmaterialistic" Possessions

Richins (1992) listed her respondents' possessions on cards, and asked 30 judges to separate the cards into "whether you think the person who mentioned the item is likely to be materialistic or not materialistic." She also permitted "can't determine" (530). Considering only those possessions for which her judges showed at least 60% agreement, she found a significant (p.<.001) positive relationship between position on the Materialism scale and ownership of stereo-typically "materialistic" possessions. This relationship reappeared in the fictional replication (Table 6).

Extension

One way to enhance the construct validity of an instrument like the Materialism scale is to contrast the personality characteristics of those who score high on it with the personality characteristics of those who score low on it. In the present study, sets of (at least four) judges who were blind to imputed Materialism scale scores and to the purpose of the investigation rated the TV characters on a 41-item adjective check list. Considering only those traits that more than half the judges checked for a given character, the adjectives showed some diagnostic differences (Table 7).

TABLE 7

TRAITS ATTRIBUTED TO LOW-MATERIALISM AND HIGH MATERIALISM TV CHARACTERS IN THE FICTIONAL REPLICATION

Table 7 shows that the low-materialism TV characters were rated more hard-working, good-natured, calm, broad-minded, cooperative, curious, imaginative and warm-and less impulsive, emotional, insecure, and irresponsible-than their high materialism counterparts. These findings enrich the construct validity of the Materialism scale, and are congruent with descriptions of materialists in other investigations (Belk 1985, McKeage 1922, Micken 1992, Williams and Bryce 1992).

Factor Analysis of the Materialism Scale

In the course of developing their Materialism scale, Richins and Dawson (1992) factor-analyzed the scale items. They identified three subscales-Centrality, Success and Happiness--that tap into separate facets of the overall dimension.

When we factor-analyzed the TV data, the eigenvalues implied a two-factor solution (Table 8). In a varimax rotation, the first factor merged Centrality and Success, and accounted for 77% of the variance. The second factor, Happiness, accounted for 7% of the variance. None of the succeeding factors accounted for more than 3% of the variance.

When we forced a three-factor solution, the first factor was predominantly Success, with two low-loading Centrality items. The second factor was entirely Happiness. The third factor included the remaining Centrality statements (Table 8). As would be expected from the factor analysis, raw score (not factor score) correlations among Success, Centrality and Happiness were high and positive: Success-Centrality,.93;Success--Happiness,.83; Happiness--Centrality, .78.

One possible interpretation of this outcome is that judges' ratings of TV characters are less fine-grained than self-revelations by real persons. It seems reasonable to assume that judges form and use overall impressions, while self evaluators are more responsive to separate nuances of individual scale items. In the present study, the Materialism scale's alpha of .98, noted earlier, supports this interpretation.

However, it is also possible that conceptually distinct facets of Materialism are positively (and maybe highly) correlated, even in self-evaluations. The alphas of .84 and .86 reported by Richins (1994) and the alphas of .80 and .88 reported by Richins and Dawson (1992) support that conclusion. If that conclusion is correct, future users of the Materialism scale will find that differences among subscales are small and relatively unreliable.

Scale Parameters

The New England adults in Richins and Dawson's (1992) standardization samples averaged 46 to 48 on the Materialism scale. The TV characters averaged 54. Fifty-four is significantly (p.<001) higher than the standardization means, and more in line with the 58 that McKeage (1992) obtained from two samples of New England college students.

McKeage (1992) concluded that his respondents were "a bit more materialistic" (141) than Richins and Dawson's (1992) adults. There is, however, another possible explanation. In Richins and Dawson's (1992) surveys, response rates were 31 to 43 percent. If, among those who received Richins and Dawson's questionnaires in the mail, the low materialists (like the low materialists in the TV stories) were more hard-working, good-natured, cooperative and curious-and less insecure and irresponsible-than the high materialists, they would be more likely to respond to the researchers' request for assistance. If this speculation is correct, low materialists were over- represented, and high materialists were under-represented, in Richins and Dawson's (1992) standardization samples and in Richins' (1994) study of possessions. We do not know whether this explanation is correct. It is only a speculation.

TABLE 8

FACTOR ANALYSIS OF THE MATERIALISM SCALE (DATA FROM THE FICTIONAL REPLICATION)

DISCUSSION AND COMMENT

We set out to extend the notion that television stories (and presumably other fictional narratives such as novels, movies, and theatrical productions) are useful sites for research on consumer behavior. An earlier "fictional replication" (Wells and Gale 1994) had shown that analysis of TV stories can complement and supplement a real-life ethnographic investigation (Otnes, Lowrey and Kim 1993). The present study focused on a measured individual difference variable-materialism, and the correlations between it and the public and private meanings of possessions.

The earlier "replication" supported many of the ethnographic findings. It was conducted by independent investigators using sharply different methods. It therefore furnished strong convergent validation (Campbell and Fiske 1959).

The earlier "replication" also extended the ethnographic findings. Because the TV narratives provided a large and diverse sample of "respondents" and situations, they yielded some findings--supported by other research (Belk and Coon 1995; Sherry 1983; Sherry, McGrath and Levy 1992)--that did not appear in the original investigation.

The present "replication" also supported many of its model's conclusions. Again, because it was conducted by independent investigators using sharply different methods, it provides strong conceptual confirmation.

The present "replication" was also an extension. Adjective check list ratings of the TV characters added meaning to the distinction between high and low materialists. Factor analysis of the Materialism scale raised cautions for future users of that

instrument. An assumption that high and low materialists in real life are similar to high and low materialists in the TV stories provided a possible explanation for the comparatively low Materialism scores among Richins and Dawson's (1992) and Richins' (1994) mail questionnaire respondents. These findings and speculations seem important enough to merit further investigation.

What's going on here? An explanation that appeals to us is that writers, directors and producers of successful TV comedies and dramas, and the actors who play popular TV characters, are accurate observers human nature. If this explanation is correct, their insights (and the insights in novels, plays, biographies and cinema) may tell us something useful about the real behavior of real consumers.

Of course fictional narratives are not mirrors (Cantor 1992, Feuer 1992, Vande Berg and Streckfuss 1992). Instead, they extract, abbreviate and amplify relationships that also occur in real behavior.

In this respect, fictional narratives resemble surveys and experiments. Surveys and experiments are not natural segments of everyday life. They extract, abbreviate and amplify relationships that carry over into real behavior. The trick is to understand what they mean, and to make valid extrapolations from them.

At this point we know very little about how to make valid extrapolations from fictional narratives. However. "fictional replications" of an ethnographic study (Otnes et al. 1993) and a psychometric study (Richins 1994) encourage further exploration. In both "replications," imaginary behavior of fictional characters complemented and supplemented real behavior of real consumers. In both "replications," fictional data from fictional narratives supported and extended real data from empirical investigations.

This experience suggests that fictional narratives can enrich

our understanding of the real behavior of real consumers. In some sense art imitates life. The trick will be to use that imitation to build bridges between the wisdom resident in the humanities and the empirical data of the social sciences.

REFERENCES

Belk, Russell W. (1985), "Materialism: Trait Aspects of Living in the Material World," Journal of Consumer Research, 12 (December), 265-280.

Belk, Russell W. and Gregory S. Coon (1993), "Gift Giving as Agapic Love: An Alternative to the Exchange Paradigm Based on Dating Experiences," Journal of Consumer Research, 20 (December), 393-417.

Campbell, Donald T. and Donald W. Fiske (1959), "Convergent and Discriminant Validation by the Multitrait-Multimethod Matrix," Psychological Bulletin, 56 (2), 81-105.

Cantor, Muriel G. (1992), "Prime-Time Television: Content and Control, Beverley Hills, CA and London: Sage Publications.

Feuer, Jane (1992), "Genre Study and Television," in Channels of Discourse, Reassembled, Robert E. Allen ed., Chapel Hill NC: University of North Carolina Press.

McKeage, Kim K. R. (1992), "Materialism and Self Indulgence: Theories of Materialism in Self-Giving," in Meaning, Measure and Morality of Materialism, Floyd Rudmin and Marsha Richins eds., Provo, UT: Association for Consumer Research, 140-148.

Micken, Kathleen S. (1992), "Materialism Research: Suggestions for New Directions," in Meaning, Measure and Morality of Materialism, Floyd Rudmin and Marsha Richins eds., Provo, UT: Association for Consumer Research, 121125.

Otnes, Cele, Tina M. Lowrey and Young Chan Kim (1993), "Gift Selection for Easy and Difficult Recipients: A Social Roles Interpretation," Journal of Consumer Research, 20 (September), 229-244.

Richins, Marsha L. (1994), "Special Possessions and the Expression of Material Values," Journal of Consumer Research, 21 (December), 522-533.

Richins, Marsha L. and Scott Dawson (1992), "A Consumer Values Orientation for Materialism and Its Measurement: Scale Development and Validation," Journal of Consumer Research, 19 (December), 303-316.

Sherry, John F., Jr. (1983), "Gift Giving in Anthropological Perspective," Journal of Consumer Research, 20 (September), 157-168.

Sherry, John F., Mary Ann McGrath and Sidney J. Levy (1992), "The Disposition of the Gift and Many Unhappy Returns," Journal of Retailing, 68 (Spring), 40-65.

Vande Berg, Leah R. and Diane Streckfuss (1992), "Prime-time Portrayal of Women and the World of Work: A Demographic Profile," Journal of Broadcasting and Electronic Media 36 (Spring), 195-208.

Williams, John, and Wendy Brice (1992), "Materialism and Care for Others," in Meaning, Measure and Morality of Materialism, Floyd Rudmin and Marsha Richins eds., Provo, UT: Association for Consumer Research, 149-154.

Wells, William D. and Kendra L. Gale (1994), "Fictional Subjects in Consumer Research," in Frank Kardes and Mita Sujan eds., Advances in Consumer Research Vol. 2 1, Provo, UT: Association for Consumer Research.

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