Gender-deconstructing consumption: Scale development and validation.

Larsen, Val ; Wright, Newell D.

INTRODUCTION

Among the many social and historical changes that have occurred in this century, none is more important or has more far reaching implications than changes in the role and economic status of women. Unsurprisingly, these changes have had and will continue to have a profound effect on business and public policy. It is, therefore, imperative that businesspeople, consumer advocates, and public policy makers attend to developments in this area. This article provides a scale which can be used to monitor changes in consumption preferences that flow from the evolution of gender ideology.

THE CONSTRUCTION OF GENDER

The idea that women are inferior and should be subordinate has a history sufficiently long that it is deeply embedded in both language and custom. Its more subtle manifestations are, therefore, difficult to detect and even more difficult to change (Smith, 1987). To take just one example, women's sphere of influence and range of career opportunities have been limited by the restriction on women in combat, a restriction framed as a beneficent protection, but a protection that no woman could wave if she wished to obtain the opportunities that were its price. Asymmetries encoded in language are, if anything, even more difficult to unmask than those which take the form of custom (Larsen, 1993).

Fortunately, Jacques Derrida (1973; Culler, 1982), a prominent post-structuralist philosopher, has provided an insight which uncovers some of the hidden asymmetries. Language, he points out, is replete with dichotomous and apparently equivalent pairs: man/woman, male/female, boy/girl, husband/wife, mind/body, north/south, positive/negative, rich/poor, competent/incompetent. Derrida's insight is that the equivalence implied by the pairing of these terms masks an actual asymmetry. In each case, the first term in the pair is preferred or "valorized." This valorization is revealed in the fact that it sounds odd to reverse the order of the terms when they are used together, i. e. say south and north, poor and rich, female and male. It is revealed, too, in the fact that the first term reflects the historical locus of power or value. Historically, the north has been more prosperous and powerful than the south (in both the US and the world), the rich more comfortable and respected than the poor, men more honored and obeyed than women. People in power have constructed the language code in such a way that it hides these asymmetries behind a falsely implied balance and, thus, makes them seem natural. Derrida and other post-structuralists unmask the conventionality of these distinctions, showing that they are rooted not in nature but in group interests and the exercise of group power.

Consumer researchers have long since pointed out that consumption, like these verbal dichotomies, is heavily gender coded (Courtney & Lockeretz, 1971; Davis & Rigaux, 1974; Gentry & Doering, 1977). It has more recently become apparent that the asymmetries uncovered by Derrida's linguistic analyses are manifest in gender coded consumption as well. For instance, Fontenelle and Zinkhan (1992) suggest that the experience of leisure is asymmetrical for women and men. The leisure of a woman is likely to be tied to that of her partner and children so that even on vacation, her needs tend to be subordinated to the needs of her family. The leisure of men, on the other hand, tends to be relatively free from constraint. In a similar vein, Wallendorf and Arnould (1991) found that Thanksgiving Day, a day of leisure and relaxation for men, often involves very hard work for women, who must cook the ritual feast, then clean up afterward. And Fischer and Arnold (1990) have shown that family Christmas shopping tends to be framed as a compulsory duty for women but as an optional pleasure for men.

Since language is at root a system of differences (Saussure, 1959), linguistic and social dichotomies are probably unavoidable. However, asymmetries in both language and custom can be "deconstructed" as well as constructed. That is, dichotomies can be reframed in such a way that they embody genuine balance or, even more instructively, in such a way that the once marginal term (or gender) becomes valorized. When the latter happens, the entire network of assumptions and socioeconomic relationships is cast in a new light, a light that discloses the hidden agenda and interests of those who created the original asymmetrical dichotomy (Culler, 1982, p. 150). A striking consumer behavior example of the more radical move is the recent actions of the Barbie Liberation Organization. This group surreptitiously switched the voice boxes on G.I. Joe and Barbie dolls. As a result, girls around the United States heard their Barbies proclaim, "Dead men tell no lies"; boys heard G.I. Joe perkily ask, "Want to go shopping?" (Associated Press, 1993). In this reversal, women are coded as powerful and violent mistresses of their domain. They are active and produce a ghoulish product--dead bodies. Men are coded as socially engaged consumers who are passive, compliant, and somewhat frivolous.

The practical, day-to-day importance of gender deconstruction in the consumer domain has recently been made clear by Frances Cerra Whittelsey (1993). Citing one example after another, Whittelsey shows that distinctions in the marketplace between male and female tend to be costly for women. She reports, for instance, that a given style of knit shirt almost always costs more in the women's department of a clothing store than in the men's. And laundering the shirt at a dry cleaner is also more expensive for women, traditionally about 27 percent more for the same style of shirt. In New York, women generally pay more than men for a basic shampoo, cut, and blow dry at a hair salon. And though women buy about half of all new cars in the United States, they are not treated equitably by dealers. Compared with white men, white women pay, on average, about $150 more and black women about $800 more for the same car. In Whittelsey's analysis, the valorization of the male in the marketplace produces a cash bonus for males and a cash penalty for females. Thus, women have a very tangible economic interest in deconstructing the existing male/female distinction. Unfortunately, for the time being, most women and supportive men too seldom make the consumption choices that would promote this interest.

MEASURES OF GENDER

Since it is a social and linguistic construct, not a biological attribute, gender cannot be equated with sex. It is most often defined as a personality disposition and most often measured with the Bem Sex-Role Inventory (BSRI). The BSRI can measure the extent to which a given woman or man's personality deconstructs or constructs the gender distinction. Thus, Frable (1989) has used the BSRI to divide people into four groups: sex typed (feminine women, masculine men), androgynous (both feminine and masculine), undifferentiated (neither feminine nor masculine), and cross-sex-typed (masculine women, feminine men). Sex typed individuals validate the traditional link between sex and gender in their own personality and in their interactions with and judgements of others (Anderson & Bem, 1981). Androgenous, undifferentiated, and, most radically, cross-sex-typed individuals deconstruct the traditional link between sex and gender (Frable, 1989).

Useful as it is, the BSRI has had limited power to predict behaviors of interest. As Bem herself has pointed out, "the literature is littered with failed attempts to use the [BSRI] and Personal Attributes Questionnaire (PAQ) to predict a wide variety of gender-related attitudes and behaviors" (Bem, 1985, p. 196). Commenting on consumer behavior in particular, Meyers-Levy (1988, p. 522) has made a similar point: "Consumer behavior researchers have met with only limited success in relating sex roles to product perceptions." Frable (1989) argues that these failures may result from attempting to use sex typing, a personality disposition, in contexts where gender ideology, a belief system, would be a more appropriate predictor. While measures of gender ideology already exist (Antil et al., 1981; Kalin & Tilby, 1978; Spence, Helmreich, & Stapp, 1973), they measure broad social and political ideology. Our objective was to develop a measure of gender ideology which was consumption-specific, one which could be used to estimate consumers' tendency to construct or deconstruct gender distinctions through their consumption choices. Such a measure must be developed if we are to assess the effects of gender-deconstructing attitudes and behaviors and the effectiveness of treatments designed to foster these attitudes and behaviors in the marketplace.

Our scale was developed using procedures outlined by Churchill, 1979) and DeVellis (1991). Following the specification of the domain, two professors and two graduate students expert in consumer behavior and/or women's studies generated 52 items which, in their judgement, tapped the tendency to construct or deconstruct gender through consumption choices. These experts were instructed to suggest consumption-related items that would produce a differential response in subjects disposed to support traditional, asymmetrical gender roles versus those disposed to question and resist traditional roles. The 52-item scale generated in this way was purified through four successive stages, each of which involved data collection, factor analysis, and the calculation of Cronbach's alpha. During the first two rounds of this purification process, twelve additional items were generated, two of which were eventually incorporated in the final scale.

To explore interrelationships among items in the initial set and to suggest new items, data were gathered and analyzed in three preliminary rounds in which 30, 27, and 65 subjects were used. In each of these three rounds, responses were factor analyzed using principal components analysis in SAS with the MINEIGEN > 1 and then successively more restrictive NFACTOR parameters. Item-to-total correlations and Cronbach's alpha were also calculated. The objective was to enhance construct cohesiveness and validity by identifying items that loaded on major factors, that held together as the number of factors was restricted, that had high item-to-total correlations, and that improved reliability (Nunnally, 1978). Items that performed poorly on these criteria were retained in successive rounds but were moved to the end of the scale so that they would not influence responses to items that performed well. There were 52 items and 10 factors with an eigenvalue greater than one in the first round, 58 items and 12 factors in the second round, and 64 items and 15 factors in the third round.

In the final round of the scale development and purification process, 67 subjects responded to the survey. The results for particular items were generally consistent with those in the preliminary rounds. Consequently, 11 items at the beginning of this scale were selected for the final scale, items that had been moved to the beginning because they had loaded on major factors and had good item-to-total correlations. Factor analysis on these 11 items yielded a two factor solution that explained 55 percent of the response variance. Cronbach's alpha was .87, and item-to-total correlations ranged from .48 to .68. When added to these 11, other items in the survey created new, relatively unstable factors and/or reduced scale reliability.

The scale was validated by examining with new samples its factor structure, its reliability, and its convergent/ discriminant and predictive validities. The subjects were students enrolled in business classes. Two sets of subjects were used to avoid an overlong survey and consequent fatigue. The first had 132 subjects (48 female, 84 male), the second 153 subjects (55 female, 98 male).

FACTOR STRUCTURE

For both validation samples as for the development sample, factor analysis yielded a two-factor solution. For the first sample, the solution explained 49 percent of the response variance, for the second 51 percent. All factor loadings above .30 are reported for both samples in Table 2. The factor structure in the first validation sample is most clearly interpretable. Factor 2 taps attitudes toward gender typing in occupational roles. Items having to do with women pilots and automobile mechanics and male housekeepers load most heavily on this factor and do not load of factor 1. The item on boys playing with dolls may also cast males in the homemaker role while that on women wearing suit coats may cast females in a businessman role. Factor 1 seems to group responses to products and services. Consistent with this interpretation, the doll and suit items load on this factor while the more unambiguously occupational items do not.

The factor structure for the second validation sample is generally consistent with that in the first. All items that loaded on factor 2 in the first sample also load on it in the second. The same is true for items that loaded on factor 1, except for item 5. While most loadings are the same, the magnitudes have changed, so the second factor structure is less unambiguously interpretable than the first. Even so, there appear to be two relatively stable factors that explain a substantial proportion of the variance in attitudes toward products or services that construct or deconstruct traditional gender codes.

RELIABILITY AND VALIDITY

Scale reliability was acceptable in all validation samples. For the first sample (n = 132), Cronbach's alpha was .83. Item-to-total correlations ranged from .37 to .64. Two weeks later, the scale was again administered to 108 of these subjects (43 female, 65 male) to assess test/retest reliability. The correlation between the first and second scores was .85. Individual item-to-item correlations ranged from .30 to .71 with most being in the .60 range. For the second sample (n = 153), Cronbach's alpha was .86. Item-to-total correlations ranged from .45 to .69.

The convergent/discriminant validity of the scale was assessed using one scale which should and two scales which should not correlate highly with genderconstructing/deconstructing consumption. The scale which should correlate is Kalin and Tilby's (1978) Sex-Role Ideology Scale (SRIS). Since our scale is designed to be a consumption-specific sex-role ideology scale, it should correlate highly with more generalized sex-role ideology scales. This expectation was confirmed by the SRIS's .61 correlation with the first (n = 132) and .56 correlation with the second (n = 153) validation scale.

Our Gender-Constructing/Deconstructing Consumption (GC/DC) scale differs from other sex-role ideology scales in its focus on consumption. On this point of difference, it resembles materialism scales which also tap a tendency to consume. To confirm that the GC/DC did not merely tap differences in materialism, we measured level of materialism using Richins and Dawson's (1992) materialism scale. Consistent with our expectation that the GC/DC has a consumption dimension but does not tap general materialism, the correlation between the scales was .18 (n = 132) and .23 (n = 153).

The GC/DC is designed to measure a tendency to support or challenge traditional gender codes. To confirm that it does not merely tap a proclivity to monitor and comply with or violate the expectations of others (i.e. be feminist among feminists, traditional among traditionalists, or visa versa), we measured level of attention to social comparison information using a scale developed by Lennox and Wolfe (1984). The .11 (n = 132) and .17 (n = 153) correlations between the two scales suggest that the GC/DC does not measure a general tendency to construct/ deconstruct social conventions but is, rather, sex-ideology specific.

The high correlation between the GC/DC and Kalin and Tilby's (1978) SRIS establishes convergent validity but also raises a question: Do we need a new scale given this high correlation with the existing one? Our answer is that, in addition to being shorter (11 items rather than 30) and more reliable (.83 to .89 Cronbach's alpha rather than the .57 to .84 reported for the SRIS--.74 and .72 for our samples), the GC/DC seems to have better predictive validity in consumer behavior contexts. We tested the predictive validity of the competing scales by regressing the composite scores on various gender-related consumption attitudes. Judging from the R2 and the p-value of the regressions, the GC/DC outperforms the SRIS as a predictor of these attitudes.

The 46 items used as dependent variables included some suggested by our experts as measures of gender deconstruction that were not included in the final scale and others generated specifically for this test of predictive validity. Items measured attitudes toward progressive or nonprogressive business practices (e.g., I am more likely to purchase a company's products if I have heard it actively seeks to promote women employees), toward the sex-typing of occupations (I like to see women move into occupations traditionally held by men and visa versa), toward the sex-typing of consumption roles (A wife should be primarily responsible for purchasing the family groceries) toward product classes traditionally marketed to one sex or the other (I enjoy reading or watching science fiction), toward products that mark gender (It is a good idea to dress baby boys in blue, baby girls in pink), and toward specific brands that have been criticized for reinforcing asymmetrical gender distinctions (Barbie, Keystone Beer, Hooters Restaurants: My attitude toward Hooters Restaurants is positive). All items were measured on a 5-point scale anchored by strongly agree and strongly disagree. Half were administered to the first, half to the second validation sample.

While neither scale significantly predicted attitudes toward science fiction, handguns, sport utility vehicles, stylish clothing, traditional weddings, or Stroh's and Keystone Beer, the GC/DC did significantly predict attitudes on 35, the SRIS on 31 of the 46 dependent variables. Focusing on the 37 variables for which one or both scales were a significant predictor, the GC/DC had an average R2 of .16, the SRIS of .10. The GC/DC, thus, appears to have considerable utility in predicting a broad range of consumption attitudes.

IMPLICATIONS FOR SOCIAL ACTION

The strong relationship between the underlying attitude tapped by the GC/DC and the broad range of consumption attitudes and behaviors measured in the tests of predictive validity have noteworthy implications for businesspeople and consumer policy makers. It suggests that business practices or policy initiatives which change this underlying attitude will have important and generally progressive marketplace consequences, i.e., more employment opportunities, better products, and better service for women. And changes in attitudes toward asymmetrical gender codes are likely to be progressive because most people have a powerful economic interest in adopting the more liberal attitudes reflected in higher GC/DC scores: deconstructing gender distinctions makes the economy more efficient and leads to an increase in aggregate economic well being.

In his pioneering work on the economics of discrimination, Gary Becker (1971) has demonstrated that discrimination based on race and sex is a form of domestic protectionism that, like protectionism in international markets, reduces aggregate economic output. When employers discriminate in hiring, they must pay a wage premium to hire from an arbitrarily restricted labor pool. In addition to hurting themselves, they waste the skills of the groups they discriminate against. Likewise, when consumers discriminate in their buying, they pay a price premium for buying from an arbitrarily restricted pool of suppliers. They reduce their own standard of living and the standard of living of the groups they discriminate against. It is true that the group favored by those who discriminate (whites, males) may benefit, but their economic gain will be only a small proportion of the aggregate economy's loss. So even the favored group may be worse off in absolute terms than they would have been in a more robust economy that was unencumbered by the costs of racism or sexism.

While all may suffer, Becker makes it clear that the group discriminated against suffers the greatest loss. And the masking power of language and ideology notwithstanding, it is difficult to hide disadvantages from a disadvantaged class. Most victims of discrimination experience dissonance which alerts them to hidden asymmetries (Frable, 1989; Showalter, 1977). So disadvantaged groups--blacks, women--are likely to see through the cultural codes that legitimize and naturalize economic discrimination against them, though there are usually some group members who embrace language and ideas that limit their opportunities.

This tendency for a disadvantaged group to deconstruct injurious distinctions was apparent in our samples. As expected, women proved to be much more inclined than men to deconstruct the gender codes that disadvantage them. In both of our validation samples, women's GC/DC scores were much higher than those of men (Table 3). The attitudes reflected in these scores provides a firm foundation for social action in business and politics, for women's numbers (more than half of the population) and control of hundreds of billions in spending give them enormous potential economic and political power. The next two sections discuss options available to businesspeople, consumer advocates, and policy makers who wish to foster the deconstruction of gender codes.

The discrimination against women that Whittelsey (1993) shows is rife among American businesses creates a major opportunity for progressive businesspeople who want to combine doing good with doing well, for businesses can profit by deconstructing asymmetrical gender distinctions. A dry cleaner who charges women 27 percent more than men to launder a shirt is either earning an exceptional profit on cleaning women's shirts or subsidizing the below-cost price for laundering men's shirts with an above-cost price for women's. Assuming that the dry cleaning market is price-sensitive, a progressive dry cleaner could corner the market on women's shirts by charging an equitable price. And the same is true in other markets where market failure takes the form of price discrimination against women (Becker 1971).

This equitable pricing strategy should work with women consumers who are price sensitive, regardless of gender attitudes. But in communities where GC/DC scores are already high or where research shows that scores rise when consciousness is raised, a business could earn good will and additional patronage by building a promotional campaign around the deconstruction of an asymmetrical gender distinction. Since the network of assumptions that undergird a distinction are most fully exposed when the implicit valorization is reversed, such a campaign might be most effective if it reversed the status of women and men. A dry cleaner might, for instance, abandon equitable pricing for one week and, with some fanfare, invert their competitors' practice by charging 27 percent more to launder a man's shirt. The same thing could be done in any market where a gender distinction harms women.

Price discrimination is an economically inefficient and relatively obvious instance of asymmetrical gender coding. As such, it may be comparatively easy to eliminate through direct competition. Less obviously asymmetric gender distinctions may be harder to identify and eliminate but may provide even greater long-term potential for a business to capitalize on progressive policies. It is easier for competitors to change to equitable pricing than it is for them to match good will won by being at the forefront of an important cultural change. Levi Strauss is an example of a company that has built a powerful brand name in part by being sensitive to cultural change (Economist, 1991). In a recent jeans commercial, they continue that tradition by positioning themselves at the forefront of efforts to deconstruct traditional gender codes. In the commercial, the figure in a dress that symbolizes a woman's restroom comes to life, takes off her skirt, and turns it into a rope. Climbing down the rope, she escapes the frozen immobility of the restroom door. As the commercial closes, she pulls on the rope and the circle that framed her on the door falls to the floor and splinters. She stand free and uncircumscribed, clad in Levi's jeans, traditional clothing of frontier males. In this commercial, Levi Strauss aligns itself and its product with women who consciously or unconsciously feel constrained by the gender codes that surround them. The company promotes the use of its product as an act of personal assertion through gender-deconstructing consumption. Similar repositioning through gender code deconstruction is possible for many businesses and products.

While market forces and business activism can foster changes in gender perceptions, the upper echelons of corporate America are still massively dominated by males, who, judging from GC/DC scores in our samples, are relatively unconcerned about deconstructing asymmetrical gender distinctions. This result is unsurprising, for compared with women, men tend to be less concerned about socially responsible consumption in a variety of domains (Roberts, 1993). But the backwardness and resistance of men may not be that important. Women control hundreds of billions in their own income and even more in family income since they still do most household shopping (Roper, 1990; Schwartz & Miller, 1991). So if women consumers can be motivated to deconstruct gender codes through their consumption choices, businesses will be compelled to be responsive. They cannot afford to ignore the power of the purse.

Through education and lobbying, consumer organizations can play an important role in motivating women to deploy their economic and political power against harmful gender distinctions. For instance, Ralph Nader's Center for Responsive Law has published Frances Cerra Whittelsey's (1993) Why Women Pay More: How to Avoid Marketplace Perils, an example of effective gender code deconstruction by a consumer advocate. In this book, Whittelsey argues in the text (and Nader in the introduction) that women need to become more reflective and demanding consumers. Whittelsey educates women on business practices and products that do not serve their interests and explains how they can avoid and punish exploitation. The book has the merit of providing concrete examples of invidious gender distinctions that harm women and of effective consumer responses that deconstruct those distinctions. But it also has one potential weakness. In emphasizing the responsibility to be an informed consumer, Whittelsey may add one more burden to the already overburdened women who now spend an inequitable amount of time on child care and household shopping.

While educating consumers to be more conscientious is important, it may be still more important to reduce the search costs of the many women who, judging from GC/DC scores in our samples, are already disposed to be gender-deconstructing consumers. Efforts to foster green marketing and green consumption show the variety of ways in which this might be done. Taking a positive tack as some environmentalist organizations have done (Ottman, 1994), feminist and consumer organizations could present awards to businesses that have excelled in hiring and promoting women and/or in developing and marketing products that deconstruct traditional gender distinctions. Recipients could be encouraged to use their awards in product promotions. A more ambitious but more effective strategy would be to develop an equivalent of the green seal that has been used to mark environmentally friendly products (List, 1993). This would require the establishment of an institution to administer the seal and of guidelines on how to earn it. It would have the advantage of identifying progressive companies and products in the marketplace and, thus, allow women to recognize and reward gender equity while incurring very low search costs. Taking the more negative tack that Greenpeace has adopted in the environmental movement, groups like the Barbie Liberation Organization can continue to subject companies that victimize women to negative publicity. Since large businesses generally dislike controversy, tactics of this sort could motivate some positive changes.

Finally, consumer advocates and individual consumers should not neglect political action. As Gregory (1987) has pointed out, the appropriate model here is the political and legal fight against racial discrimination. Since women constitute political majorities in most developed countries, they have been and will continue to be big winners when governments intervene to limit discrimination through laws such as The Equal Credit Opportunity Act of 1974 and to remedy past discrimination through regulations requiring affirmative action (Gleckman, Smart, & Dwyer, 1991). Even if laws do not explicitly forbid practices such as the kinds of gender-based price discrimination that Whittelsey (1993) has identified, government action can raise the perceived price of a legal challenge and, thus, make businesses more responsive to consumer pressures.

LIMITATIONS AND DIRECTIONS FOR FUTURE RESEARCH

The principal limitation of this study is its use of a demographically restricted sample in the development and validation of the GC/DC scale--college business students. While young professionals are very important to most businesses and to policy makers, it is possible and even likely that these consumers may differ from other consumers, especially the elderly. So the most obvious extension of this study would be to validate the GC/DC scale with other demographic groups. Once validated with a relevant group, activists in business, government, and consumer organizations could use the instrument to evaluate efforts to encourage gender-deconstructing behavior. The level of GC/DC scores among various populations and the degree to which scores can be raised by initiatives such as those discussed above may indicate the degree to which political and consumer pressures can be mobilized to deconstruct gender distinctions that harm the interests of women.

REFERENCES

Anderson, S. & Bem, S. (1981). Sex typing and androgyny in dyadic interaction: Individual differences in responsiveness to physical attractiveness. Journal of Personality and Social Psychology, 41, 74-86.

Antil, J. K., Cunningham, J. D., Russell, G. & Thompson, N. L. (1981). An Australian sex-role scale. Australian Journal of Psychology. 33, 169-183.

Associated Press. (1993). 'Terrorist' act has G.I. Joe shopping. Roanoke Times & World News. 30 December.

Becker, G. S. (1971). The economics of discrimination. Chicago: University of Chicago Press.

Bem, S. (1985). Androgyny and gender schema theory: A conceptual and empirical integration. Nebraska symposium on motivation: Psychology and gender. Lincoln: University of Nebraska Press, 179-226.

Churchill, G. (1979). A paradigm for developing better measures of marketing constructs. Journal of Marketing Research. 16, 64-73.

Courtney, A. E. & Lockeretz, S. W. (1971). A woman's place: An analysis of the roles portrayed by women in magazine advertisements. Journal of Marketing Research. 8, 92-95.

Culler, J. (1982). On deconstruction. Ithaca, New York: Cornell University Press.

Derrida, J. (1973). Speech and phenomena: And other essays on Husserl's theory of signs. Evanston: Northwestern University Press.

Davis, H. & Rigaux, B. (1974). Perception of marital roles in decision processes. Journal of Consumer Research. 1, 51-61.

DeVellis, R. F. (1991). Scale development: Theory and applications. Newbury Park, CA: Sage.

Economist. (1991). A comfortable fit. The Economist. 319 (June 22nd), 67-68.

Fisher, E. & Arnold, S. J. (1990). More than a labor of love: Gender roles and Christmas gift shopping. Journal of Consumer Research. 17, 333-345.

Fontenelle, S. M. & Zinkhan, G. M. (1992). Gender Differences in the Perception of Leisure: A Conceptual Model. Advances in Consumer Research. 20, 534-540.

Frable, D. E. S. (1989). Sex typing and gender ideology: Two facets of the individual's gender psychology that go together. Journal of Personality and Social Psychology. 56, 95-108.

Gentry, J. & Doering, M. (1977). Masculinity-femininity related to consumption choice. Advances in Consumer Research. 5, 423-437.

Gleckman, H., Smart, T. & Dwyer, P. (1991). Race in the workplace. BusinessWeek. July 8, 50-63.

Gregory, J. (1987). Sex, race and law. London: Sage Publications.

Kalin, R. & Tilby, P. J. (1978). Development and validation of a sex-role ideology scale. Psychological Reports. 42, 731-738.

Larsen, Val. (1993). A sociolinguistic approach to gender and personal selling. Advances in Consumer Research. 20, 48-51.

Lennox, R. D. & Wolfe, R. N. (1984). Revision of the self-monitoring scale. Journal of Personality and Social Psychology. 46, 1349-1364.

List, S. K. (1993). The green seal of eco-approval. American Demographics. 15 (January), 910.

Meyers-Levy, Joan (1988). The influence of sex roles on judgment. Journal of Consumer Research. 14, 522-530.

Nunnally, J. C. (1978). Psychometric Theory. New York: McGraw-Hill.

Ottman, J. (1994). Nominations sought for 'green' awards. Marketing News. 28 (19), 19.

Richins, M. L. & Dawson, S. (1992). Materialism as a consumer value: Measure development and validation. Journal of Consumer Research. 19, 303-316.

Roberts, J. (1993). Sex differences in socially responsible consumers' behavior. Psychological Reports. 73, 139-148.

Roper Poll (1990). The environment: Public attitudes and individual behavior. New York: The Roper Organization.

de Saussure, F. (1959). Course in general linguistics. New York: McGraw-Hill.

Schwartz, J & Miller, T. (1991). The earth's best friends. American Demographics. 13 (February), 26-35.

Showalter, E. (1977). A literature of their own. Princeton, NJ: Princeton University Press.

Smith, D. (1987). The everyday world as problematic. Boston: Northeastern University Press.

Spence, J. T., Helmreich, R. & Stapp, J. (1973). A short version of the Attitude Toward Women Scale (AWS). Bulletin of the Psychonomic Society. 23, 210-220.

Wallendorf, M. & Arnould, E. J. (1991). 'We gather together': Consumption rituals of Thanksgiving Day. Journal of Consumer Research. 18, 13-31.

Whittelsey, F. C. (1993). Why women pay more: How to avoid marketplace perils.

Val Larsen, Truman State University Newell D. Wright, James Madison University

Table 1

1. If a couple is buying a car, the man should take the lead.

2. I would not want my boy to play with dolls.

3. I would be reluctant to fly on a plane that had a woman pilot.

4. A motorcycle is a male, not a female product.

5. The use of power tools like chain saws and electric drills is
more appropriate for men than for women.

6. I believe women's hair should generally be cut and styled
differently from men's hair.

7. * Were one available, I might use the services of a female
automobile mechanic.

8. * A magazine like Good Housekeeping is proper reading for a man.

9. Women should not wear suit coats cut like those customarily
worn by men.

10. On the whole, it is better that sewing be left to women.

11. Plastic surgery is more appropriate for women than for men.

* Starred items are reverse coded.

Table 2

 Sample 1 (n = 132) Sample 2 (n = 153)

Item Factor 1 Factor 2 Item Factor 1 Factor 2

5 .76 1 .79
10 .73 4 .69 .36
4 .68 .33 10 .64 .41
11 .67 3 .60 .50
6 .62 6 .59
1 .56 7 .52 .32
3 .78 11 .51
7 .77 2 .49 .38
8 .64 5 .83
2 .33 .42 8 .67
9 .34 .39 9 .65

Principal components analysis with a varimax rotation

Table 3

Sample Sex Mean SD p-value of t-test

n = 132 female 30.79 7.89 0.0001
 male 22.84 6.41

n = 153 female 28.96 7.92 0.0001
 male 22.06 7.47