Effect of Sex of Owner and Personal Circumstances on Attitudes Toward a Service Establishment

ABSTRACT - Women-owned businesses are an important social phenomenon. This experimental study finds that occupational sex-stereotyping does exist in the marketplace. The finding holds true whether subjects are male or female, sex-role traditional or contemporary, or psychologically masculine or feminine. Implications for further research are discussed with special emphasis on theoretical approaches which may provide greater insight into this and related areas.

Citation:

John Martin and Mary Lou Roberts (1983) ,"Effect of Sex of Owner and Personal Circumstances on Attitudes Toward a Service Establishment", in NA - Advances in Consumer Research Volume 10, eds. Richard P. Bagozzi and Alice M. Tybout, Ann Abor, MI : Association for Consumer Research, Pages: 339-344.

Advances in Consumer Research Volume 10, 1983      Pages 339-344

EFFECT OF SEX OF OWNER AND PERSONAL CIRCUMSTANCES ON ATTITUDES TOWARD A SERVICE ESTABLISHMENT

John Martin, Boston University

Mary Lou Roberts, Boston University

ABSTRACT -

Women-owned businesses are an important social phenomenon. This experimental study finds that occupational sex-stereotyping does exist in the marketplace. The finding holds true whether subjects are male or female, sex-role traditional or contemporary, or psychologically masculine or feminine. Implications for further research are discussed with special emphasis on theoretical approaches which may provide greater insight into this and related areas.

INTRODUCTION

Much has been written in recent years about the influx of women into occupations traditionally stereotyped as masculine. Empirical research in this area has focused almost exclusively upon women in professional and managerial occupations. The research on female managers, however, has not dealt with an important sub-set of managers -- those who own their own businesses. Since there is evidence that women respond to blocked career pathways (or to their perception that career paths are or will be blocked) by establishing their own businesses, the trend toward increasing female business ownership should continue and perhaps even accelerate.

The basic objective of this study is to ascertain whether or not differential expectations concerning male vs. female performance, a phenomenon which is well-documented in the behavioral sciences literature, applies in a marketing context. If there is evidence of differential expectations or evaluations, further research is warranted. A secondary objective of this study, then, is to propose promising theoretical and empirical approaches to an issue which has so-far been largely ignored in the marketing literature.

This is an important issue because it affects the strategies by which female managers and entrepreneurs relate themselves to their environments. Specifically, it affects their approaches to dealing with the providers of business resources such as financial institutions, to the selection of target markets, and to the strategies for establishing effective exchange relationships with those markets.

LITERATURE REVIEW

We could locate only two relevant studies in the marketing literature, both of which deal with perceptions of female salespeople. Robinson and Hackett (1977) found that male salespeople held stereotypical attitudes toward female salespeople. In addition, the attitudes of sales managers (almost all of whom were male) were even more strongly stereotypical than those of the male salespersons. Banks (1979) conducted an experiment which found that failure of both male and female salespersons was attributed to internal causes (e.g., lack of ability). These college student subjects behaved in a manner counter to much of the existing behavioral science literature by not attributing the failure of male salespersons to external factors (e.g., bad luck).

Most of the relevant literature in the area of occupational sex-stereotyping does come from the behavioral sciences. Goldberg conducted the classic experiment in 1968. He concluded that Women are prejudiced against female professionals and, regardless of the actual accomplishments of these professionals, will firmly refuse to recognize them as the equals of their male colleagues." (Goldberg, 1968, p. 30). A replication found women who were presented with artistic work of both males and females who had no known professional credentials devaluing the women's work relative to the men' s. The work of women who had successfully accomplished was rated equally with men with proven accomplishment (Pheterson, Kiesler-and Goldberg 1971). Mischel (1974) found that journal articles were rated most favorably by judges of both sexes when the sex of the author was in accordance with the normative expectation for that profession (e.g., primary education for women; law for men). Etaugh and Kasley (1981) found female job applicants being devalued by both males and females, but especially by males.

Taking a slightly different approach, Feather examined the anticipated consequences of success or failure for both males and females in a variety of occupations. He concluded that "reactions to male and female success and failure for an occupation depend upon the perceived appropriateness of the occupation for the sex concerned." (Feather, 1975, p. 546).

Two final studies throw light on circumstances that may affect sex-based evaluations. Hagen and Kahn (1975) found that the type of experimental interaction affected males' liking of a competent woman. Also, both males and females were less likely to include a competent woman than a competent man in their experimental group and were more likely to include an-incompetent man than an incompetent woman. Isaacs (1981), in another study which varied the purported authorship of journal articles, did not find women devaluing the work of other women but found males devaluing the work of women in a "masculine field unless high status was attributed to the female author. Isaacs hastens to point out that this research was conducted on the campus of the University of California at Berkeley, a bastion of contemporary lifestyles, among student subjects whose responses to the Spence and Helmreich Attitudes Toward Women Scale identified them as ultra-contemporary with regard to women's roles. Apparently considering this group a "leading indicator" of social change, she hypothesizes that stereotyping of occupational appropriateness and achievement may indeed be decreasing.

Taken as a whole, however, these studies present strong evidence that sex-stereotyping of occupational achievement still flourishes. What remains is to see if the stereotyping transfers into the domain of the female entrepreneur and if it affects intended patronage of the woman-owned business.

As strong as this evidence is, however, many behavioral scientists argue that psychological masculinity and femininity is a more important determinant of attitudes and behavior than is sex per se. We will focus on the Bem sex-role inventory (Bem. 1974) and the Attitudes Toward Women Scale (Spence and Helmreich, 1978), two potentially competing explanations for the expectations that are held regarding male/female occupational appropriateness and achievement. Both scales are widely accepted and extensively used by researchers in the behavioral sciences. In addition, the BSRI has been used in a limited number of studies which have related sex-role self-concept to intentions to purchase sex-typed products (see Gentry and Doering, 1977; Allison, Colden, Mullet and Coogen, 1978; and Golden, Allison and Clee, 1379).

The basic hypothesis to be drawn from the research generated by the BSRI is well-stated by Kelly and Worell. Their review finds that "androgynous orientations are associated with greater behavioral flexibility across situations than are sex-typed orientations ... that feminine-typed persons should behave in a more expressive, affective manner ...whereas masculine-typed persons should excel...in situations in which instrumental, assertive, and goal-directed qualities are appropriate." They add that persons who describe themselves as neither expressive nor instrumental (undifferentiated) should "exhibit behavioral deficits across situations" (Kelly and Worell 1977, p. 110). Recent research has continued to support their interpretation (Orlofsky and Windle 1978. Orlofsky and Stake 1981).

Taken together with the research on sex-biased evaluations, the androgyny literature would seem to imply that feminine-typed persons would have a preference for persons, as well as situations, who are expressive in nature while masculine-typed individuals will prefer persons who are instrumental (e.g., entrepreneurs). This preference may or may not be mediated by attitudes toward women's roles as the did in the Isaacs (1981) study.

HYPOTHESES

This study is concerned with expectations of male vs. female performance in a market context. Market-related measures of performance include expectations concerning behavior of male and female objects, such as the service and products delivered, and the outcome of this behavior such as market demand and image. Since the hypotheses are based on literature just reviewed, they are stated first.

H1A: Both males and females expect the same level of performance from proven men and proven women.

H1B: Both males and females expect a lower level of performance from unproven women than from unproven men.

These hypotheses are consistent with the literature. H1B reflects the general tendency to devalue the work of women relative to men. H1A reflects the special situation in which the female has achieved professional recognition thereby increasing the likelihood that she will be evaluated similarly to an equally successful male.

H2A: Both males and females who hold traditional sex-role attitudes expect a lower level of performance from women than from men under both proven and unproven conditions.

H2B: Both males and females who hold contemporary sex-role attitudes expect the same level of performance from women and from men under both proven and unproven conditions.

Sex-role attitudes as commonly measured deal with feelings about the degree of freedom which women should be allowed to exercise. Do these specific attitudes reflect a generalized favorability toward women's endeavors? The findings of the single relevant study which we could locate (Isaacs 1981) suggest that they do. H2A and B, then, posit a straightforward translation of sex-role attitudes onto evaluations of occupational achievement.

H3A: Psychologically feminine individuals expect higher levels of performance from unproven men and women than from proven men and women.

H3B: Psychologically masculine and undifferentiated individuals expect higher levels of performance from proven men and women than from unproven men and women.

H3C: Androgynous individuals expect similar levels of performance in all four experimental conditions.

Hypothesis 3A and the "psychologically masculine" portion of hypothesis 3B are applications of the expressive/instrumental relationships found in the literature. Undifferentiated individuals are also expected to attribute greater likelihood of success to proven individuals. This is a more "traditional" attribution, and we interpret the "behavioral deficit" concept from the androgyny literature to mean that these individuals will follow the path of least resistance in many aspects of their lives.

We interpret the "behavioral flexibility" that androgynous persons are supposed to exhibit as meaning they can accept and/or adapt to a wide variety of situations. The unproven treatment should therefore not particularly affect them. Neither should the woman-owned business have any effect. Therefore, we conclude that the experimental treatments, singly and interactively, are not likely to strongly affect these individuals.

THE EXPERIMENTAL STUDY DESIGN

In order to provide a valid test of the hypotheses a laboratory experiment was designed with the following characteristics.

Market Based Stimulus and Treatments

It was necessary to select a stimulus which would be "neutral," in a gender sense, but which could be varied in terms of gender and affective overtones. The stimulus chosen was a new seafood restaurant. The decision to "eat out" at a moderately priced restaurant is a reasonably important and visible decision for students who commonly frequent fast food outlets. This should increase the likelihood of strong expectations being formed about performance.

In order to provide sex and "success" based manipulations for the testing hypotheses, two treatments were formulated and imbedded into a flyer-type advertisement describing the restaurant. Each treatment had two conditions giving a two by two factorial design.

For the first treatment, the owner of the restaurant was represented as being either male or female, Rick or Ruth. All of the staff in the male condition was also described as being male (the same was true of the female condition). Gender-distinguishing words occurred six times in the flyer. This treatment provides a direct sex manipulation of the stimulus.

The second treatment involved creating dichotomous conditions which would change the affective value of the stimulus. It was necessary to portray an entrepreneur entering a new business in which he/she had no experience so as not to directly precondition performance expectations. This was accomplished by using professional golfer as the previous occupation. This was appropriate in a sports-oriented community where a number of well-known sports figures are proprietors of local establishments. The polar conditions were labeled "proven" and "unproven", and were created by changing the physical, financial and social circumstances of the owner and staff. In the proven condition, the owner had achieved widespread recognition and financial success as a professional golfer. In the unproven condition, the owner was also a professional golfer who had experienced physical and financial set-backs, through no fault of his/her own, and had no widespread recognition. In both cases, the employees were parallel to the owner in the described conditions. Otherwise, the description of the stimuli were identical and portrayed a moderately priced seafood restaurant with pleasant decor and friendly, efficient service.

Sample

Subjects were male and female graduate and undergraduate business students from a large urban university. Business student subjects were deemed appropriate because they represented a relatively homogenous group. Calder, Phillips and Tybout (1981) argue strongly that a homogenous subject group provides the most rigorous test when the objective of the study is theory falsification as it is in this instance. Approximately 220 students participated in the study; however, because of missing segments and data the effective sample size is 125 for the experimental manipulation.- It was possible to use larger samples for construct validation of the various measures.

Experimental Procedure

Since several lengthy scales were administered, the study was conducted in three segments. The first two segments, administered approximately one week apart, covered individual background factors (socialization influences) and included widely-used measures of sex-role attitudes (the Attitudes Toward Women Scale) and sex-role self-concept (BSRI). The third segment, administered approximately one week later, presented the experimental treatment and obtained attitudinal and patronage intentions data for the restaurant. In addition, respondents were asked to record their cognitions regarding the stimulus (unaided cognitive responses).

CONSTRUCT OPERATIONALIZATION

In addition to the two treatments, a number of constructs had to be selected or formulated to measure dependent and other independent variables.

Expectations concerning performance were the dependent variables for the test. In order to measure these, a number of perceptual domains were identified as relevant to restaurant patronage based on previous studies of store image assessment (see Lindquist 1974-75 for an extensive review.) These domains or constructs were "marketplace success of the restaurant," motivation of the owner and staff," "restaurant atmosphere," "acceptability of the food," and "personal service." Constructs with both objective and subjective overtones were included.

The dependent variables were defined, as shown in Table l, and statements suitable for use in a seven point Likert-type scale were assembled to tap each one. Four judges (two marketing professors and two graduate marketing students) were used to select the most appropriate items (content validity), and a final set of 29 statements was used in the study. The sets of items for each construct were factor analyzed for the total sample as well as for the subgroups of male and female respondents (tests for convergent validity and instrumentation threat). All items were also factor analyzed together to examine the interdependence of constructs (a test for discriminability). Reliability for each scale was measured using Cronbach's alpha. As a consequence of this analysis, the "personal service" construct was eliminated because of poor convergent validity, and "restaurant atmosphere" because of a low alpha. This left three measures with Cronbach alphas in the 0.80s: "marketplace success" and food acceptability, which are objective in nature, and "staff motivation' which has distinct affective overtones. The items for each construct were summated and averaged to maintain a seven point measure.

The Bem Sex Role Inventory (Bem, 1974) was used as an independent measure to tap respondents' sex role self-concepts. A number of studies have been conducted which indicated that Bem's masculine and feminine scales are multi-dimensional not undimensional as originally conceptualized by Bem (Whetton and Swindells, 1977; Pedhazur and Tetenbaum, 1979). The scale was factor analyzed to test Bem's proposed structure of feminine, masculine and neutral dimensions. The results produced a set of factors which are consistent with the studies just cited. For each of the masculine and feminine dimensions, a set of items was extracted which had a single factor (eigen value> 1) and Cronbach's alphas in the range of 0.86 to 0.91 for the total sample and male and female subsamples.

As a further test for independence of the masculinity and femininity dimensions, Pearson correlations were calculated between them for the total sample and each sex subsample. In all instances, the correlations were close to zero and non-significant.

TABLE 1

TESTS FOR CONVERGENT VALIDITY, INSTRUMENTATION THREAT AND RELIABILITY: DEPENDENT VARIABLES

Spence and Helmreich's (1978) Attitudes Toward Women Scale was used to tap sex-role attitudes, another independent measure. The scale was factor analyzed to explore the underlying structure, and found to have four sub-dimensions: male dominance, job equality, female rights and freedom, and courtship equality.

The sets of items representing each of the four attitudinal constructs were factor analyzed ant, where necessary, items were deleted so that single factor situations were obtained for each of the male, female, and the total sample groupings. Cronbach's alpha ranged from 0.60 to 0.85 for the four sub-scales. Scores on the four constructs were summated and averaged to give each equal weight in a single composite score for sex-role attitude.

RESULTS

Manova with covariates was used for the analysis. The three dependent variables were the performance expectations shown in Table l. The independent variables were "sex of the owner," sex of the respondent," the proven/unproven" conditions, "sex role attitude," and "sex-role self-concept" of the respondent. This last set of variables was entered as a main effect or as a covariate depending on the nature of the hypothesis.

In the majority of tests, the affective dependent variable of "staff motivation" was the most responsive. In order to simplify reporting, results are mainly given for staff motivation. Likewise, since no covariates were significant in all tests, only the use of covariates are reported, not the numeric results. Means reported for dependent variables are based on a seven point scale, ranging from 1 ta 7 with 7 the most desirable and l the least.

Manipulation Check

A manipulation check was conducted to see if the two inductions were effective. This was done by having respondents record their cognitive responses after reading the stimulus. These unaided responses are a lower bounded measure of inductive effectiveness. The results showed 66% of respondents correctly recording cognitions directly related to the unproven/proven treatment and 45% correctly making direct references to the male/female manipulation. Only two incorrect responses were recorded. A total of 35% of respondents correctly made direct references to both manipulations. This provides a strong indication that the experimental manipulations were indeed noted by the respondents.

TABLE 2

RESULTS OF MANOVA TESTS

The two manipulations were tested separately using Manova. In the case of the success manipulation there was a significant main effect with and without covariates, with staff motivation the most influenced dependent variable (Sig < .01) and food acceptance also showing some reaction. Means for staff motivation were 5.7 (n = 64) for the proven treatment and 5.1 (n = 61) for the unproven. This clearly shows a significant main effect of the success treatment. The main effect of the sex of owner treatment was not significant for any of the three dependent variables. Again, the largest effect was seen in the means for the-staff motivation construct -- x = 5.5 (n = 61) for the male owner and x = 5.3 (n = 64) for the female owner.

Test of Hypotheses

In order to test the hypotheses, two Manova's, one without covariates and one with, were conducted for each. Covariates were used to control for possible effects of other experimental treatments or conditions. Since in none of the analyses were the covariates significant, their effects are not reported. The results of the analysis are summarized in Table 2 which presents sample sizes, levels of significance, and the covariates used in each analysis

The Proven-Unproven Manipulation. Hypothesis H1A states that both males and females expect the sane level of performance from proven men and proven women. There were 64 respondents in the proven treatment. The results support H1A since there is significant main effect for sex of subject nor was there an interaction effect between the sex of owner treatment and sex of subject. The results for the staff motivation construct are illustrative:

TABLE

While these data show some evidence of an interaction effect. the effect sizes are small.

H1B states that both males and females expect a lower level of performance from unproven women than from unproven men. There were 61 subjects exposed to the unproven treatment. Again, there are no main or interactive effects meaning that the hypothesis should be rejected. However, the results for the staff motivation construct show a tendency for both males and females to have lower expectations of the unproven women.

TABLE

Effect of sex-role attitudes. H2A and H2B deal with the effect of traditional vs. contemporary sex-role attitudes on performance expectations. Respondents were divided at the median, using the Spence and Helmreich scale, into a traditional and a contemporary group. For both groups, there is no main effect for sex of owner, but there is a significant effect for the proven/unproven treatment. The results are similar with or without covariates. In neither instance is there a significant interaction effect.

For our sample, in fact, both sex-role traditional and contemporary subjects hold higher expectations of proven individuals regardless of their sex. It is also important to note that using sex of respondent as a covariate does not change this interpretation. The means for the staff motivation construct show that in both the traditional and contemporary subsamples, the unproven female owner is devalued. Our data. then. does not support either H2A or H2B.

The Sex-Role Self-Concept. H3A, B and C deal with the effects of psychological masculinity/femininity on performance expectations.

Table 2 shows that there are significant main effects of the proven/unproven treatment for both the feminine and masculine subsamples but not for the undifferentiated and androgynous subsamples. There are no main effects of the sex of owner treatment for any of these four subsamples. The only significant interaction effect is between the sex of owner treatment and the sex of subject conditions for the masculine subsamPle.

The significant main effects for feminine individuals are as hypothesized. However, examination of the means for staff motivation shows i - 5.84 (n = 19) for the proven treatment and x - 4.97 (n a 12) for the unproven treatment. Clearly, feminine individuals did not inflate their expectations of unproven women but instead devalued them just as did masculine individuals.

The lack of significant effects for undifferentiated individuals are not as hypothesized. Means for staff motivation are very similar for both the unproven (x = 5.26," = 14) and proven ax - 5.47," - 18) treatments. Apparently undifferentiated individuals simply did not react to the experimental conditions.

The lack of significant effects for the androgynous individuals are as hypothesized. However, examination of the means for staff motivation shows some difference with the unproven treatment at x = 5.1 (n = 17) and proven treatment x = 5.7 (n = 16).

The masculine sub-sample, with a significant main effect for the proven/unproven treatment and an interaction effect between proven/unproven and sex of owner, provides an interesting insight. It means that not only does the masculine group have lower performance expectations for unproven people relative to proven people, but unproven females are particularly devalued. In other words, the effect of psychological masculinity reinforces the effect of the male owner, relative to the female owner, especially in the unproven condition.

DISCUSSION AND RESEARCH IMPLICATIONS

Even though many of our findings were consistent with the hypotheses we established, three aspects of the results were particularly unexpected.

The first was the strength of the "success manipulation. We expected it to be effective for only some of the subsamples. Instead, it was the only treatment or condition to be consistently effective throughout the analysis. While the fact that it did have an effect was predictable from the literature, we can offer no real explanations for its strength and consistency. More research is needed to determine causal factors generating this type of expectation and the situations in which it is most likely to occur.

The second unexpected result was that contemporary vs. traditional sex-role attitudes had no effect. This was particularly surprising in view of the fact that 87: of our sample scored above the midpoint of the Attitudes Toward Women Scale meaning that the vast majority held mildly to strongly contemporary sex-role attitudes. We expected that type of distribution for a group of northeastern area college students and anticipated that it would cause them to look more favorably on female entrepreneurs. On the other hand, it can be argued that our sample was so homogeneous that no effect of truly traditional sex-role attitudes could be observed. However, we would hasten to point out that even though our sample was skewed toward the contemporary end of the continuum, they held traditional expectations (i.e., they tended to devalue women overall and strongly devalued unproven women). Our conclusion, then, is that sex-role attitudes do not necessarily generalize onto other attitudes even when they seem to be intrinsically related.

The third aspect concerns the fact that only the staff motivation construct was consistently affected by the manipulation of independent variables. What is clear is that this is the only dependent variable that is expressive in nature; the remainder are instrumental. What is not clear is why the effect on the expressive variable did not "carry over" to the instrumental variables. In other words, if respondents believed that "unproven" persons would indeed try harder, why didn't they also believe that these people would be more successful? Future research would do well to maintain the expressive/instrumental dichotomy and to further probe the effects of these two types of variables on attitudes and expectations as well as intended and actual behavior.

The results obtained from the use of the BSRI were quite gratifying. This is not simply because they tended to support our hypotheses. Primarily it is because we believe that the concept has more to offer to the study of consumer behavior than do the more commonly used sex-role attitude scales, and our results offer support to this belief. We therefore encourage its use where appropriate. At the same time we warn against scoring it in the manner originally proposed by Bem (1974). Scoring it into four groups (i.e., the addition of the undifferentiated category) is consistent with currently accepted practice.

Another research implication from this study relates to what Sawyer (1981) has called an underpowered test. Sawyer argues strongly, and correctly we believe, that it is effect size, not statistical significance per se, which is important. There are two instances within this study in which a larger sample would have caused the existing effect size to be statistically significant. The first instance is in the devaluation of unproven females by female respondents (H1B). We did not predict this result, but it is reminiscent of Kanter's "iron maiden" female role in organizations. The second instance is the higher evaluation given by the androgynous sub-group to unproven individuals (H3C). This result was also not predicted but it could be interpreted as a manifestation of the greater "behavioral flexibility" of androgynous individuals. It would be interesting to further investigate both phenomena. Even more importantly, however, future research should exercise care to obtain an adequate sample size, especially where complex analysis can quickly lead to small cell sizes as it did in this instance.

The final research implication concerns the need for more theoretical approaches to the study of occupational sex-stereotyping as it applies In the marketing context. The literature from the behavioral sciences is intriguing but atheoretical. Two areas of theory seem to have great potential applicability.

The first is attribution theory. A limitation of the current study is that it did not determine the attributions made by respondents as they evaluated expected performances under the various experimental manipulations. Knowing the causal factors which respondents attached to their expectations would have been very helpful in interpreting the results.

The second theory area is information processing. When one considers the results of this study together with the literature cited earlier on male/ female-typed products, an intriguing question emerges. Is information processed differently when products have masculine or feminine images than it is when products are gender-neutral?

While it is inappropriate to develop managerial implications on the basis of one exploratory study, further research could produce information of great value. This could be in the area of "credentialling." For instance, should male and female managers and/or entrepreneurs present credentials in the same manner? Other implications may lie in the area of communications source credibility. Is the credibility of male vs. female spokespersons evaluated differently, especially when the produces have gender-linked or gender-neutral images?

CONCLUSION

With regret, we must conclude that sex-stereotyping seems to be alive and well, even among a seemingly contemporary group of subjects. The attributions generated by sex-stereotyping in other situations and/or by other groups of subjects represents a subject deserving of further study by marketers.

REFERENCES

Allison, Neil K., Linda L. Golden, Gary M. Mullet, and Donna Coogan (1979) "Sex-Typed Product Images The Effects of Sex, Sex Role Self-Concept, and Measurement Implications," Advances in Consumer Research Vol III ed by Olsen, Jerry C Ann Arbor 604-609.

Banks, Sharon (1979) "An Attributional Experiment on Sexual Biases Towards Salespeople," AHA Educators' Conference Proceedings ed by Beckwith Neil et al Chicago, 431-434.

Calder, Boby J., Lynn W. Phillips, and Alice M. Tybout (1981) "Designing Research for Application Journal of Consumer Research September 197-207.

Etaugh, Claire and Delen C. Kasley (1981) "Evaluating Competence Effects of Sex Marital Status, and Parental Status," Psychology of Women Quarterly Winter 196-203.

Feather, N. T. (1975) "Positive and Negative Reactions to Male and Female Success and Failure in Relation to the Perceived Status and Sex-Typed Appropriateness of Occupations," Journal of Personality and Social Psychology, 31, 536-548.

Gentry, James W., and Mildred Daring (1977) Masculinity-Femininity Related to Consumer Choice ' ASA Educators' Conference Proceedings ed by Greenberg Barnett A and Danny" Bellinger, Chicago 423-427.

Goldberg, Phillip A. (1968) "Are Women Prejudiced Against Women?" Transactions April, 28-30.

Golden, Linda L., Neil Allison, and Mona Clee (1978), "The Role of Sex Role Self-Concept in Masculine and Feminine Product Perceptions," Advances in Consumer Research, Vol VI, ed by Wilkie William, Ann Arbor, 599-604.

Hagen, Randi L., and Arnold Kahn (1975) "Discrimination Against Competent Woman." Journal of Applied Social PsYchology 5, 362-376.

Isaacs, Marla Beth (1981) "Sex Role Stereotyping and the Evaluation of the Performance of Women Changing Trends," Psychology of Women Quarterly 6 187-195.

Kelley, Jeffrey A. and Judith Worell (1977) "New Formulations of Sex Roles and Androgyny A Critical Review," Journal of Clinical and Consulting Psychology, 45, 1101-1115.

Lindquist, Jay D. (1974-75), "Meaning of Image," Journal of Retailing, 50 29-37.

Mischel, Harriet N. (1974), "Sex Bias In the Evaluation of Professional Achievements," Journal of Educational Psychology 66 157-166.

Orlofsky, Jacob L. and Jayne E. Stake (1981) "Psychological Masculinity and Femininity Relationship to Striving and Self-Concept in the Achievement and Interpersonal Domains," Psychology of Women Quarterly 6 218-233.

Orlofsky, Jacob L. and Michael T. Windle (1978) "Sex Role Orientation Behavioral Adaptability and Personal Adjustment," Sex Roles 4, 1978.

Pedhazur, Elazar J., and Toley J. Tetenbaum (1979) "Bem Sex Role Inventory A Theoretical and Methodological Critique," Journal of Personality and Social Psychology, 37 996-1016.

Pheterson, Gail I., Sara B Kiesler, and Philip A. Goldberg (1971), "Evaluation of the Performance of Women as a Function of Their Sex Achievement and Personal History," Journal of PersonalitY and Social Psychology 19, 114-118.

Robinson, Dan J., and Donald W. Hackett (1977), "Saleswomen Perceptions, Problems and Prospects," Journal of Marketing, 61, 66-71.

Sawyer, Alan G. (1981), "Statistical power and Effect Size in Consumer Research," Advances in Consumer Research Vol IX. ed by Mitchell Andrew, Ann Arbor, 1-7.

Spence, Janet T. and Robert L. Helmreich (1978), Masculinity and Femininity, Austin, University of Texas Press.

Whetton. C. and T. Swindells (1977), "A Factor Analysis of the Bem Sex Role Inventory," Journal of Clinical Psychology, 33, 150-153.

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