wwmmw I?! . . fl ._‘.:- - . ~ ~‘h ’- sV._ In. I... .:‘,I.:; ..,- “"h'u‘é; “‘33:. 1:1; . .141: ”.13 uh».: kJMC.v.I—~. v-I‘. ‘34—: .4 "”04 II- I»! . .:..I' I. . ‘ . ""“""'Z" . ‘4...“ n..- .1; 4‘. .;~' u . i" J . ' ii fl. “12.: {2.1.511 1 -1... .. .I , :mlf' -.'I'I'.II I.“ W'll" -'I,uu [Isl-.1)! Ir r": 4-»..- m, 1 I. ’1 ~ *w. W»; rwmhw 4-». w‘::fl"“?‘"*‘;", wfifixx +3 ‘32:-.. 21w “\ibh‘ #5 v“. 53.-.. .5.“" 3.-.”-..3.-. T‘u": ‘L. 43.».w ., .- .mu‘ .1; I4 :;,_4.:. ..§:._.:...:_.. ‘5: - ‘33 I. W.» h 3... . 7:32.331:- .13.»: . .‘:~‘-‘.. ~ e 21?. ,.‘ 3 “1...... . 5:7.-.TJ.» CHL- ‘ .I .....:- I‘m-J... --.-.-- -._.: ...I...-.....‘..:.I.-zuu , .53?" .351 “.2. it... ... a...“ :dn::tm L’Vk‘fl .... - -3 “‘ ft. ¥Wh‘ l _.. I‘. . v J ‘3 22.“... -... I..-:.3.-..c.~£nk “.5, 1-... ”RH .35.“.-. " I. 3 ‘ “.3“. maul..- . .- .. .. 7 .3. I. .m- 4*. .4 1-». 1 d.h‘:—-é:-..“;.4.A . Matti—pum- -3. . m5. .1 " , .31. . . '17. .3 A .. i‘ .1‘ 1“,; v-~ «IA-I1. w 3—.» -4“... . 3. «0“. ‘ML ...I.v ‘n-. :w':‘ “”315 W‘m .33.?qtrmm".1~ M :36 “k \ v_a...~:u..m~« "Huh. can. 1.4.... A :o. . . a --..' .. . C‘ -?.€- - . ‘.7;:.-£ .‘fiu‘fiufi ‘ ‘- :t-. .I. ”2. 4..-.1 ,.. mm. «A; .3J.xu". .. .“u. ,3" m:7'.-.:;i.4.h -..~:J nu -4 u “h 4. - ¢...~ . I: -3-” 3. .. I .- - x... .4... -_ . r»: 5‘}- ~ 3 . ~ ”Au.“— MJ u 3 ’- an. I: .yuv:.ma‘~. ~. -5: «:5»? J’nfi ‘ 4 IN N .- =4 . ‘- “I’ll-.5...“ ff. .- r ._.. I. -u 0.7:}..5‘ 41:... 3...:«I w 41?...— . “4...... ....,...\.‘Z. t ' 3-” “In“ ”I . _. “ ,..., 1“)..." . v'w’fu 'v- ‘ .. , d l ...,¢r,I,.np . In . “up”: .7. .,. f ., “.5. V' 5 ugh-u."- ,i..",'/-t' ,. , .‘U .4... . ~., 4" ...I - a ‘ffi‘ _ - a: .f, .1, ‘.."- . .v 3.4!? 1'. ,4! ‘ W: r H wt 9 8:1“ ”1:- l ‘ “"uun-WfiA A\ n. 3...-.. ,h" .'.'E ’3. .. ‘4-1 W‘ n .""' ’.....“. .' .m . .'.' Z 'i '5‘»? .I .33....NJ." N. .. m, "a... -1 _A J:lr-!‘v: - .r n-r-I "IESIS CHISGAN IIIIIIIII IZIIIIII IIIIIIIIIIII IIIIIIIIL 300792 6524 This is to certify that the dissertation entitled Sex-Role Differentiation Among African-Americans and Anglo-Americans: A Consumer Behavior Study presented by Allen C. Harris has been accepted towards fulfillment of the requirements for Ph.D degree in Mass Media Date g/ZO /98 MSU is an Affirmative Action/Equal Opportunity Institution 0-12771 WY Ilium mu University PLACE IN RETIJRN BOX to remove this checkout from your record. TO AVOID FINES return on or before data due. DATE'DUE DATE DUE DATE DUE it I I flf MSU Is An Affirmetivo Action/Equal Opportunlty IMItution emu“.- SEX-ROLE DIFFERENTIATION AMONG AFRICAN-AMERICANS AND ANGLO-AMERICANS: A CONSUMER BEHAVIOR STUDY By Allen C. Harris A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Mass Media PhD. Program 1992 4% I w !72- 36 ABSTRACT SEX-ROLE DIFFERENTIATION AMONG AFRICAN-AMERICANS AND ANGLO-AMERICANS: A CONSUMER BEHAVIOR STUDY By Allen C. Harris A theoretical framework was offered to (I) examine among African- Americans and Anglo-Americans the nature and prevalence of sex-role differen- tiation based upon inherent personality characterizations for men and women; (2) compare the attitudes of African-Americans and Anglo-Americans toward equality between the sexes; and (3) test whether flexible attitudes toward a variety of purchase decisions are related to non-traditional sex-role attitudes, and whether flexibility in purchase decisions is associated with a more flexible definition of oneself in terms of sex-role norms. As predicted, results of this study offer support for the assumption that the sex-role identities of African-American males and females are more androgynous than the sex-role identities of Anglo-American males and females. The most "androgynous" group was the African-American female group, with Anglo-American males displaying the lowest sex-role identity score (least androgynous group). Sex-role identity scores for the African-American males and the Anglo-American females were nearly equal. Anglo-American males were also found to be more traditional in their sex-role attitudes than were African-American males. However, the difference between the mean sex-role attitude scores of the Anglo-American female and the African-American female groups was not significant. Women of both races tended to be more flexible in their sex-role attitudes than were men of either race. African-American subjects were also found to be less likely than Anglo- American subjects to assume responsibility in purchase decisions based on traditional norms governing the sex appropriateness of the decision. It appears that among African-Americans, the influence to make a purchase decision may be determined less by the role position of the individual (husband/wife) and more by other, as yet unexplored, factors. Also, as predicted, both sex-role identity and sex-role attitude were significantly correlated with purchase decision responsibility among Anglo-American subjects. However, only sex-role attitude was significantly correlated with purchase decision responsibility among the African-American subjects. Sex-role attitudes were, however, somewhat better at predicting purchase decision responsibility for gender "inappropriate" rather than gender "appropriate" decisions among African-American subjects. Copyright by ALLEN C. HARRIS 1992 Dedicated to Mary Virginia Grandison 1947-1988 ACKNOWLEDGEMENTS There are a number of friends, colleagues, and advisors who have given generously of their time and knowledge to make this dissertation possible. For her encouragement and guidance throughout this project, I offer my sincerest thanks and gratitude to my advisor, Dr. Gina M. Garramone. FOr their editorial comments and critical anaylses, I wish to thank each of my committee members, Dr. Bruce VandenBergh, Dr. Bonnie che, and Dr. Tony Atwater. For initially exposing me to the field of communication and culture, and for his continued support throughout my professional career, I am especially grateful to Dr. Thomas Kochman. For his valuable help in the area of data collection techniques, I wish to thank Dr. James Danowski. I also owe a great debt to Kevin L. King for his care and proficiency in the layout and design of the manuscript. I My deepest thanks go to my family members, Inez, O.C., Toni, and Sharon Harris, as well as to Joan Parker, Robert Parham, Dr. Barbara Wood, Dr. Victor Harnack, and Dr. Thomas Kochman for their emotional support and encouragement throughout the course of this dissertation. TABLE OF CONTENTS CHAPTER I ....................................................... 1 INTRODUCTION .................................................. l JUSTIFICATI'ION FOR STUDY .............................. 1 THE NEED FOR SEGMENTATION STUDIES WITH RESPECT TO THE AFRICAN-AMERICAN CONSUMER MARKET ............................................. 2 THE "MYTH" OF THE BLACK PAST AND ITS EFFECT ON MARKETERS ..................................... 3 RESEARCH PROBLEM ...................................... 6 STATEMENT OF RESEARCH PROBLEM ............... 8 PURPOSE OF STUDY ....................................... 9 CHAPTER I NOTES ......................................... 16 CHAPTER II ...................................................... 17 REVIEW OF LITERATURE ON SEX-ROLE ORIENTATION .......... 17 THE TRADITIONAL VIEW OF SEX ROLES .................. l7 IDENTIFICATION THEORY ............................ l9 SOCIAL LEARNING THEORY ......................... 20 COGNTTIVE DEVELOPMENTAL THEORY .............. 22 THE CONCEPT OF ANDROGYNY ........................... 23 GENDER SCHEMA THEORY .......................... 24 SELF-SCHEMA THEORY .............................. 25 SEx ROLE TRANSCENDENCE MODEL ................. 27 AFRICAN-AMERICAN SEX-ROLE ORIENTATION ............ 29 AFRICAN-AMERICAN EARLY CHILDHOOD DEVELOP- MENT ................................................ 29 UNITY/SYNTHESIS v. DUALISM ................. 31 SEX-ROLE ORIENTATION AND PURCHASE DECISIONS ..... 35 CHAPTER III ..................................................... 39 THEORETICAL FOUNDATION AND METHOD ..................... 39 THEORETICAL MODELS .................................... 39 HYPOTHESES ......................................... 44 HYPOTHESIS 1 .................................. 44 HYPOTHESIS 2 .................................. 44 HYPOTHESIS 3 .................................. 45 HYPOTHESIS 4 .................................. 45 RESEARCH QUESTION ................................ 46 OPERATIONALIZATION OF MEASURES ............... 47 HOUSEHOLD PURCHASE DECISION RESPON- SIBILITY ........................................ 47 SEX—ROLE IDENTITY ............................ 48 SEX-ROLE ATTITUDES .......................... 49 METHODS .................................................. so SU BJ ECTS ............................................. 50 PROCEDURES ......................................... 52 BSRI ITEM REDUCTION PROCESS .............. 53 CHAPTER III NOTES ........................................ 57 CHAPTER IV ..................................................... 58 DATA ANALYSES ................................................ 58 HYPOTHESIS 1 .............................................. 58 ANALYSIS OF VARIANCE ....................... 58 ITEM ANALYSIS ................................ 60 SEX-ROLE ATTITUDES ...................................... 62 ANALYSIS OF VARIANCE ....................... 64 ITEM ANALYSIS ................................ 66 RELATIONSHIP BETWEEN HOUSEHOLD PURCHASE DECISION RESPONSIBILITY AND SEX-ROLE ORIENTATION . 68 ANALYSIS OF VARIANCE ....................... 69 HYPOTHESIS 2 .............................................. 71 NULL HYPOTHESIS 2-1 ............................... 72 ANALYSIS OF VARIANCE ...................... 74 ITEM ANALYSIS ................................ 76 NULL HYPOTHESIS 2-2 ......................... 83 ITEM ANALYSIS ................................ 86 HYPOTHESIS 3 .............................................. 90 ANALYSIS OF VARIANCE ....................... 90 ITEM ANALYSIS ................................ 92 HYPOTHESIS 4 .............................................. 96 PRODUCT MOMENT CORRELATIONS ........... 98 RESEARCH QUESTION ...................................... 104 CHAPTER V ...................................................... 105 CONCLUSION .................................................... 105 GENERAL DISCUSSION ..................................... 105 STRENGTHS AND LIMITATIONS OF STUDY .................. 112 ADVERTISISNG\MARKETING IMPLICATIONS ................ 114 RESEARCH IMPLICATIONS ................................. 116 LIST OF REFERENCES ........................................... 121 APPENDIX ........................................................ 132 vi TABLE 1: TABLE 2: TABLE 3: TABLE 4: TABLE 5: TABLE 6: TABLE 7: TABLE 8: TABLE 9: TABLE TABLE TABLE TABLE TABLE TABLE TABLE TABLE 10: ll: 12: 13: 14: 15: 16: 17: LIST OF TABLES Percentage Distribution of Subjects by Demographics .......... 51 Factor Analysis (BSRI Masculine Items), Orthogonal Rotation . . . 54 Factor Analysis (BSRI Feminine Items), Orthogonal Rotation . . . 56 Sex-Role Identity by Race/Sex Group ........................ 59 BSRI Item Analysis for Male Subjects ....................... 61 BSRI Item Analysis for Female Subjects ..................... 63 Sex-Role Attitude by Race/Sex Group ....................... 65 Sex-Role Attitude Item Analysis by Sex and Race .............. 67 Sex-Role Identity by Sex-Role Attitude .................... . 70 Sex-Role Attitude Item Analysis by Sex-Role Identity .......... 72 Wife Dominant Purchase Decisions by Sex-Role Identity ........ 75 Husband Dominant Purchase Decisions by Sex-Role Identity . . . . 77 Purchase Decision by Sex-Role Identity Item Analysis for Female Spouses ......................................... 79 Purchase Decision by Sex-Role Identity Item Analysis for Male Spouses ........................................... 80 Mean Difference Score for Androgynous Husbands and Wives . . . 81 Mean Difference Score for Sex-typed Husbands and Wives ...... 82 Wife Dominant Purchase Decisions by Sex-Role Attitude ....... 84 vii TABLE 18: TABLE 19: TABLE 20: TABLE 21: TABLE 22: TABLE 23: TABLE 24: TABLE 25: TABLE 26: Husband Dominant Purchase Decisions by Sex-role Attitude ...... 85 Item Analysis (Purchase Decision by Attitude for Wives) .......... 87 Item Analysis (Purchase Decision Attitude for Husbands) ......... 88 Wife Dominant Purchase Decisions by Race/Sex ................ 91 Husband Dominant Purchase Decisions by Race/Sex ............ 93 Item Analysis (Purchase Decision'by Race for Wives) ............ 94 Item Analysis (Purchase Decision by Race for Husbands) ........ 95 Correlations (Sex Roles with Purchase Decisions Responsibility) . . 99 Correlations (Sex Roles with Purchase Decisions Responsibility) . . 101 viii LIST OF FIGURES FIGURE 1: Five Lifestyle Segments of Age 18-49 Black Women .......... 11 FIGURE 2: Sex-Role Orientation and Purchase DecisioResponsibility ..... 40 FIGURE 3: Scale for Measuring Purchase Decision Responsibility ......... 47 CHAPTER I INTRODUCTION JUSTIFICATION FOR STU DY If the African-American1 consumer market was considered a separate country, it would rank as eleventh largest in the free world with buying power equal to that of Canada (1990 Census data). The importance of this segment to the consumer behavior discipline is evident when one considers the potential impact that African-Americans have on the US. economy. For example, ap- proximately 30 million people, or about 12 percent of the total US. population are African-American. About 40 percent of all major American cities have African-American populations of more than 100,000. In metropolitan New York alone, there are over 2 million African-Americans -- the population of a city about the size of Philadelphia. African-Americans’ median incomes have risen steadily over the past several decades, growing at a faster rate than the median income of Anglo-Americans. Furthermore, African-Americans have a median age of 21.4 years, compared to a 28.7 median age for whites. Almost one out of every two African-Americans is under eighteen. African-Americans are also going to school in greater numbers, staying longer than ever before, and widening curriculum choices as well, with more choosing careers in busi- ness and industry. THE NEED FOR SEGMENTATION STUDIES WITH RESPECT TO THE AFRICAN-AMERICAN CONSUMER MARKET Despite these facts, however, researchers in marketing and consumer be- havior have generally tended to ignore the African-American segment, allowing it to remain somewhat of a mystery. The trend has been for marketers general- ly to view all African-Americans as a single, homogeneous market, differing from white consumers only in overt areas of physical differences such as hair texture and skin color, or sociological differences such as income and level of education. For example, marketers frequently position products to the African- American segment, but in so doing fail to recognize the diversity of this con- sumer market. As Clarence Smith, president of Essence, states: "the advertising industry doesn ’t ha ve a problem ackno wledgmg the concept of target marketing when it comes to the general market, but the industry does ha ve a problem doing this Where black consumers and black media are concemed” (Allen, 1981, p. 8-52). Wall (1979) makes a similar observation concerning marketers of cos- metics. Many cosmetic houses, says Wall, claim to have products specifically designed for black women, but, having such products, they assume that they are positioned to the black market. However, it would be a marketing "bowler, " adds Wall, if a company introduced a product for white consumers and assumed that by doing so, it had positioned that product in the general market. "To announce ’1 ha ve a product for Whites’ is hardly positioning" (p. 59). The failure of marketers to recognize a distinct and diverse African- American market is due in part to the "myth" surrounding the African- American past and the historical denial of black culture as an identifiable and legitimate culture. THE 'MYTH' OF THE BLACK FAST AND ITS EFFECT ON MARKETERS The "rnyt " of the black past was, as Herskovits (1941) indicated, that blacks had no past; and this view of black people dominated the thinking of so- cial scientists and the public up to the mid 1960s. AS Glazer states as late as 1963: "fire Negro is only an American and nothing else. He has no values and culture to guard and protect" (Glazer and Moynihan, p. 53). Here, the underlying belief was that any Afro-centric culture belonging to American blacks had been all but destroyed by slavery. What remained was but a pale imitation of Anglo-American culture. This belief prevailed despite earlier scholarly arguments against it by Woodson (1963), DuBois (1924), and Herskovits whose own efforts attempted to "comprehend the entire picture of the Negro, in Africa and the New World, in its historical and functional set- ting" (p. 2). Nonetheless, the view that blacks had no past became the prevail- ing one with profound implications for society and marketing as well. For example, because black people were viewed as having no ethnic cultural past of their own, the term "culture," when it was used with respect to African- Americans, was often viewed as the culture of the white middle-class of which African-Americans were deprived. Thus, terms such as "culturally deprived" and "culturally disadvantaged" were commonly used in the 19603 to describe African-Americans and other urban minorities. Lewis (1969) called this situa- tion a "culture of poverty." Thus, the traditional stereotype of black Americans was that of an economically depressed group, uneducated, nondiscriminating in their buying practices, primarily concerned with emulating white values, and not worthy of special marketing considerations. The proposition that blacks had no distinctive African or African-derived ethnic—cultural heritage also left social scientists and historians with the task of having to account for obvious differences in behavior between blacks and whites in other ways. Nineteenth century observers favored the genetic view: often accurately describing ways in which the culture of black people was dif- ferent from that of whites, but then incorrectly attributing such differences to in- ferior genes. The nineteenth century view was admittedly racist. Much of the observation of black people was done by southern whites; and racial differen- ces, therefore, were generally used to justify the conditions of slavery, discrimina- tion, and poverty that black people were made to suffer. Twentieth century observers, including most liberals, reacted to the ob- viously racist implications of nineteenth century descriptions of black behavioral differences in two ways. One, they either explained such differences as pathological responses to the oppressive forces of caste and class (e.g., Myrdal, 1944), or, two, they saw differences as deviations of white behavior, distorted because of blacks’ social distance from the white mainstream where such be- havior regularly occurred (Stewart, pp. 1-2). As such, no search for a rudimentary structure to black behavior was undertaken, because none was presumed to exist. Underlying this view were well-intentioned efforts to repudiate the implications of "inherent (genetic) inferiority" that earlier ob- servers had attributed to African-Americans to account for behavioral differen- ces. However, a difficulty with this view was that it still agreed that the distinctive behavioral patterns of African-Americans were wrong, or sick; if not as a result of "bad genes," then, as a result of "bad environment." Thus, black people who behaved differently from whites were still stigmatized for doing so. AS a result of these beliefs, there developed a social etiquette that con- sidered it impolite to discuss minority-group differences in public. This rule emerged over a period when such differences were regularly used as evidence of minority-group inferiority. To resist this implication, minority-group members felt it necessary to divert attention away from ways in which they were distinc- tive. This included generally working to prevent public discussion of differen- ces. Some even went so far as to deny that such differences existed. Liberal-minded whites cooperated with minorities in these efforts, since they shared the general view that these differences were signs of minority-group in- feriority. Consequently, they felt that the public discussion of racial differences would be, as Baratz and Baratz (1972) said, "tantamount to discussing a hunchbackis hump with him" (p. 13). In many respects, this view is still held today. When someone attempts to talk about group differences, members from such groups, or others, try to discredit any such observations or conclusions by accusing the observer of "stereotyping." The basis for this attitude is the fear that for members of a group to acknowledge that their group is different is to concede the implication of inferiority -- genetic or social -- that differences have historically been offered as "proof" of. And so, herein lay a crucial marketing dilemma, the unspoken fear among marketers, especially white practitioners, of being openly offensive to African-American consumers by publicly identifying and exploiting those ways in which the behavior patterns of African-Americans and Anglo-Americans differ. This is a critical factor that has hindered past ef- forts to meaningfully segment the African-American consumer market. Furthermore, Since African-Americans were not seen as having a distinc- tive culture of their own, any attitudes or values they held were assumed to be those of the white middle-class. Thus, a major assumption among some marketers was that blacks aspired to be like whites and that they consumed in a manner which reflected a desire for assimilation into mainstream culture. Such beliefs were reinforced by early research on the black consumer market. Bullock (1961), for example, wrote that blacks attempt to surround themselves with symbols of "whiteness," while Bauer and Cunningham (1970) felt that blacks were fighting to attain full membership in American society and as such used consumption of a socially visible nature as a means of showing that they had arrived. Assumptions of this nature provided little, if any, justification for treating African-Americans as a separate market, worthy of specific marketing considerations. AS a result of researchers’ inability to accurately depict black life in America, there now exists a serious need for segmentation studies within the African-American consumer market according to demographic, psychographic and other determinants of purchasing behavior. This would permit effective positioning of existing products and services, new product development to meet the needs of the African-American consumer, and intelligent media selection (Berkman and Gilson, 1986). Efforts in this area have been slow to develop and are, therefore, long overdue. It is this kind of study that the present disser- tation proposes to conduct, specifically with regard to the distribution of sex- role orientations within and across cultural groups. Of primary concern, are sex role differences between the African-American and the Anglo-American cul- tures. RESEARCH PROBLEM A number of researchers in marketing and consumer behavior have focused extensively on the relationship between sex roles and consumer be- havior variables. Consequently, some operationalizations of sex-role identity have appeared in the marketing literature (e.g., Prakash and Flores, 1989; Jaffe, 1990). For example, using Gough’s California Psychological Inventory, Fry (1971) showed that behavior was consistent with sex-role identity. Specifically, more feminine men were more likely to smoke feminine cigarettes than were men identified as masculine. In a study by Gentry, Doering and O’Brien (1978), the CPI Fe Scale (Gough, 1975) and the Personal Attributes Questionnaire (Spence, Helmreich and Stapp, 1974) were utilized to explore male and female perceptions of a wide variety of products and leisure activities and to explore whether sex-role identity within each sex is related to product perceptions. Analyses found only weak support for a congruency between sex-role identity and the use of products with a neutral sexual identity or one opposite of the individual’s sex- role identity. Sex-role identity has also been the focus of a number of studies in advertising. Two such studies (Jaffe and Berger, 1988, and Leigh, Rethans and Whitney, 1987) examined the interaction effect of sex-role orientation and positioning (modern versus traditional) on the advertising effectiveness of ads for low involvement products. Both studies indicated that there is a strong two- way interaction effect between a woman’s sex-role orientation and positioning in advertising. Jaffe and Berger found that women who scored high on masculine per- sonality traits, as measured by the Bern Sex Role Inventory (Bem, 1974), have a higher purchase intent for modern positionings than for traditional position- ings, in comparison to women who score high on feminine personality traits who favor the traditional over the modern positionings. Leigh, Rethans and Whitney segmented women into groups of modern versus traditional based on a life-style inventory which measured sex-role orientation. The researchers found a positive association between the woman’s own sex-role orientation and the role she favored in the advertisements (e. g., women classified as modern in terms of their own sex-role orientation had a higher purchase intent when the ad portrayed a modern woman). Jaffe (1990) also investigated the advertising effectiveness for different positionings of financial services aimed at the women’s market. Using an experimental design, the researcher determined the impact of modern versus traditional positionings on the purchase probabilities of financial services among women with different sex—role identities. Analysis indicated that there is a significant two¢way interaction effect between positioning and sex-role identity on the purchase probability of financial services. Women who score higher on maculinity have a higher purchase probability for the modern positionings com- pared to the traditional positioning, in comparison to women who score lower on masculinity who Show no difference for the modern or traditional positioning. STATEMENT OF RESEARCH PROBLEM Researchers in marketing have shown a great deal of interest in specula- tion about the changing sex roles in society and their possible effects on con- sumer behavior. This is especially noticeable in research regarding the impact of sex-role orientation on the outcome of family purchase decisions. Here, recent studies have shown a strong relationship between sex-role orientation and the degree of household influence, preference agreement, mode of conflict resolu- tion, and decision outcome (e. g., Qualls, 1987). Few studies, however, have focused extensively on race and ethnicity as intervening variables in the relationship between sex-role orientation and the out- come of family purchase decisions. In fact, the entire field of sex-role research in relation to African-Americans, has been termed "the single biggest blind spot in existing sociology" (Hochschild, 1973, p. 1023). PURPOSE OF STU DY That we should expect to find differences in sex-role orientation between African-Americans and Anglo-Americans is well founded in the race and cul- ture literature. For example, Millham and Smith (1981) have noted that: ”In American culture, the limitations placed on full social, economic, and political access and participation of minority groups such as Blacks might be expected to diminish the viability for Black Americans of sex-role differentia tion based upon trait personality competencies. Those ’masculine ’ traits that are coexten- sive with competence, ambitious strivings, and competitiveness in the White culture would probably become positive values for all members of the Black group, which could not afford the luxury of limiting the competition for resources to one gender class. The function of sex-role drllerentia tion might be diminished even fur- ther with the need for sharing responsibilities of home and family, which, in White culture, could be left to the noneconomicall y com- peting sex " (p. 79). Such views stem from a belief that the development of sex role values in Anglo-American culture is based on a pervasive sexual division of labor, in which the nonpossession of a set of trait competencies by males/females does not threaten the economic or the social survival of the group. In societies in which survival is more heavily dependent upon broadly based individual com- petencies, failure to develop significant competencies because of gender would diminish greatly the viability of that group. Evidence that conceptions of masculinity and femininity differ for various ethnic groups in the United States and that sex-typed personality traits are likely to be less prevalent among African-Americans is numerous (e.g., Lad- ner, 1972; Petigrew, 1964; Lewis, 1975; Kochman, 1981). For example, Ladner (1972) notes that young African-American women are taught to be independent, strong, and prepared to take on the responsibility of supporting as well as rais- ing their families -- qualities noticeably different from traditional stereotypes of 10 femininity. Likewise, Epstein (1973) describes black professional women as self- assured, highly career-oriented, and more positive toward one another than are white professional women. Nor are these findings limited to African-American women alone. For example, Petigrew (1964) reported that, when the Min- nesota Multiphasic Inventory (MMPI) was administered to black and white Alabama prisoners and to Wisconsin working class veterans in a Veterans Ad- ministration hospital, black males scored higher than white males on the test’s measure of femininity because they agreed most often with such "fenrinine" choices as "I would like to be a singer, " and "I think that I feel more intensely than most people do. " The test designers rated the black males as more "feminine" than the white males because they used only Anglo cultural norms as references. But as Erikson (1968) notes, "to be a singer and to feel intensely may be facets of a masculine ideal gladly admitted 17' you grew up in a southern community, or for that matter, in Naples [Italy]" (p. 306). Moreover, in her study of black family socialization and sex-roles, Lewis (1975) has posited a range of behavioral traits that whites would consider "mas- culine" or "feminine" but that blacks would consider common to both sexes. These traits are aggressiveness, independence, self-confidence, non-conformity, sexual assertiveness, nurturance, emotional expressiveness, and focus on personal relationships. Whites would consider the first five traits "masculine" and the last three as "feminine." At least one empirical study supports Lewis’ claim. Using the masculine and feminine items of the Bern Sex Role Inventory (1974), African-American and Anglo subjects were asked to rate the desirability in American society of 19 of the scale’s feminine (affectionate, cheerful, childlike, compassionate, does not use harsh language, eager to soothe hurt feelings, flatterable, gentle, gul- lible, loves children, loyal, sensitive to the needs of others, shy, soft spoken, 11 $1256 Conservative/Traditionally 32% Views sex, drugs, and liquor as steps along the path of degeneration. Fashion Conscious 31% Liberated in ideas about sex, liquor, and drugs: likes to try new hair styles and would rather dress for fashion than comfort. Independent 16% Outgoing and on the way up. Financially secure and independent. Prefers living in suburbs than city. Girl Next Door 1 2% Has strong moral values; likes trying new hair styles as much as she likes to bake, which she often does. Conservative Thinker 9% Shops for values, disapproves of in- stallment purchases and believes that men should rule the household while women take care of it. Figure 1: Five Lifestyle Segments of Age 18-49 Black Women 12 sympathetic, tender, understanding, warm, yielding) and 19 of the scale’s mas- culine (acts as a leader, aggressive, ambitious, analytical, assertive, athletic, com- petitive, defends own beliefs, dominant, forceful, has leadership abilities, independent, individualistic, makes decisions easily, self -reliant, self-sufficient, strong personality, willing to take risks, willing to take a stand) personality traits for a man and for a woman. All but three of the masculine traits (in- dividualistic, makes decisions easily, and strong personality) were rated by both male and female Anglo subjects as significantly more desirable for a male than for a female. In contrast, none of the masculine items were rated as significant- ly more desirable for a man than a woman by both male and female African- American subjects. Similar results were found for both groups’ ratings of the feminine items (Harris, 1991). If such behavioral traits are accurate in describing African-Americans, then the sex-role identities of both African-American males and African- American females might best be characterized as "androgynous. " As such, we might predict that black consumers and white consumers who are androgynous would exhibit similar behavior patterns. Such notions are logical suppositions and, therefore, appear worthy of empirical investigations. It might be a mistake, however, to assume that African-American culture is any more homogeneous than Anglo culture in the area of sex-role attitudes and identities. For example, a 1981 survey (Figure 1) reveals five separate life- style segments (consa'va ti ve traditionalist; fashion conscious; independent; girl next door: conservative thinker) among black women alone. Furthermore, some empirical studies suggest that traditional sex-role attitudes and sex-role identities may be as common among blacks as among whites. For example, in an analysis of black women’s and white women’s sex-role attitudes, Gump 13 (1975) found that black women were more likely to define their identity with respect to the roles of wife and mother and were more home centered and more submissive than were their white counterparts. Likewise, Rosen (1977) found that both African-American and Anglo women, on the average, were fair- ly non-traditional, but African-American women tended to be significantly more traditional, in general, than did Anglo women. Moreover, Hershey (1978) of- fers evidence against the assumption that African-Americans are more androgynous than Anglo-Americans in their sex-role identities. She found the differences between African-American men and African-American women in sex- role identities to be as great as those separating Anglo women and Anglo men. According to Hershey, the tendency for women to identify more strongly with "feminine" qualities crosses racial lines, as does the tendency for men to see themselves in more traditionally masculine terms. The fact that these studies report different, and often conflicting, findings can be explained, in part, by the different operationalizations of dependent measures in each study. For example, Hershey’s conclusions are based on each subjects’ ratings of the BSRI masculine and feminine items as each item applies to hirnself/herself specifically. Unlike Hershey, however, the Harris study draws its conclusions from subjects’ cultural definitions of appropriate behavior for men and women in general. In other words, whereas Hershey asked subjects to indicate "how mre" of the person each characteristic was, Harris simply asked subjects to indicate "how desirable" in American culture each characteristic was for a man and for a woman. Moreover, the Harris study was originally designed to test the BSRI as a valid indicator of cultural definitions of masculinity and femininity among African-Americans and Hispanics. Testing the BSRI in this manner was 14 thought to be an important initial step before attempting to apply the instru- ment to minority groups who might, in fact, have different cultural definitions of what constitutes masculinity and femininity. In such instances, the concept of "androgyny, " (as defined in terms of individual personality traits) may have little or no conceptual meaning in that particular culture. Thus, although the goal of the two studies was the same, the methodological processes by which each researcher sought to attain this goal were different. Furthermore, in both the Gump and the Rosen studies the researchers measured "sex-role attitudes" (the extent to which the individual believes that roles should be assigned strictly on the basis of gender) rather "sex-role iden- tity" (the level of "masculine" and "feminine" characteristics adopted and ex- hibited by an individual). Although sex-role identity and sex-role attitudes are both theoretically and logically related, the extent of this relationship has not been extensively investigated, especially among African-Americans. As such, the failure to distinguish between sex-role identity and sex-role attitudes could make any attempt to characterize African-Americans as more or less traditional than Anglo-Americans difficult if not impossible. Questions about cultural differences in sex-role orientation are central to the understanding of consumer behavior. It is argued, for example, that consumers’ sex and sex-role orientations are major determinants of their con- sumer behavior habits (e. g., Prakash and Flores, 1989). But suppose that race and ethnicity are crucial elements in determining a person’s sex-role attitudes and identities; and suppose further that sex-role differentiation among African- Americans is much less restrictive than, and fundamentally unlike, patterns found in traditional Anglo society. If these assumptions are true, then our present knowledge regarding the impact of sex-role orientation on the outcome 15 of family home purchase decisions, which are based mainly on Anglo cultural patterns, may need to be re-examined. The point is that only recently has there been enough contrastive cultural data to enable researchers to make such determinations. Thus, the purpose of this study is to examine, empirically, the impact of sex-role orientation on the outcome of family home purchase decisions among African-American versus Anglo-American consumers. 16 CHAPTER I NOTES 1 The terms African-American and Anglo-American are used throughout this study to designate the cultural patterns and perspectives of black people and the white middle-class, respectively. As such, the term African-American is, at times, used interchangeably with the term black; while the term Anglo-American is used interchangeably with the term white. CHAPTER II REVIEW OF LITERATURE ON SEX-ROLE ORIENTATION Psychologists and sociologists have long understood that the extent to which traditional sex-role attitudes and identities are accepted as desirable by an individual, whether black or white, is directly associated with the ways in which sex roles are initially developed and valued within a given cultural framework. THE TRADITIONAL VIEW OF SEX ROLES The early labeling of physiological sex begins the sex-typing process, whereby a person acquires and values the particular characteristics considered appropriate for his or her sex in that culture (Mischel, 1966). These sex-linked constellations of personality traits, attitudes, preferences, and behaviors that a person learns through the sex-typing process are known as sex (genda') roles (Cook, 1985). As such, sex role is a multidimensional concept whereby "mas- culinity" refers to those characteristics that have traditionally been associated with men and "femininit y " refers to those characteristics that have traditionally been associated with women (e.g., Broverman, Broverman, Clarkson, Rosenkrantz, and Vogel, 1970; Rosenkrantz, Vogel, Bee, Broverman, and Broverman, 1968). The result of this sex-typing process is an individual’s sex role identity: the pattern and level of masculine and feminine characteristics adopted and exhibited in some manner by a person (Cook, 1985). In accordance with sex-role identity, a person also develops sex-role attitudes which govern the extent to which the individual believes that roles should be as- signed strictly on the basis of gender (Hershey, 1978). Thus, differences in the behavior of men and women, and in the meanings and values each assigns to 17 18 the behavior of others, are linked to but not determined by physiological sex. Sex is innate, but sex role for a person is learned. Early sex role research assumed that "healthy" individuals adopted the sex role appropriate to their gender, and demonstrated only those traits judged desirable for that gender by society (e. g., Bandura, 1969; Mischel, 1966; Bron- fenbrenner, 1961; Kolberg, 1966). This assumed process of sex-typing has been referred to as the linear model of sex drlferentiation (e. g., Kaplan and Bean, 1976) whereby it is presumed that genetic differences, physiological differences, and ultimately psychological differences somehow form a logical progression. Here, masculinity and femininity are not only typical in men and women, respectively, but expected and healthy. Variations from the norm in a man’s or woman’s characteristics are therefore viewed as deviant. The biological basis for sex differences posited in the linear model of sex differentiation has not been supported in research (see e.g., Cook, 1985). Social scientists have generally recognized that learning begun after birth is primarily responsible for enduring psychological differences. Up to the mid 19703, how- ever, traditional theories in sex role development (e.g., Bandura, 1969; Kagan, I964; Kohlberg, 1966) still emphasized that the ideal sex roles for men and women were quite different and to a large degree opposite. These traditional theories of sex role development can be distinguished in terms of the types of variables and processes considered most central to the sex-typing process. The most influential of these theories have been identification, social learning, and cognitive-dc velopment. 19 IDENTIFICATION THEORY Identification can be defined as ”a particular kind of imitation: the spon- taneous duplication of a model’s complex, integrated pattern of beha vior without Specific training or direct reward but based on an intirna te rela tionship between the identifier and the model" (Mussen, 1969, p. 718). Identification theory proposes that identification with the same sex parent (or caretaker) presumably permits the child to learn elaborate yet subtle sex-typed patterns of characteristics without direct training, as the child attempts to become more similar to the parent in ideals, attitudes, behaviors, and feelings (Cook, 1985). Original conceptions of identification were based in Freud’s psychoanalytic theory of child personality development whereby Freud proposed a sequence of stages that all individuals supposedly pass through toward appropriate sex-role identity (Roopnarine and Mounts, 1985). During the first two stages (the oral and the anal), children form attachments to and identify primarily with their mothers. The mother is perceived as the primary love ob- ject and the source of warmth and nurturance. During the third stage (Phallic- urethral), identification with the same sex parent is motivated by fear; fear of the loss of love in girls and fear of retaliation in boys. According to Freud, when boys discover that girls do not have penises, they reason that girls once possessed them but that they were cut off. Boys interpret this as proof of what might happen to them. Since the boy’s primary love object is the mother, he perceives the father as a competitor for the mother’s love and fears that the father might castrate him. However, upon realizing that he cannot con- quer the father to win the mother, the young boy represses his sexual desires for his mother and identifies with the strength and authority of the father. 20 Through identification with the father, the aggressive parent, the boy acquires masculine behaviors. For a girl, when she discovers that boys have penises and she does not, She feels cheated and disappointed. AS a result, she turns away from the mother, blaming her for the lack of a penis. Since the mother is the one responsible for sending her ill-equipped into the world, the young girl transfers her love to the father. Despite such renouncement, she realizes that her desires for the father cannot be fulfilled and once again identifies with the mother whose love she cannot afford to lose. The wish for a penis is replaced by the desire to have children. Thus, proponents of identification theory assume that children acquire an extensive range of characteristics through an intimate relationship with their same-sex parent. In the child’s efforts to become more like the parent, Sex appropriate characteristics are obtained. SOCIAL LEARNING THEORY Although Freudian theory may have provided the requisite foundation for a comprehensive theory of psychosexual development, it has always been dif- ficult to prove or disprove (Roopnarine and Mounts, 1985). Consequently, it was viewed by social scientists as inadequate for explaining the process of sex- role development. Social learning theorists believe that parents and significant other adults in the child’s life shape gender-role behavior by rewarding or prais- ing gender appropriate behaviors while punishing or discouraging gender inap- propriate ones (Mischel, 1966; Maccoby and Jacklin, 1974). Research on differential treatment of children, however, has produced conflicting results. For example, while Snow, Jacklin, and Maccoby (1983) reported that fathers treated their infant sons and daughters differently, Roopnarine (1986) failed to 21 find such differences. Additionally, while Fagot (1978) reported differential parental reactions to the sex-typed activities of children, Smith and Daglish (1977) failed to find sex related patterns of socialization. In an attempt to restructure the early conceptions of social learning theory, both Mischel (1977) and Bandura (1977) emphasized the role of mental processes as mediating variables in learning and in the acquisition of behavior. Mischel argued that observational learning and cognitive processes are especially important for acquisition of sex-typed behaviors, but actual performance of be- haviors that the child has already learned is strongly affected by reinforcement. Individuals learn of the different consequences for sex-typed behaviors through observation of these consequences following from others’ behaviors as well as being directly rewarded or punished for certain behaviors themselves. The sexes rapidly learn sex-linked expectations for behaviors by perceiving their parents’ and others’ attitudes and attributes. Gradually, sex-typed behaviors acquire dif- ferent value and meaning for the sexes, and individuals learn to regulate their own behavior by their self-evaluations. Thus, social learning theory emphasizes how the sexes develop different meaning, valuing, and frequency of behavior through observation of others, direct or indirect reinforcement of their own or others’ behavior, and their own cognitive processing of these differences. Reinforcement is primarily important in determining an individual’s choice among possible behaviors in a situation. Later, personal self-evaluation in terms. of an individual’s personal standards and rules becomes more important. 22 COGNITIVE DEVELOPMENTAL THEORY Kohlberg (1966) speculated that changes in sex-typing are the " roduct of general motives to structure, and adapt oneself to, physical-social reality, and to preserve a stable and positive self-image" (p. 166). AS such, children actively create their gender identity and associated sex stereotypes and values through their efforts to understand the world around them (Cook, 1985). In cognitive developmental theory, the sex-typing process is rooted in more general changes in the child’s cognitive understanding of physical objects, which occurs with age. Kohlberg proposed that this common developmental process occurs be- cause children define their sex-role concepts fundamentally in terms of physical sex differences. Sex-typing therefore represents a child’s increasingly sophisti- cated cognitive organization of the world around him or her in sex-linked terms. The ability to categorize boys and' girls occurs at around age three. Gender constancy, the cognitive understanding that one’s gender is invariant, is achieved in conjunction with the ability to conserve other properties. Follow- ing the attainment of gender constancy, children imitate the model that is ap- propriate for their own sex. According to Kohlberg, because we live in a very sex stereotyped world, boys will Show a strong tendency to imitate the masculine behavior of their fathers, while girls will Show a keen tendency to im- itate the feminine behavior of their mothers. Thus, according to the cognitive-developmental theory, the motives for ac- quiring sex-typed patterns of behavior seem to relate to the child’s need to develop a sense of consistency and to boost self-esteem. As such, sex stereotypes, sex-typed values, and same-sex identification are the consequences of a sex-typed identity rather than the cause (Cook, 1985). 23 In accordance with these traditional theories of sex role development, traditional measurements of sex role characteristics viewed masculinity and femininity as the negatively correlated ends of a single continuum. When so defined, high masculinity logically implies low femininity; the absence of mas- culine characteristics necessarily implies the presence of feminine ones. This view is common to each of the three theories discussed above. Furthermore, although each theory generally recognizes the importance of cultural and sociological factors in determining the nature and extent of expression of sex- linked behavior and attitudes, they largely ignore these dimensions in the operationalization of gender categorization. THE CONCEPT OF ANDROGYNY In a departure from previous methods, Bern (1974) developed measures of masculinity and femininity that were based upon the culturally defined value of characteristics for a particular sex and upon the logical and empirical inde- pendence of masculinity and femininity. Bern was originally interested in explor- ing how possession of masculine and feminine characteristics may have an impact on a person’s flexibility of behavior across Situations. Masculinity and femininity were viewed by Bern not as polar opposites but as complementary groups of positive traits and behaviors. Society labels these mutually exclusive, heterogeneous categories of attributes as more characteristic of, and desirable for, one or the other sex. According to Bern, however, although many in- dividuals play sex-typed roles, role repertoire may be conceptualized more mean- ingfufly as "feminine," "masculine," or "androgynous." The highly sex-typed individual, whether feminine or masculine, would have internalized society’s sex- appropriate standards for desirable behavior to the relative exclusion of the 24 other sex-appropriate characteristics. Sex-typed persons will, therefore, be more motivated to correspond to society’s definitions of desirable male and female be- haviors in their behavioral expectations, self descriptions, and actual behaviors. Conversely, the androgynous person is one who is much less sensitive to these definitions of desirable behavior and thus is freed from the need to conform to them. She is able to adapt to a variety of situations, rather than using these sex-typed standards as the only guides for personally desirable behavior (Bem, 1974, 1979). GENDER SCHEMA THEORY (Bern, 1981b) AS in her original view, gender schema also assumes that individuals dif- fer in the extent to which they use culturally based definitions for appropriate female and male behaviors as guides for evaluating their own and others’ be- haviors. The gender schema theory, however, is more specific about how these differences occur, and what their implications for behavior are. Bem defines schematic as having a readiness to sort information into categories on the basis of some particular dimension despite the existence of other dimensions that could serve equally well as a basis for categorization. Thus, being schematic with respect to gender means "spontaneous! y sorting attributes and beha viors into masculine and feminine categories or equivalent classes despite there dif- ferences on a variety of dimensions unrelated to gender" (Bern, 1982, p. 1192). Bem proposed that the process of sex-typing by which children become psychologically masculine or feminine occurs in several steps. First, children learn of the existence of sex-based distinctions throughout society, and which at- tributes are more appropriate to their own sex. This learning process teaches children the network of sex-linked associations, which is the content of the 25 gender schema. Once learned, the gender schema will predispose the child to perceive the world in gender linked terms. Second, the gender schema func- tions to shape individuals’ evaluations of themselves. Children learn that certain aspects of the personality are appropriate for their sex alone and, thus, themselves. As a result, they begin to evaluate their own adequacy as in- dividuals in terms of the gender schema they have recently developed. The gender schema serves as a standard for evaluating personal characteristics and behavior in line with society’s definitions for the sexes. Individuals are motivated to conform to these definitions and as they observe their own con- formity to these standards, the sex-based differentiation of the self-concept is strengthened. A traditional sex-role identity is the result. Gender schema theory, therefore sees sex-typed individuals as having a general tendency to partition the world into masculine and feminine categories and, in particular, to decide on the basis of gender which personal attributes are to be associated with their self-concepts and which are to be disassociated from their self concepts. Sex-typed individuals are thus gender schematic in the sense that they have "a genaahized readiness to encode and organize informa- tion -- including information about themselves - in terms of the culture’s defini- tions of masculinity and femininity, definitions to which they are highly attuned" (Bern, 1982, p. 1193). SELF-SCHEMA THEORY (Markus, etaL. 1982) Markus, Crane, Bernstein, and Siladi (1982) propose a slightly different, and more complex concept of schematic processing in relation to gender. They argue that being schematic means having "summaries and constructions of past beha vior that enable individuals to understand their own social experience and 26 to organize a wide range of information about themselves" (p. 38). Therefore, although virtually everyone will have some understanding and representation of attributes and meanings that could be characterized by the general terms mas- culinity and femininity, not everyone will structure their self -concept around knowledge relevant to masculinity or femininity. For some individuals, how- ever, these networks of meanings are used in thinking about, describing, and evaluating the self. When this occurs, a gender based self-schema is thought to exist. Markus, et a1. postulate that the individual can have a masculinity based self -Schema, a femininity based self-schema, both, or neither. As such, sex- typed persons are seen as being schematic with respect to either masculinity or femininity, but not both; androgynous individuals are seen as schematic with respect to both masculinity and femininity, while undifferentiated individuals are seen as aschematic with respect to gender. AS explained by Markus, et al., the gender schematic individuals who think of themselves as distinctly masculine or feminine are assumed to have a large network of schema-relevant cognitions that are retrieved as a unit when the schema is activated. For example, for the individual with a femininity schema, all of these cognitions are related to the concept of femininity and thus are available to working memory when the feminine schema is activated. For androgynous individuals both masculine and feminine attributes are associated with the self-concept. These individuals have not sharply difl‘eren- tiated themselves with respect to gender and relate some attributes of both mas- culinity and femininity to their self-concept. They are, therefore, able to respond in some instances like individuals with a femininity schema and in other instances like those with a masculinity schema. 27 For the low sex-typed (undifferentiated) individuals, neither masculine nor feminine attributes are associated with the self-concept, and if these in- dividuals have structures of masculine or feminine knowledge, they are not well developed or elaborated. They are therefore seen as aschematic, or truly without self-schemas reflecting conventional aspects of masculinity or femininity. However, exactly how this aschematic state would be manifested is not clear. Unlike Bem’s gender schema model, the self-schema model postulates that male and female sex-typed individuals do not have equivalent gender schemata. That is, the gender connotations of both masculine and feminine stimuli are not equally available to them. Thus, they do not process masculine and feminine information with equal efficiency. However, since, logically, both claims could be correct, both theories can be useful in explaining differences in sex-role attitudes and identities cross-culturally. SEX ROLE TRANSCENDENCE MODEL (Rebecca, at. al., 1976) Rebecca, Hefner, and Oleshansky (1976) assumed a somewhat different approach to the concept of androgyny. According to their transcendence model, sex-role development proceeds through three stages: from undifferen- tiated conception of sex roles (Stage I) to polarized oppositional views of sex roles (Stage II) to flexible dynamic transcendence of sex roles (Stage III). In Stage I, undifferentiated conception of sex roles, the child’s thinking is characterized by globalness. Here, the child is busy organizing perceptual in- formation from the environment. She/he does not yet know that society has imposed a sex-delineated system whereby individuals make judgements on the basis of gender. However, the child soon learns, through a process of differen- tiation, that rudimentary opposites exist (e.g., big/small) before grasping that 28 individuals are either male or female and that society has prescribed Specific roles for each sex. In the polarized Stage II, socialization by parents and society assures the adoption of conventional prescriptions and behaviors. Here, the child becomes cognizant of the need to behave and think in accordance with his/her sex of as- signment and an equal desire to actively reject the opposite pole. In this stage, individuals view the "fit to the stereotypes" as a necessary and a major step in gaining acceptance into adult society. As in Stage I, a transition occurs, only now the individual experiences a "crisis" that is in sharp contrast to prevailing views of the sex-role dichotomy. The individual is now forced to make some comparisons between his/her existing "arsenal" of sex-typed conceptions and some alternative view (e. g., the rights of women). Unfortunately this change in conceptual thinking is not given enough support in a society that is still very sex-typed. The researchers state that, "in our society there is virtually no sup- port for [the shift from stage II to stage III] comparable to that of the I to I] shill... This makes the Stage III shit? all the more difficult to accomplish and very dramatic when it does happen " (p. 204). As such, not every person will make the Shift from stage II to stage III. Finally, in Stage III, sex role transcendence, the individual is supposedly flexible in his/her behaviors and attitudes regarding sex-typing; the individual can move freely from situation to situation and behave and feel appropriately and adaptively. Choice of behavioral and emotional expression is not deter- mined by rigid adherence to culture appropriate sex-related characteristics. The struggle to maintain flexibility continues as the individual grapples with her own conscience and societal insistence on gender based distinctions. This view, therefore postulates that we move from sex-typed conceptions to a stage in which introspection and conflict resolution assist us to confront 29 and deal with ingrained gender dichotomies. Although this approach is plausible, empirical support for the mechanisms it proposes is scarce. As such, the theory’s ability to fully explain how sex-role orientation relates to consumer behavior variables, such as household decision behavior, is not clear. Both gender-schema and self-schema, however, propose that household decisions perceived to be congruent with individuals’ views of themselves as either masculine or feminine is preferrable to household decisions perceived in gender-schema-discrepant terms. As such, both theories are plausible explana- tions of the relationship between sex roles and household decision behavior. AFRICAN-AMERICAN SEX-ROLE ORIENTATION The androgyny models discussed in the previous section all assume that culture-based child-rearing processes, to some degree, are central in defining adult sex-role attitudes and identities. A major concern, then, is whether African-American and Anglo-American child-rearing practices offer Significantly contrasting bases for differentiating behavior in the area of sex role develop- ment. AFRICAN-AMERICAN EARLY CHILDHOOD DEVELOPMENT A high valuation for personal uniqueness is a significant configuration in African-American culture. This is in Sharp contrast to Anglo culture. Koch- man (1971), for example, cites the importance of personal attributes rather than status or office in defining the individual in African-American culture. In Anglo culture, for example, personal interaction and degree of authority are de- pendent upon the social position or office which an individual occupies, while in African-American culture, these factors are determined by the individual’s 30 personal attributes, such as his verbal ability, personality, wit, strength, and intelligence. The cultural significance of these traits can be seen in the ways in which African-American children are socialized to display these behaviors. For example, Young (1970) indicates that behavior in young infants which is defined as random and reflexive in Anglo-American culture is inter- preted as "a manifestation of a specific motive or personality trait " in African- American culture. Thus, infant explorations with hands, feet, and mouth are viewed as "hitting or biting." They are considered expressions of individual traits and the infant is said to be "mad" or "mean." As Young describes, "a mother will duck the flying two month old fist and say ’she sure is mad at me’. " Young also describes common forms of interplay between adults and in- fants and adults and toddlers in which the children are encouraged to be asser- tive, initiating and defiant within bounds. According to the researcher, "the baby is treated as though willful and assertive beyond his natural inclination and able beyond his natural abilities. He is highly stimulated and admired for his assertiveness, and his acceptance of authority is expected to be defiant" (p. 285). Lewis suggests that defining the infant as having a highly developed character has consequences not only for the baby but for the other youngsters in the home who are socialized to view babies in this manner. Children brought up to attribute to infants well defined wills and unique personalities are, at a very early age, developing a conceptualization of the nature of the per- son as an independent, initiating being, which profoundly affects their own self- image and their expectations of others as they move into adulthood. 31 UnityISynthesis v. Dualism Another important concept in African-American life which may explain how sex roles are defined within the culture is the "bringing together of polarities." According to Lewis (1964), this stands in direct contrasts to the Anglo concern with "dualism. " Anglo culture is understood in the estab- lishment of linguistic, analytic, and moral dichotomies, such as subject/object; mind/body; good/bad; and male/female, to name a few. African-American cul— ture, however, is characterized by "unity" and 'lsynthesrs. " The black tradition affirms that good and bad, creative and destructive, wise and foolish, up and down, are inseparable facets of existence (e.g., Bennett, 1964). Therefore, these polarities are not conceptualized as dichotomies. Bennett finds that the existen- tial unity expressed in "good is bad," is in conflict with the Anglo-American dichotomy, "either good or bad." Dualism is most evident in the way traditional sex roles are defined in Anglo-American culture (e.g., Kagan, 1964). Not only does a sex role dictate the adoption of different responses or traits for boys and girls (i.e., boys are assertive and girls are passive) but along many dimensions these traits are not merely contrasting, they are mutually exclusive. They represent a differentiation of polarities. Kagan argues that the concept male and female and the attributes associated with "maleness" and "femaleness" are basic to the language and sex role identity of Anglo-Americans. He writes: "By the time a child is four, he has already dichotomized the world into male and female people and is concerned with boy-girl drfIIerences. By the time he is seven he is intensely committed to molding his beha vior in accordance with cultural standards ap- propria te to his biological sex and he shows uneasiness, anxiety, and even anger when he is in danger of beha ving in ways regarded as characteristic of the opposite sex" m 162). 32 But the unity and synthesis which characterizes African-American culture assumes no such dichotomy in the definitions of masculinity and femininity. Thus, polarities such as assertive/passive and independent/dependent which are seen as mutually exclusive gender indicators in Anglo culture, are viewed as in- dependent personality characteristics which can be applied equally to both men and women in African-American culture. The concept of dualism is also depicted by the manner in which Anglo- American culture handles the parent-child attachment process. Lewis (1964), for example, believes that attachment is a significant component in parent-child relationships and notes that in Anglo culture the mode of attachment varies ac- cording to the sex of the child. Here, he distinguishes between two types of at- tachment behaviors: proximal and distal. Proximal behaviors are those involving physical contact such as touching, holding, and rocking. Distal be- haviors are those performed at a distance such as looking at, smiling, and talking to. Lewis has found that for the first two years, Anglo children are trained to move from what is considered an infantile type of interaction, the proximal mode, to what is viewed as a more adult type of interaction, the dis- tal mode (p. 234-235). However, there is a significant difference in the ages and rates at which this occurs for boys and girls in Anglo culture. While girls are encouraged to spend more time near their mothers, are touched significantly more often than boys, and are trained much less severely and much more slow- ly to move from a proximal to a distal mode of interaction, boys are discouraged from close physical contact and encouraged at an early age to move into a more adult type of interaction. Lewis also found that for the fust two years, girls receive both more proximal and more distal behavior from parents than boys, while from the age of six months boys receive a marked reduction in proximal behavior. He further believes that this differentiation in 33 socialization is intended to deliberately stimulate autonomy and independence in boys at an early age and to encourage dependence in girls. Both Lewis and Bern, therefore, agree that these differential expectations help to define the child’s schema in terms of gender differentiation. Conversely African-American families handle sex role training and role expectations in ways that are reflective of the African-American insistence on the unity of opposites. Thus, unlike the Anglo child, the African-American child is not inculcated with standards which polarize behavioral expectations ac- cording to sex. Attributes such as age and order of birth prove more crucial in differentiating behavior than does biological sex. In its early development, the African-American baby, whether male or female, has a very close personal relationship with its parents and others. According to Young, this attention lasts for about the first three years (p. 282). At this time, an abrupt change oc- curs. Parental attention virtually stops and the child is pushed into the group of older children and comes under the authority of the oldest child who is in charge of the group. Here, age, rather than sex elicits the differences in the way the child is treated and in the way he is expected to behave. It is at this point that the refinement of various role characteristics occurs. Older children, as determined by order of birth, and irrespective of sex, assume the role traditionally assigned to males in Anglo culture; that of assertive leaders and care givers. Younger children, on the other hand, assume the role generally reserved for females in Anglo culture; that of passive followers and care recipients. Thus, many of the behaviors which Anglo-Americans see as more appropriate to one sex than the other, African-Americans view as equally ap- propriate or inappropriate to both sexes because, in their culture, such be- haviors are determined by factors unrelated to sex. 34 Moreover, those sex role differences that do exist are seen more in the nature of contrasts than mutually exclusive traits. For example, Young found that in the southern black community she studied, females as well as males are viewed as individualistic and non-conforming in their behavior. Both husband and wife have authority in the home and both are responsible for the economic support of the family. Also both males and females tend to display confidence and a sense of worth, and both are independent in sexual behavior. Young found that not only is behavior considered appropriate for males in Anglo cul- ture displayed by both women and men in African-American culture, but be- havior that is traditionally associated with females in Anglo culture is characteristic of both sexes in African-American culture. In African-American culture, both males and females display similar styles of child care: they are both nurturing and highly interactive physically with children. Both men and women value personal relationships, both are expressive emotionally and both are more adept at handling the world of interpersonal relationships than the world of objects and the physical environment. This is not to say that sex-role differentiation does not exist in African- American culture. Rather, research tends to indicate that to the extent to which such roles do exist, they do not include widespread differentiation on the basis of personal qualities and characteristics. According to Millham and Smith (1981), it may be that sex-role differentiation among African-Americans is more related to how and in what social and interpersonal contexts these com- petencies are displayed than it is to limiting the acquisition of such qualities in members of a particular sex. The studies in this section all present possible explanations for proposed differences in sex-role orientation between African-Americans and Anglo- Americans. However, empirical support for most suppositions presented in 35 these studies is tenuous at best. For example, the Young (1970) study was con- ducted among members of a small black southern community over twenty years ago. Thus, the generalizability of these findings may be suspect. Furthermore, since the study was conducted in 1970, it is also uncertain whether her initial observations are valid in the 1990s. SEX-ROLE ORIENTATION AND PURCHASE DECISIONS As indicated earlier, a relationship between sex-role orientation and household decision behavior is logical since sex-role learning is believed to be basic to people’s attitudes, abilities, expectations, and behavior (e.g., Bern, 1974, 1975; Bem and Lenney, 1976). For example, Bern predicted and found that androgynous sex-role identities are associated with greater flexibility across situa- tions than are sex-typed sex-role identities. It is also logical, according to Bern, that feminine sex-typed persons should behave in a more expressive, affective manner than masculine sex-typed persons, whereas masculine sex-typed persons should excel, relative to feminine sex-typed individuals, in situations in which in- strumental, assertive, and goal-directed qualities are appropriate. Bern (1975) examined the behavior of sex-typed and androgynous in- dividuals in two situations: The first was a conformity deception paradigm, and the second involved the opportunity to play with a baby kitten. These situations were designed to evoke stereotypically masculine (independence) and feminine (nurturant-expressive) responses. In the conformity study, individually tested students were asked to rate the humor of cartoons, some of which had been objectively judged to be low in humor. Attempts to induce conformity con- sisted of the presentation of a false consensus that non humorous cartoons were rated by others as funny. As hypothesized, Bern found that masculine 36 sex-typed and androgynous subjects of both sexes conformed less frequently to the social pressure of false norms than did feminine sex-typed subjects. In a second experiment of the same series, Bern provided different stu- dents with the opportunity to play with a tiny kitten. During part of the ses- sion, subjects were Specifically instructed to play with the animal; later, subjects had the opportunity to interact with the kitten spontaneously. It was predicted that on this stereotypically "feminine" task, both androgynous and feminine sex- typed subjects would play with the animal more than masculine sex-typed sub- jects. For males, the hypothesis was confirmed. However, among females, feminine sex-typed subjects played with the animal significantly less than androgynous persons. Bern and Lenney (1976) investigated whether sex-typed persons would sys- tematically avoid engaging in cross-sex behaviors relative to androgynous or "sex-reversed" (feminine sex-typed male, masculine sex-typed female) individuals. Subjects representing each BSRI category (male and female androgynous; sex-typed males and sex-typed females; and sex- reversed males and females) were individually recruited to perform a number of brief activities, ostensibly for the purposes of being photographed; it was stressed that proficiency in a task was unimportant. Subjects were required to select one activity from each of thirty pairs of tasks. For some pairs, subjects had to choose between (a) a neutral and a masculine activity, (b) a neutral and a feminine activity, or (c) a masculine and a feminine activity. Small cash payments for performing the tasks were arranged such that sex-inappropriate choices always paid more. Ex- amples of stereotypically "masculine" tasks included nailing boards together and baiting fish hooks; stereotypically "feminine" tasks included ironing cloth nap- kins and winding balls of yarn. As predicted, the researchers found that for each gender, sex-typed subjects were significantly more stereotyped in their 37 activity choices than androgynous or sex-reversed subjects even when sex-ap- propriate tasks paid them less money than they would have received for cross- sex tasks. Post activity questionnaire data also indicated that sex-typed subjects felt more uncomfortable when performing cross-sex activities than androgynous or sex-reversed subjects. Such behavioral validation studies indicate that sex-role identity is capable of predicting individuals’ choices of sex-stereotypical activities. Logical- ly, this observation should extend to consumer behavior activities as well. Accordingly, a number of consumer behavior studies have shown that tradition- al household decision tasks and responsibilities appear to be distributed accord- ing to gender (e.g., Cunningham and Green, 1974; Davis and Rigaux, 1974; Green and Cunningham, 1975). These finding have indicated sex-role differen- ces in decision making with respect to many products. Decisions regarding some products, such as groceries, tend to be the domain of the wife, while the husband makes decisions on another set of products, such as life insurance and automobiles. Thus, one can assume that cultural role expectations, within Anglo- American society, ascribe the influence or power to make a decision to the role position rather than to the individual. This should not, however, be the case within African-American culture. In African-American culture, the personal uni- queness factor suggests that the influence or power to make a decision will be determined, not by the role position of any individual, but by the individual’s personal attributes. Logically, therefore, one might expect to find greater egalitarianisrn in family household purchase decisions among African-American families than in Anglo-American families. One consumer behavior study (Middletown and Putney, 1960) offers sup- port for these assumptions. Middletown and Putney studied 40 families 38 (10 white middle-class families, 10 black middle-class families, 10 white working- class families, and 10 black working-class families). The researchers found that proportionately more black families (17 of 20) made joint decisions than white families (13 of 20). With reference to social class, middle-class black families were more likely to make joint family decisions than any of the other race-class groups studied. According to the researchers, 9 of the 10 black middle-class households resolved their differences in a cooperative, "give-and-take" manner that denoted dominance neither by the husband nor the wife. This was the ex- perience for 8 of the 10 working-class black families. A pattern of patriarchal dominance appeared in l of the 10 black middle—class families and of matriar- chal dominance in 2 of the 10 black working-class families. The Middletown study provides important information concerning the ways in which family household-purchase decisions may be affected by race. However, Since no measure of sex-role orientation by racial group was included, it is not possible to determine whether difference found to exist between the black and white groups were due to difl'erences in sex-role identities or some other intervening variable. CHAPTER III THEORETICAL FOUNDATION AND METHOD The present study examines among African-Americans and Anglo- Americans the nature and prevalence of sex-role differentiation based upon the sex-role orientations of men and women in each of the two groups. It does so within the framework developed by Bem (1974; 1981b), Markus, Crane, Bernstein, and Siladi (1982) and Rebecca, Hafner, and Oleshansky (1976) whose basic premise is: Since sex roles are founded on cultural definitions of appropriate beha vior for males and females, gender is not a uniform! y salient stimulus for all individuals in attributing traits to others. Consequently, gender indicators assume less importance for those individuals whose selflconcept is not organized around gender. Moreover, there is a positive relationship between a person is sex- role orientation (gender based v. non genda' based) and her abilities, expectations, and beha viors. THEORETICAL MODELS Past research has shown that household purchase decision and respon- sibilities appear to be distributed according to gender (e.g., Cunningham and Green, 1974; Davis and Rigaux, 1974; Green and Cunningham, 1975); and that sex-role orientation is instrumental in defining the purchase decision role respon- sibilities of husbands and wives (e.g., Rosen and Granbois, 1983: Qualls, 1984). In the proposed models (Figure 2), the nature of husband’s and wife’s household purchase decision responsibility is structurally determined by the individual’s sex-role orientation. Here, sex-role orientation is defined in terms of two concepts: The first, which is termed sex-role identity is conceptually 39 40 SEX-TYPED MODEL Most Characteristic of Anglo-Americans Traditional Sex-role : Attitudes Traditional Gender-Based Household F Purchase Decision Responsibility ANDROGYNOUS MODEL Most Characteristic of Afiican—Amencans Androgynous ------ - Traditional Sex-role Sex-role Identity Attitudes l . I Flexrble I NonGender-Based I Household I TTT— Purchase - ——————————————— .. Decision Responsibility Figure 2: Sex-Role Orientation and Purchase Decision Responsibility 41 defined as "e valuations of oneself} how an individual describes himself or her- self in relation to cultural indicators of masculinity and femininity. " The second, termed sex-role attitudes is conceptually defined as 'evaluations of traditional sex roles (sex stereotypes): a person ’s approval or disappro val of the traditional standards of appropriate beha vior for men and women. " The importance of measuring both sex-role identity and sex-role attitudes (sex stereotypes) in determining household purchase decision responsibility is il- lustrated by the models. As shown by the model representing sex-typed in- dividuals, the person who defines her identity in terms of traditional sex-role norms (sex-typed sex-role identity) is also likely to approve of the traditional sex stereotypical standards of behavior for the sexes (traditional sex-role at- titudes). Accordingly, sex-typed individuals of both sexes are likely to evaluate an activity or task in terms of traditional stereotypes which define the sex-ap- propriateness or inappropriateness of the activity or task. This is indicated by the arrow extending from sex-typed sex-role identity to traditional sex-role at- titudes. Moreover, traditional sex-role attitudes constantly help the sex-typed in- dividual to reinforce his/her traditional sex-role identity. This process is indicated by the arrow extending from traditional sex-role attitudes to sex-typed sex-role identity. Since this individual’s identity is largely based on traditional sex-role stereotypes, she is, therefore, likely to assign household purchase decision responsibility to the husband or wife according to the traditional gender-based norms which define the sex-appropriateness of that individual task. This is illustrated by the arrow extending from sex-typed sex-role identity to traditional gender-based household purchase decision responsibility. For ex- ample, the sex-typed individual would be more likely to see the wife as being primarily responsible for buying groceries for the family, and the husband as 42 being primarily responsible for purchasing the family car or products related to the automobile. However, unlike the sex-typed individual, the person who has an androgynous sex-role identity does not actively engage in sex-role stereotyping. In other words, her evaluations of individuals or Situations is not based on traditional standards of appropriate or inappropriate behavior for men and women. She may, indeed, be aware of society’s definitions of appropriate be- havior for the sexes, but the effects of such stereotyping on her personal iden- tity is, at best, tenuous. This is indicated by the broken arrows connecting androgynous sex-role identity and traditional sex-role attitudes. Since the androgynous individual’s identity is not based on, nor reinforced by, stereotypical sex-role attitudes, she is likely to be flexible in assigning household purchase decision responsibility to both husband and wife. This is illustrated by the arrow extending from androgynous sex-role identity to flexible, non gender-based household purchase decision responsibility. For example, the androgynous individual would be as likely to view the husband, as the wife, as being primarily responsible for buying groceries for the family or purchasing products related to the family car. Sex-role identity in Anglo-American society has traditionally been defined in terms of culture-based expectations of appropriate and inappropriate behavior for the sexes. Therefore, as indicated in Figure 2, the "Sex- Typed Model" will be more representative of sex- role identities among Anglo- Americans than among African-Americans. Conversely, research indicates that sex-role identity in African-American culture is much less restrictive than tradi- tional patterns found in Anglo-American culture. Accordingly, the sex-role iden- tities of most African-Americans might well be characterized as androgynous. 43 As such, the "Androgynous Model" will be more representative of sex-role iden- tities among African-Americans than among Anglo-Americans. Moreover, as illustrated by the models, the relationship between sex-role identities and sex-role attitudes among Anglo-Americans is stronger than the relationship between these two variables among African-Americans. The as- sumed relationship between sex-role identity and sex-role attitudes among Anglo- Americans suggests that both constructs should be accurate predictors of sex-role related behavior, such as household purchase decision responsibility, among members of this group. This is indicated by the arrows extending from both sex-role identity and sex-role attitudes to household purchase decision responsibility in the Anglo-American model. Conversely, research indicates that sex-role differentiation among African- Americans may be more related to how and in what social and interpersonal contexts gender-related competencies are displayed than it is to limiting the ac- quisition of such characteristics in members of a particular sex. Thus, although most African-Americans might be characterized as 'androgynous" in their sex- role identities, such identities may prove to have little or no value in predicting an individual’s evaluations of traditional standards of appropriate attitudes or behavior for men and women. This tenuous relationship is illustrated by the broken arrows connecting sex-role identity and sex-role attitudes in the African- American model. Here, it is hypothesized that among African-Americans, sex- role attitudes (evaluations of sex-stereotypical roles) will be a more accurate predictor of sex-related behavior than will sex-role identity alone. This is il- lustrated by the arrow extending from sex-role attitudes to household purchase decision responsibility and the broken arrow extending from sex-role identity to the same variable. 44 HYPOTHESES Based on the theoretical framework, several observations are predicted. These predictions are formulated in the following hypotheses. Hypothesis 1 A major assumption underlying the relationship between sex-role orienta- tion and sex-role related behavior for African-Americans is that African- Americans are, in fact, more androgynous in their sex-role identities than are Anglo-Americans. It is important, therefore, to test this assumption empirical- ly. Thus, the first hypothesis is as follows: lhe sex-role identities of African-American male and female sub- jects will be more androgynous than the sex-role identities of their Anglo-Ameriran male and female counterparts. Hypothesis 2 The theoretical models presented in this section also hypothesize that a spouse’s household purchase decision responsibility will be determined by that person’s sex-role orientation. The second hypothesis tests this assumption as follows: lhere is a significant relationship bet ween sex-role orientation and household purchase decision responsibility such that spouses with sex-typed sex-role identities and traditional sex-rule attitudes will as- sume more responsibility in purchase decisions traditionally defined as gender appropriate and less responsibility in purchase decisions traditionally defined as gender inappropriate than wrll spouses with androgynous sex-role identities and flexible sex-role attitudes. 45 Hypothesis 3 A third hypothesis is offered to test race as a factor in determining household purchase decision responsibility. In accordance with the models, the third hypothesis is as follows: There is a significant relationship bet ween race and household pur- chase decision responsibility such that Anglo-American spouses will assume more responsibility in purchase decisions traditionally defined as gender appropriate and less responsibilit y in purchase decisions traditionally defined as gender inappropriate than will African-American spouses. Hypothesis 4 Finally, a fourth hypothesis is posited in relation to the African- American and the Anglo-American models presented in Figure 3. As illustrated by these models, both sex-role identity and sex-role attitudes are accurate predic- tors of household purchase decision responsibility among Anglo-Americans. However, among African-Americans, sex-role attitudes are hypothesized to be better predictors of household purchase decision responsibility than is sex-role identity. Thus, the fourth hypothesis states: Both sex-role identity and sex-role attitudes are accurate predictors of household purchase decision responsibility for Anglo-American husbands and wives. However, among African-American husbands and wives, sex-role attitudes Will be a more accurate predictor of household purchase decision responsibility than wrll sex-role iden- tity. In hypotheses 2, 3 and 4, household purchase decision responsibilit y is conceptually defined as "the degree to which husbands and wives dominate household decisions." Here, the nature of the husband’s and wife’s dominance is structurally determined by three variables: sex-role identity, sex-role attitudes 46 and race. Based on the literature (e.g., Green and Cunningham, 1975; Qualls, 1987), decisions related to groceries, furniture, and clothing for the children traditionally tend to be wife-dominated, while decisions related to automobile, major appliances, and family savings traditionally tend to be husband- dominated. These Six general categories have, therefore been selected for use in this study. An attempt was made, however, to make the product categories of furniture and major appliances more "gender specific." For example, instead of questioning subjects about purchase decisions related to furniture in general, they were asked specifically about decisions related to the purchase of bedroom furniture. The general category of furniture might include items that could traditionally fall within the husband’s domain (e.g., television set, recliner). It was felt (face validity), however, that decisions related to the purchase of bedroom furniture might be more within the traditional domain of the wife. Furthermore, purchase decisions regarding some major appliances such as refrigerators, ovens, and dishwashers, would appear to fall more within the traditional domain of the wife than the husband. Therefore, on the question- naire, air conditioner, hot water tank, and furnace were offered as examples of major appliances. RESEARCH QUESTION It has been suggested in the race and culture literature that among African-Americans the order of birth of a child proves more crucial in differen- tiating role behavior than does biological sex. Although no formal hypothesis regarding the relationship between birth order and family household purchase decision responsibility is posited, sociological data will be analyzed in an 47 attempt to determine if this variable accounts for differences in sex-role identity among African-Americans. The research question is: Is the birth order of an individual a significant indicator of sex- role identit y among African-Americans? OPERATIONALIZATION OF MEASURES Household Purchase Decision Responsibility Subjects were presented with a list of the six products and services2 selected for use in this study (groceries, bedroom furniture, major appliances, family savings, automobile, children’s clothing) together with a number of decisions that have to be made when purchasing each product or in performing each service; For example, "For groceries, who decides when to shop?" "For bedroom furniture, who decides when to buy?. " On a scale ranging from 0 to 100 percent (converted to 1-10 for purposes of analysis), subjects are then asked to indicate his/her level of perceived influence for each decision, as illustrated in Figure 3. Scores were then summed for each product decision. Thus, opera- tionally, household purchase decision responsibility (HPDR) is defined as "the GROCERIES 0% 50% 100% When to shop For purposes of I I I I I I I I I I analysis I1IZI3I4|5|6I7I8I9IlO Figure 3: Scale for Measuring Purchase Decision Responsibility 48 degree to which decision responsibrlit y is assumed by the husband or the wife for each of the product related decisions. " Sex-Role Identity Sex-role identity will be measured by subjects’ response to the "mas- culine" and ”feminine" items on a modified version of the Bem Sex-role Inven- tory (Bern, 1974; 1981a). The BSRI asks respondents to indicate on a seven-point scale how well each personality trait (20 feminine and 20 masculine) describes himself/herself. (NOTE: The masculinity and femininity scales used in the present study will consist of only 10 masculine and 10 feminine items, respectively. See BSRI Item Reduction Process under the Methods section.) The scale ranges from 1 (never or almost never true) to 7 (always or almost al- ways true). The total weighted subject population of both sexes can then be split at the median of both the femininity and the masculinity scales. Subjects can then be assigned to one of four categories: high-high (androgynous), low- low (undifferentiated), and two sex-typed categories representing those who have a predominance of one set of characteristics (male/female). Thus, opera- tionally, sex-role identit y is def med as "the amount of masculine and feminine personality traits possessed by an individual. " During the present study, coefficient alpha was computed separately for the masculinity and the femininity scores of both the Anglo-American and the African-American subjects. The results showed both scores to be reliable for both groups (masculinity = .8409, for Anglo-American and .7824 for African- American subjects; femininity = .9239 for Anglo- American and .9343 for African-American subjects). These results are comparable to Bem’s findings for two normative samples during the initial construction of the BSRI 49 (masculinity = .86, .86; femininity = .80, .82). Wilson and Cook (1984) also found similar results for both scales (masculinity = .88; femininity = .78). Bem also found test- retest reliability of the BSRI to be .90 after a four week inter- val. Sex-Role Attitudes Sex-role attitudes were measured by subjects’ responses to a modified ver- sion of Hershey’s (1977) Sex Stereotyping Scale. This instrument is designed to specifically measure the degree to which a person engages in sex-role stereotyp- ing. Here, subjects were asked whether each of ten tasks commonly stereotyped by sex (playing football, taking care of children, cleaning up the house, commanding soldiers on the battlefield, teaching elementary school, making out the family budget, making up the budget of a large corporation, working as a brain surgeon, running a government agency) is appropriate for 1 The scale for each task ranges from 1 (much more appropriate for either sex. one sex than the other) to 7 (appropriate for either sex). Thus, operationally, sex-role attitudes is defined as "the degree to which an individual actively engages in traditional sex-role stereotyping. " Coefficient alpha was computed for the scores of both the Anglo- American and the African-American subject groups. The results showed scores to be highly reliable for each group (.9239 for Anglo-American males, .9074 for Anglo-American females; .850 for African-American males, .8541 for African- American females). 50 METHODS SUBJECTS Subjects were selected by means of a national stratified sample, con- ducted by Market Facts, Inc., a Chicago based consumer mail panel. A com- pletely random national survey would generally not have included sufficient numbers of African-American respondents for comparative analysis. Therefore, to obtain comparative data for this study, African-Americans were oversampled in order to ensure a generous distribution of subjects along racial lines. The survey produced a sample consisting of 1,740 respondents in the fol- lowing race/sex categories: 451 (25.92%) Anglo-American males; 378 (21.72%) African-American males; 481 (27.64%) Anglo-American females; and 430 (24.71%) African-American females. As illustrated in Table l, respondents are distributed nation wide, with the highest concentrations in the Mid Atlantic, East North Central, South Atlantic, West South Central, and the Pacific. An assessment of subjects’ marital status by race/sex category reveals that 358 (79.38%) of the Anglo-American males are married, 1.3% are widowed, 4% are divorced, less than 1% are separated, and 14% indicate that they are single (never married); 267 (70.63%) of the African-American males are married, less than 1% are widowed, 5.56% are divorced, and 22.2% indicate that they are single; 332 (69.02%) of the Anglo-American female subjects indicated that they are married, 10.6% are widowed, 8.7% are divorced, just over 1% are separated and 10.4% indicated that they have never been married (1 subject did not respond); 189 (43.95%) of the African-American female subjects are mar- ried, 12.56% are widowed, 20.7% divorced, 6.7% separated, and 6.05% have never been married. 51 Table 1: Percentage Distribution of Subjects by Demographics Male Female Anglo African Anglo African American American American American (11 =451) (112378) (n=48l) (n=430) New England L ‘ 4 :‘H‘. . 4* . . . ”44;“. ., . ”.1 a .. . u |s88 Mid Atlantic 15.96 15.18 13.72 E.N. Central 17.74 17.67 18.84 W.N. Central 7.54 8.52 1.16 South Atlantic 15.74 16.63 16.28 E.S. Central 4.88 5.82 20.00 W.S. Central 11.09 10.40 19.30 Mountain 5.99 6.03 0.93 ............................ ....79 38. .69 02,, .. .43 95 . Widowed 1.33 10.60 12.56 Divorced 3.99 8.73 20.70 Separated 0.67 1.04 6.74 Never Married 14.63 10 40 16.05 education 1801368001196 o. , ................ _ .7 ... .., .... 021 ......................................... Grade schad- . 8 years 0.67 2.91 1.16 High School 1—3 years 5.32 6.24 6.98 4 years 41.24 39.71 25.35 College l-31ears 31.26 39.42 30.15 25.81 4 years 11.53 14.55 10.60 13.95 5-8 years 9.76 21.96 10.40 26.05 No response 0.22 *Numbers represent percentage distributions 52 The between group as well as the within group variation in socio- economic status is relatively small. For example, as illustrated in the table, the education levels for the subjects is relatively even. The reported median income of African-American male subjects is the highest of all four groups ($37,500), followed by Anglo-American males ($33,750), Anglo-American females ($31,250), and African-American females ($28,750). African-American male sub- jects are also the youngest of the four groups, reporting a median age range of 30-39 years of age. The median age range of the other three groups are all in the 40-49 category. PROCEDURE Data were gathered by way of a self-administered, standardized question- naire (Appendix). Each questionnaire contained four parts; the Bem Sex-Role Inventory; the sex-stereotyping scale; the household purchase responsibility measure; and questions designed to elicit demographic and socioeconomic infor- mation. All households surveyed are members of ”Consumer Mail Panel " Respondents were introduced to the questionnaire by the following messages: Dear Panel Member, Please have the ADULT MALE [ADULT FEMALE] head of your household complete the questions beginning on the other side of this page. Then return the completed questionnaire to us in the enclosed postage-paid en velope as soon as possible. Thank you, in advance, for your help. 53 BSRI item Reduction Process In a recent study to determine the construct validity of the BSRI (Har- ris, 1991), I asked 159 subjects to respond to 19 of the original 20 masculine items, and 19 of the original 20 feminine items of the BSRI.3 Specifically, respondents were asked to utilize a 7-point scale, ranging from 1 (not at all desirable) to 7 (extremely desirable) in order to rate the desirability in American society of each of the personality traits. (For example, "In American society, how desirable is it for a man to be assertive?" "In American society, how desirable is it for a woman to be assertive7") Each of the respondents was asked to rate the desirability in American society of each of the 38 personality traits for a man, or each of the 38 traits for a woman. No subject was asked to rate both. A factor analysis of the intercorrelations among the desirability ratings of the 38 masculine and feminine items was performed. Table 2 shows results for the 19 masculine items and Table 3 shows results for the 19 feminine traits. In each table, traits are presented in alphabetical order. As indicated by the Scree test performed on the eigenvaules, there are four meaningful factors in these data. Using a factor loading .45 as criterion for meaningfulness, it will be noted that all 19 of the masculine items load positively on the first factor. Thus, they all appear to be adequate measures of what we might call the "mas- culine" personality. However, most of these items also logically appear to fall into two distinct categories: items representing instrumentality (e. g., acts as a leader, competitive, dominant) and items representing personal autonomy (e. g., defends own beliefs, independent, individualistic).4 It was thought to be impor- tant, therefore that items on the reduced masculinity scale be representative of both these classifications. Based on this criteria, 8 items (acts as a leader, 54 Table 2: Factor Analysis (BSRI Masculine Items), Orthogonal Rotation FACTORS m _. ._ ., w ....... I ....... 2 3 a Acts as a leader .85274 -. 15922 -.10883 -.06520 Aggressive .49566 -.20757 -.63365 .13839 Ambitious .81996 -.00879 -.15144 -. 14942 Analytical .69815 .02595 .21200 -.05405 Assertive .83004 -.01531 -.15284 -.02189 Athletic .63450 -. 16001 .03776 .07525 Competitive .83347 -.18989 -.07595 -.08536 Defends own beliefs .77783 .04317 -.02101 -.18283 Dominant .78814 -.23697 -.01 192 .06274 Forceful .68271 -.22834 . 12297 .01716 Has leadership abilities .84409 -.0617 -.02600 -.18316 Independent .53886 -.23715 -.58647 .05307 lndividualistic .74079 -.06301 .15621 -.06966 Makes decisions easily .65646 .05636 .01803 -.02360 Self-reliant .72927 -. 10666 -.10808 -.04857 Self-sufficient .78692 -.08927 -.17855 .04175 Strong personality .62350 .01483 .08761 -.18341 Willing to take a stand .85399 -.00072 -.08343 -. 12963 Willing to take risks .64872 -. 12323 -.03660 .05800 55 ambitious, assertive, competitive, dominant, has leadership abilities, individualis- tic, willing to take a stand) were selected for the masculinity scale. Despite their lower loadings, two additional items (independent and athletic) were also selected because each has traditionally been used, in popular culture, as stereotypical indicators of differences between African-Americans and Anglo-Americans (e.g, " African American males are more athletically inclined than are Anglo-American males" and "African-American women are more inde- pendent than Anglo-American women"). Most of the feminine traits (Table 3) load positively on the second fac- tor. These 12 items (affectionate, compassionate, eager to soothe hurt feelings, gentle, loves children, loyal, sensitive to the needs of others, sympathetic, tender, understanding, warm, yielding) loading on the second factor are all rela- tively positive or socially desirable personality traits while those feminine items loading positively on the third factor (flatterable, gullible, soft-spoken) and the fourth factor (childlike, shy) are relatively negative or socially undesirable. Ac- cording to Bern (1979), the negative feminine items were included on the BSRI in order to balance the overall social desirability of the feminine and masculine attributes. However, as the concept of androgyny has evolved, explains Bern, "it seems increasingly inappropriate to define androgyn y in terms of these rela- tively undesirable attributes. " Thus, those feminine items loading on the third and fourth factors were not included on the femininity scale in the present study. The feminine items composing the femininity scale for the present study are affectionate, compassionate, eager to soothe hurt feelings, gentle, loves children, sensitive to the needs of others, sympathetic, tender, understanding, and warm. 56 Table 3: Factor Analysis (BSRI Feminine Items), Orthogonal Rotation FACTORS Affectionate ‘ -.04727 .82333 -.04004 -.02904 Cheerful -.35430 .26495 -.71695 -.05311 Childlike -.09413 .09955 .07686 .80585 Compassionate -.08692 .79433 .14040 .047 54 Does not use harsh language -.l303 .22522 .18958 .17557 Eager to soothe hurt feelings -.09931 .71536 .00284 .29483 Flatterable . 12718 .21257 .49305 .15964 Gentle -.12782 .80924 .10136 .17630 Gullible -.00632 .12224 .73447 .33570 Loves children -.07153 .55148 -.16984 -.04934 Loyal .00541 .64046 -.02905 -.20735 Sensitive to needs of others -.13771 .71851 .0195] -.O7507 Shy -.21696 .27384 .20659 .68628 Soft spoken -.19201 .36252 .63880 .14406 Sympathetic -.15225 .72376 . 18826 .17063 Tender . 13358 .78103 .06534 .25796 Understanding .00046 .81787 .24530 -.01817 Warm .01595 .77925 -.031 36 . 10348 Yielding -.20927 .56525 . 12707 .43 792 57 CHAPTER 111 NOTES Since Hershey originally designed the scale in 1977, a pretest was conducted to determine whether or not the activities were still sex related. Two additional activities were also included in the pretest: working as a receptionist (feminine activity) and digging ditches (masculine activity). However, neither of these items proved to be any more nor less sex re- lated than the original items. Thus, they were not used to replace any of the original ac- trvrtres. Research indicates that measures of influence in purchase decisions should be product specific (e.g., Davis, 1977). The items ”mascuh'ne" and "feminine" were omitted from the questionnaire due to vagueness in meaning and their potential biasing nature. This is in accordance with the fac- tor analysis by Pedhazur and Tetenbaum and others (e.g., Berzins, Welling, and Wetter, 1978; and Gaudreau, 1977) and Bem’s more recent studies (1979, 1981a, 1981b) which indi- cate that the two items are not highly correlated with their own total scale scores, but serve primarily as gender markers. The two items also do not load highly with other items on the BSRI; rather they form a bipolar factor of their own, a factor correlated in several analyses with gender. 4 See e.g., Pedhazur & Tetenbaum, 1979; Ratliff & Conley, 1981; and Ruch, 1984. Typically, the masculinity scale has been shown to be composed of two major factors encom- passing dominanoe/leadership/assertiveness/instrumental activity and independence/autonomy, whereas the femininity scale is represented by a single major factor of empathy/nurturancefrn- terpersonal sensitivity. CHAPTER IV DATA ANALYSIS HYPOTHESIS 1 The first hypothesis addresses the predicted differences in sex-role iden- tity among African-American males and females and their Anglo-American counterparts. The formal hypothesis is as follows: The sex-role identities of African-American male and female sub- jects will be more androgynous than the sex-role identities of their Anglo-American male and female counterparts. Analysis of Variance As stated in chapter 3, sex-role identity is operationally defined as the degree of masculine and feminine traits possessed by the individual. Androgyny is measured by subjects’ responses to the masculine and feminine items on the Bem Sex Role Inventory. Responses for each subject can range . from 1 to 7 for each item, creating a total possible maximum score of 140 for the combined 20 items. As such, the higher a subjects’ score for the 20 items, the more androgynous he/she is considered. A one-way analysis of variance, with race/sex (African-American male, Anglo-American male, African-American female, Anglo-American female) as the independent variable and sex-role iden- tity score as the dependent variable, was performed on the data to test the following null hypothesis: Null Hypothesis: Ho; (11 = (12 = u: = u4 lhe sex-role idenu'ties of Afiican-American male subjects, Afiican- American female subjects, Anglo-American male subjects, and Anglo-American female subjects are equal. 58 59 TABLE 4: Sex-role Identity by Race/Sex group Ho: ul=u2=u3=u4 Ha: At least one of the population means differ ANOVA TABLE Between“ H A. 7418.3326 5 Within 1736 379785.6828 218.7706 T 1739 402040.6805 102.9800 -American 107.3730 females 481 107. 1 13. * PLANNED MULTIPLE COMPARISONS Comparison 1 Anglo-American male/African-American male Comparison 2 Anglo-American male/Anglo-American female Comparison 3 Anglo—American female/African-American female Comparison 4 African-American male/African-American female ‘ The PMC procedure used here is the Dunn or Boa/erroni t method. 60 As indicated by the significant F ratio (Table 4), the null hypothesis. that all population means are equal, is rejected. The mean sex-role identity score for the African-American male subjects is 107.3730 compared to a mean score of 102.98 for Anglo-American male subjects. For African-American female subjects, the mean sex-role identity score is 113 compared to a mean score of 107.3576 for the Anglo-American female subject group. Based on the multiple comparison (Dunn) tests, each of these differences is significant. Thus, The first hypothesis, (the sex-role identities of African-American males and females are more androgynous than the sex-role identities of their Anglo- American male and female counterparts) is supported. Item Analysis In order to offer a more vivid illustration of the differences in sex-role identity between African-American and Anglo-American subjects, separate analyses of each BSRI scale item by race/sex group is provided. Table 5 shows the mean androgyny scores and significance tests results for each of the BSRI masculine and feminine items for Anglo-American and African-American male subjects. As illustrated, African-American males scored significantly higher than An glo-American males on 7 (acts as a leader, assertive, athletic, competi- ti ve, dominant, has leadership abilities, willing to take a stand) of the 10 mas- culine dimensions. Only the differences in the mean scores for the masculine items ambitious, independent, and individualistic failed to reach significance at either the .01 or the .05 level. African-American males also scored significantly higher than Anglo-American males on only 4 (eager to soothe hurt feelings, loves children, sensitive to the needs of others, sympathetic) of the 10 feminine items. Although differences in the mean scores of 3 additional feminine traits TABLE 5: BSRI Item Analysis for Male Subjects. 61 Anglo African American. ,American_.,_ . MASCULlNElTEMSs . . - . -- Acts as leader Ambitious 5.3060 5.4683 Assertive 5.0222 5.3942 Athletic 4.0754 4.5053 Competitive 5.0643 5.5476 Dominant 4.3370 4.7566 Has leadership abilities 5.5388 5.9630 Independent 5.8847 6.0159 lndividualistic 5.5366 5.6402 .Wil'm. 19 take? $1309., _. . _ _ .. 58093 FEMtNtNEiTl—EMS " Affectionate ’ ""5197: Compassionate 5.2306 Eager to soothe hurt feelings 4.8803 Gentle 5.1508 Loves children 5.6341 Sensitive to the needs of others 4.7849 Sympathetic 5.0887 Tender 4.9712 Understanding 5.2927 Warm 5.1042 denotes 1 significant at .01 level of significance. denotes t significant at .05 level of significance. N = 827 (449 AnglocAmeriean, 378 African-American). 62 (tender, understanding, warm) failed to reach significance at either the .01 or the .05 level, they were nonetheless in the predicted direction. Table 6 shows the mean androgyny scores and significance tests results for each of the BSRI items for African-American and Anglo-American women. Similar to the male subjects, African-American female subjects scored significant- ly higher than Anglo-American females on all 10 (acts as a leader, ambitious, assertive, athletic; competitive, dominant, has leadership abilities, independent, indivrdualrstic, willing to take a stand) of the masculine items. However, African-American females scored higher than their Anglo counterparts on only 1 of the feminine items (understanding), and actually scored significantly lower than Anglo-American women on two items (loves children, tender). Here, the data shows that, as expected, it is the masculine dimensions, more so than the feminine dimensions that account for differences in overall sex-role identity scores of the African-American and Anglo-American female groups. Here, evidence supports Lewis’ (1974) claim that personality traits associated exclusive- ly with masculinity and femininity in Anglo-American culture are common to both sexes in African-American culture. SEX-ROLE ATTITUDES Since the first hypothesis is supported by the data, an investigation of possible differences in sex-role attitudes between African-American and Anglo- American subjects was warranted. Thus, a test of the possible differences in sex-role attitude scores between the four race/sex groups was conducted. As noted earlier, sex-role attitude is operationally defined as "the degree to which an individual actively engages in traditional sex-role stereotyping. " Sex-ro 1e stereotyping is measured by subjects’ responses to the items on the Sex Role TABLE 6: BSRI Item Analysis for Female Subjects. 63 Anglo African t Value American. w. American. . ., q Acts as leader Ambitious 5.1726 5.8140 -7.790 * Assertive 4.8046 5.5233 -8.018 * bthletic 3.3971 3.6791 -2.436 * \Competitive 4.4782 4.9233 -4.516 "‘ Dominant 4.0541 4.5419 -4.482 * Has leadership abilities 5.1185 5.9000 -8.677 "‘ lndependent 5.8254 6.6372 41.094" lndividualistic 5.5094 6.1000 -7.278 "‘ Willing-1 to take a stand”. semmrierrms 5 51.57 . 5,1767 .1, Affection ate Compassionate 5.9085 5.8930 .185 Eager to soothe hurt feelings 5.7526 5.6488 1.091 Gentle 5.7401 5.6837 .634 Loves children 6.1143 5.9209 2.153+ Sensitive to the needs of others 5.7193 5.8349 -1.257 S mpathetic 6.0665 6.0116 .666 Tender 5.9044 5.7256 2.044 + Understanding 5.9272 6.1744 -3.l42 + Warm 5.7942 5.7558 .458 denotes 1 significant at .01 level of significance. denotes 1 significant at .05 level of significance. N = 910 (480 Anglo-American, 430 African-American). 64 Stereotyping Scale. The total scale score for the combined 10 tasks for each subject can range from 7 to 70. The lower a subjects’ score for the scale, the greater tendency he/she has to engage in sex-role stereotyping. Conversely, the higher the score, the more flexible the subjects’ attitudes are regarding the gender appropriateness of these tasks. Analysis of Variance A one-way analysis of variance, with race/sex (African-American male, Anglo-American male, African-American female, Anglo-American female) as the independent variable and sex-role attitude score as the dependent variable, was performed on the data to determine which, if any, of the four groups differ in their sex-role attitudes. Results of the ANOVA are presented in Table 7. The significant F ratio indicates that the sex-role.attitude scores of at least one of the groups is significantly different from the scores of the other groups. The mean sex-role attitude score for the African-American male subjects is 50.7910 compared to a mean score of 47.9424 for the Anglo-American male subjects. The mean sex-role attitude score for the Anglo-American female group is 52.1 while the mean sex-role attitude score for the African-American female group is 53.5. As indicated by the multiple comparison tests, the difference between the mean scores of the African-American male group and the Anglo-American male group is significant, indicating that Anglo-American males are more traditional in their sex-role attitudes than are African-American males. The difference be- tween the mean scores of the Anglo-American female (52.1227) and the African- American female (53.5442) groups, however, is not significant. Moreover, a further comparison of group means shows that women of both races are more flexible in their sex-role attitudes (52.1227, Anglo American; 53.5442, 65 Table 7: Sex-role Attitude Score by Race/Sex group Ho: ul=u2=u3=u4 Ha: At least one of the population means differ ANOVA TABLE Sesame Dr*sttmsrsnnaresMeanSqueresFRano ”my Between Groups 3 7604.7837 2534.9279 17.0676 .0000 Within Groups 1736 257835.4140 148.5227 hotel 1739 265440.1977 47.9424 .791 52. 1227 53.5442 African-American African-American * PLANNED MULTIPLE COMPARISONS Comparison 1 Anglo-American male/African-American male Comparison 2 Anglo-American male/Anglo-American female Comparison 3 Anglo-American female/African-American female Comparison 4 African-American male/African-American female .,-3.352..... . , - . 3 -l.758 -3.204 " The PMC procedure used here is the Dunn or Bonferroni t method. 66 African-American) than are men of either race (47.9424, Anglo-American; 50.7910, African-American). Item Analysis All groups showed a tendency to stereotype the jobs/activities, playing football and commanding soldiers in the field by sex (Table 8). As illustrated, the mean for the activity playing football is 3.1245 for Anglo-American males; 2.5 for African-American males; 3.0229 for Anglo-American females; and 2.6535 for African-American females. For the activity, commanding soldiers in the field, the mean score is 4.0710 for Anglo-American males; 3.7407 for African- American males; 4.2557 for Anglo-American females; and 3.9628 for African- American females. Thus, all four groups scored considerably lower on these two activities than they did on any of the other eight activities. Note, how- ever, that current US. practice is to limit the command of soldiers on the bat- tlefield to men; moreover National Football League and National Collegiate Athletic Association rules restrict the participation of women in football at these levels. Consequently, subjects’ scores on these two items may have been influenced not only by their attitudes regarding the sex appropriateness of the tasks, but by policies regarding the ability of women to legally participate in these activities. However, since these two scale items alone serve only to reduce the mean sex-role attitude scores of each of the race/sex groups, the scale itself should still act as an adequate indicator of subjects’ tendency to engage in sex—role stereotyping. Moreover, both of these activities produced ex- treme within group variation, indicating that while some subjects rated the tasks as much more appropriate for one sex than the other (item score = 1), some other subjects viewed these activities as appropriate for either sex 67 TABLE 8: Sex-Role Attitude Item Analysis by Sex and Race. FEMALE 'ismnmimi ” " . ' ‘ ' ' ' Americas. Anglo . .. _ _ , _ 3.0229 2.6535 * Taking care of children 5.4761 5.8698 * Cleaning up the house 5.5899 "' 5.6590 6.1326 “ Repairing telephone wires 4.8413 * 4.4179 5.3558 " Teaching elementary school 5.6984 "‘ 5.6736 5.8977 Making up the budget of a large corporation 5.6852 * 5.9376 5.9163 Working as a brain sur- geon 5.6667 5.9709 5.7047 Running a government agency 5.5741 " 5.9023 6.0907 Making up the family budget 5.6905 5.8067 5.9605 Commanding soldiers on the battlefield 3.7407 + 4.2557 3.9628 + ‘denotes t significant at .01 level. + denotes 1 significant at .05 level. N = 1740 (451 Anglo-American male; 378 African-American male; 481 Anglo-American female; 430 African-American female) 68 (item score = 7). That these two activities produced a wide range of scores is consistent with public opinion regarding the gender appropriateness of each task. For example, although women may be "legally" restricted from playing football at the collegiate and professional levels, a number of girls have success- fully challenged such rules at the high school level. Also, laws restricting women’s full participation in all military activities has prompted a national debate since the Gulf War. THE RELATIONSHIP BETWEEN HOUSEHOLD PURCHASE DECISION RESPONSIBILITY AND SEX-ROLE ORIENTATION A basic premise underlying the theoretical model presented in the pre- vious chapter is the relationship between sex-role identity and sex-role attitudes. Here, it is assumed that a person who defines her personality in terms of tradi- tional sex-role norms is also likely to approve of the traditional sex stereotypi- cal standards of behavior for the sexes. Thus, we should find that sex-typed subjects of both sexes exhibit a greater tendency to engage in sex-role stereotyp- ing than androgynous subjects of either gender. In this particular analysis, androgyny is treated as an ordinal level measure for classification purposes. The total weighted subject population of both sexes was split at the median (masculinity = 55, femininity = 60) of the femininity and the masculinity scales of the BSRI. Female subjects scoring at or above the median on both scales were assigned to the female androgynous category; males scoring at or above the median on both scales were assigned to the masculine androgynous category; females scoring below the median on the masculinity scale and at or above the median on the femininity scale were assigned to the female sex-typed category; and males who scored below the median on the femininity scale and 69 at or above the median on the masculinity scale were assigned to the masculine sex-typed category. A total of 697 subjects were omitted. These included male and female subjects who scored below the median on both the femininity and the masculinity scales (undifferentiated); male subjects scoring at or above the median on the femininity scale and below the median on the masculinity scale (sex-reversed males); and female subjects scoring at or above the median on the masculinity scale and below the median on the femininity scale (sex-reversed females). The procedure rendered a total of 235 androgynous males; 234 sex- typed males; 297 androgynous females; and 307 sex-typed females. Analysis of Variance A one-way analysis of variance with sex-role identity (androgynous males, sex-typed males, androgynous females, sex-typed females) as the inde- pendent variable and sex-role attitude (sex-role stereotyping score) was per- formed to determine if sex-typed subjects are more likely than androgynous subjects to engage in sex-role stereotyping (T able 9). As illustrated, the mean sex-role stereotyping score for the androgynous male subjects is 55.3660 com- pared to a mean score of 48.3547 for the sex-typed male subjects. The sex-role stereotyping score for the androgynous female subjects is 55.6397, while the mean score for the sex-typed female subjects is 53.2443. As indicated by the planned comparison tests, each of these comparisons is significant (p.<.000, androgynous males/sex-typed males; p.<.015, androgynous females/sex-typed females). Notice that the difference in the mean scores of the sex-typed male and female subjects is also significant (p.<.000) whereas the difference in the mean scores of the androgynous males and the androgynous females is almdst negligible (p.<.796). The sex-typed male subjects are, therefore, shown to be 70 TABLE 9: Sex-role Identity by Sex-role Attitude H01ul=u2=u3=u4 Ha: At least one of the population means differ ANOVA TABLE .............................................. Between Groups 3 8352. 5191 2784 1730 18.9103 0000 Within Groups 1069 157389.2162 147.2303 Total 1072 165741.7353 Androgynous males Sex-typed males Androgynous females Sex-typed females 307 48.3547 297 55.6397 53.2443 *PLANNED MULTIPLE COMPARISONS Comparison 1 Androgynous male/Sex-typed male Comparison 2 Androgynous female/Sex-typed female Comparison 3 Androgynous male/Androgynous female Comparison 4 Sex-typed male/Sex-typed female . 11206 . .9876 1.0 1.0533 * The PMC prowdure used here is the Dunn or Bonférroni t method. 71 the most likely of any of the race/sex groups to engage in sex-role stereotyping. Furthermore, although androgynous females are shown to be less traditional than sex-typed women in their sex-role attitudes, the attitudes of both groups are fairly flexible. Table 10 shows the mean score for each of the 10 activities among sex- typed and androgynous subjects. As illustrated in the table, all the groups ap- pear to be fairly flexible in their sex-role attitudes. Androgynous men, however, are significantly more flexible in their attitudes than sex-typed males on all but two of the activities, playing football and commanding soldiers in the field. The mean differences between the androgynous women and the sex- typed women are significant on only 4 (cleaning up the house, repairing telephone wires, teaching elementary school, and commanding soldiers on the battle field) of the 10 activities. Notice that, although each of the groups were fairly traditional in respects to the activity commanding soldiers on the battle lie/d, androgynous women were more flexible in their attitudes regarding this ac- tivity than any of the other groups. HYPOTHESIS 2 The second hypothesis tests the prediction that a spouse’s household pur- chase decision responsibility will be determined by that person’s sex-role orienta- The formal hypothesis is as follows: There is a significant relationship bet ween sex-role orientation and household purchase decision responsibility such that spouses with sex-typed sex-role identities and traditional sex-role attitudes will as- sume more responsibility in purchase decisions traditionally defined as gender appropriate and less responsibility in purchase decisions traditionally defined as gender inappropriate than will spouses with androgynous sex-role identities and flexible sex-role attitudes. tion. TABLE 10: Sex-Role Attitude Item Analysis by Sex-Role Identity. MALE FEMALE AIVITY dreaSexTy" Playing Football 3.0000 2.8162 2.8599 Taking care of children 6.3702 5.2179 * 5.6384 Cleaning up the house 6.3149 5.0385 “' 5.8176 * Repairing telephone wires 5.5702 4.3419 * 4.8632 Teaching elementary school 6.2213 5.0855 * 5.8469 " Making up the budget of a large corporation 5.9872 5.4060 * 5.9805 Working as a brain sur- geon 5.9745 5.3248 * 6.0098 Running a government agency 5.9319 5.4359 * 6.1616 6.0977 Making up the family budget 5.9702 5.6068 * 6.0168 5.9349 Commanding soldiers on the battlefield 4.0255 4.0812 4.7710 4.1954 * I"denotest significant at .01 level. 73 (NOTE: On the questionnaire, respondents were asked to indicate whether they are married, widowed, divorced, separated, or never married. For purposes of this study, respondents who indicated that they are widowed, divorced, or separated were assigned to the husband/wife groups. These in- dividuals were asked to respond to the purchase decision responsibility ques- tions based on past experiences with their spouses.) The following null hypotheses were tested. Null Hypothesis 2-1: H01: U] = 112 = 113 = M The amount of responsibility in household purchase decisions tradi- tional] y defined as wife dominant and traditionally defined as hus- band dominant is the same for husbands with androgynous sex- role identities, husbands with sex- typed sex-role identities, wives with androgynous sex-role identities, and wives with sex- typed sex- role identities. Null Hypothesis 2—2: H02: u1 = u2 = u; = u4 The amount of responsibility in household purchase decisions tradi- tionally defined as wife dominant and traditionally defined as hus- band dominant is the same for husbands with traditional sex-role attitudes, husbands with flexible sex-role attitudes, wives with tradi- tional sex-role attitudes, and wives with flexible sex-role attitudes. Null Hypothesis 2-1 The total weighted subject population (all subjects who indicated on the questionnaire that they were either married, widowed, divorced or separated) of both sexes was split at the median (masculinity = 53, femininity = 57) of the femininity and the masculinity scales of the BSRI. Female subjects scoring at or above the median on both scales were assigned to the female androgynous category; males scoring at or above the median on both scales were assigned to the masculine androgynous category; females scoring below the median on the masculinity scale and at or above the median on the femininity scale were 74 assigned to the female sex-typed category; and males who scored below the median 0n the femininity scale and at or above the median on the masculinity scale were assigned to the male sex-typed category. A total of 510 subjects were omitted. These included male and female subjects scoring below the median on both the femininity and the masculinity scales (undifferentiated); male subjects scoring at or above the median on the femininity scale and below the median on the masculinity scale (sex-reversed males); and female subjects scoring at or above the median on the masculinity scale and below the median on the femininity scale (sex-reversed females). The procedure rendered 191 androgynous males; 210 sex-typed males; 300 androgynous females; and 259 sex- typed females. Additionally, the six products along with the purchase decisions for each were separated based on the traditional gender appropriateness of each pur- chase decision. Thus, purchase decisions regarding groceries, bedroom furni- ture, and clothing for the children were grouped together under the heading of Vlad/tional Wife Dominant Decisions. " Likewise, purchase decisions regarding major appliances, automobile, and family savings were combined to create what is called "Traditional Husband Dominant Decisions. " Analysis of Variance A oneway analysis of variance with sex-role identity (androgynous hus- bands, sex-typed husbands, androgynous wives, sex-typed wives) as the inde- pendent variable and purchase decision score for the wife dominant decisions as the dependent variable was performed to determine if the four groups differ in their responsibility for making these purchase decisions (Table 11). The sig- nificant F ratio indicates that the mean score of at least one of the groups is 75 Table 11: Wife Dominant Purchase Decisions by Sex-role Identity H0: ul=u2=u3=u4 Ha: At least one of the population means differ ANOVA TABLE , 1 . .:'.1 372'... 1....11' . .‘ ‘_.' .'1.>. 1‘ ‘.',‘,‘ . :.: ‘ .,‘,‘.' ‘. ‘3‘ " .' > ' -'r’.’ .‘1 ’ M1“ '.: 1'3, .‘1‘1 1 . . .‘." ::.: '. .'.‘ ‘. ‘. ‘. 4»- . ' ' 1. o Between Groups 3 66814 5720 "'222715240 1361618 . Within Groups 956 156369.7239 163.5667 Total 959 223184.2958 48.2147 58.2467 41.081 62.7954 *PLANNED MULTIPLE COMPARISONS Comparison 1 Androgynous husbands/Sex-typed husbands Comparison 2 Androgynous wives/Sex-typed wives Eigj'iompanson 1 . ...‘ 1.. .. ... .' .. .1..‘...,.'.‘, ‘1': “5.... ’. _ .. .. ..1....~v...t .. 1....“._, ..‘."._._...“.1.‘,‘ is .' 'e '4 71‘ ' '. ’ ‘. .‘ ‘.‘.: ‘.-.:.‘.".'.;.'.:.“.'.:.‘.:1",'.f.:.". :._ _ .‘i' ' '7 , '1‘ :.’.'.’ 7.:i' '.:.'.“.:1'.".'.:.:.:.‘ :.;.:. 1. I15 :-:‘:'- 2. . , .sar-t-iv A n l AIL... AAA. grglxgnhAna Li..ll IL A VA . .. 5,579 giiiifemparisena 4.5487 4.193 * The PMC procedure used here is the Dunn or Bonfencni t method. 76 significantly different from the mean scores of the other groups. The mean score for husbands with androgynous sex-role identities is 48.2147 compared to a mean score of 41.0810 for sex-typed husbands, indicating that, as predicted, androgynous husbands assumed more responsibility than sex-typed husbands in purchase decisions traditionally defined as wife dominant (comparison 1). Also as predicted, comparison 2 shows that androgynous wives assumed less respon- sibility (71258.27) in purchase decisions traditionally defined as wife dominant than did sex-typed wives (x=62.7954). Table 12 shows the results of the ANOVA and the multiple comparison test for sex-role identity by purchase decision score for the traditionally defined husband dominant purchase decisions. The mean score for the androgynous husbands is 54.9738 compared to a mean score of 59.7524 for the sex-typed hus- bands. This shows that, as predicted, androgynous husbands assumed less responsibility than sex-typed husbands in purchase decisions traditionally defined as husband dominant (comparison 1). Also, as predicted, comparison 2 illustrates that androgynous wives assumed more responsibility (112572100) than did sex-typed wives (x=50.9228) in traditionally defined husband dominant pur- chase decisions. Thus, null hypothesis 2-1 (the amount of responsibility in household purchase decisions traditionally defined as wife dominant and tradi- tional] y defined as husband dominant is the same for androgynous husbands, sex-typed husbands, androgynous wives, and sex-typed wives) is rejected. Item Analysis An analysis of each of the 18 purchase decisions was also performed to determine how the groups varied along each dimension. Results of the analysis for the female spouses are presented in Table 13 while results for the male 77 Table 12: Husband Dominant Purchase Decisions by Sex-role Identity H0: ul=u2=u3=u4 Ha: At least one of the population means differ ANOVA TABLE Source F Sumoquwes Mean 8g Between Groups 3 013 .7815 3379.5938 . . Within Groups 956 127750.2185 133.6299 Total 959 137889.0000 57.21 00 Sex husbands 59.7524 Sex wives 50.9228 ‘PLANNED MULTIPLE COMPARISONS Comparison 1 Androgynous husbands/Sex-typed husbands Comparison 2 Androgynous wives/Sex-typed wives §ErrorTProbl 1.1558 .000 .000 .9805 * The PMC procedure used here is the Dunn or Bonferroni 1 method. 78 spouses are presented in Table 14. As illustrated in Table 13, only the decision "how much to spend for furniture" failed to reach significance as rated by the androgynous and the sex-typed wives. All other mean differences between the two groups are significant at either the .01 or the .05 level. Differences be- tween the androgynous and the sex-typed wives along two dimensions (when to buy furniture, and how much to spend on furniture), although significant, were not in the predicted direction. The sex-typed wives actually assumed less responsibility in these purchase decisions than did the androgynous wives. All differences between the means of the sex-typed and the androgynous husbands (Table 14) were in the predicted direction. All but three of these differences (which style of furniture to buy, which make and model of automobile to buy and how to invest the family savings) reached significance at either the .01 or the .05 level. Interestingly, the data also indicate that the traditional wife dominant purchase decisions (groceries, bedroom furniture, and clothing for children) con- tinue to be much more within the domain of wives than husbands, irrespective of sex-role identity. For example, mean differences in purchase decision scores among androgynous husbands and androgynous wives (Table 15) are less than those between sex-typed husbands and sex-typed wives (Table 16) for each of the wife dominant decisions except one (which style of bedroom furniture to buy). However, with the exception of which style of bedroom furniture to buy, the mean scores for the androgynous wives are higher than those for the androgynous husbands, indicating that androgynous wives assume greater responsibility in these decisions than androgynous husbands. These results also indicate that androgynous wives tend to assume a greater amount of responsibility in traditionally defined husband dominant decisions than do androgynous husbands. Except for decisions regarding 79 Table 13: Purchase Decision by SRI Item Analysis for Female Spouses. Wu Animations W t radrtrenal Ire ommant ,, “ ' roceries When to shop 6.4100 7.6216 -6.800 " How much to spend 6.7833 7.5367 -4.382 * Where to shorL 6.5133 7.4826 -5.278 * room Furniture I When to buy 6.6767 6.3089 2.154 + I How much to spend 6.1333 6.2510 -.713 [ Which style to buy 6.2300 6.9421 -3.807 * lClothing for children When to buy 6.3067 6.8842 -3.061 * What styles to buy 6.2300 6.9421 -3.807 * I- How much to spend _ ,w 694.33. .. __.7_.‘o.193.,,_. , 4829.: Iliajor Appliances When to buy 7.1167 6.0000 6.665 * How much to spend 6.2167 5.7529 2.958 * Which brand to buy 7.2900 6.1622 6.243 " Automobile When to buy 6.2433 5.3591 4.924 "' How much to spend 6.1500 5.3166 4.660 "' Make and model 5.4433 4.8996 3.001 "' [Family Savings When to save 6.2267 5.8687 2.590 * I How much to save 6.3200 5.9382 2.531 * How to invest 6.2033 5.6255 3.496 ' denotes 1 significant at .01 level. + denotes 1 significant at .05 level. Table 14: Purchase Decision by SR1 Item Analysis for Male Spouses. 80 I When to shop 5.0105 4.0286 4.674 * I How much to gmnd 5.0314 4.0762 4.713 * I Where to shop 4.9319 3.8667 4.920 * oom Furniture . When to buy 5.6806 5.2905 1.938 + How much to spend 5.8848 5.5000 1.978 + Which style to buy 5.4346 5.1429 1.492 Clothing for children When to buy 5.0419 4.2476 3.572 * What styles to buy 5.6230 4.2286 6.324 * meuch to. Spend. w . _ .. ‘_ .. ., 55759 When to buy 6.0471 6.6333 -2.968 * How much to spend 5.8796 6.5524 -3.640 * Which brand to buy 6.0157 6.9524 -4.398 * Automobile When to buy 6.1571 6.8810 -3.419 * How much to spend 6.5707 7.0048 -2.059 + Make and model 6.7906 7.1048 -1.471 [Family Saviggs When to save 5.6178 6.0143 -2.434 * I How much to save 5.7906 6.2524 -2597 * How to invest savings 6.1047 6.3571 -1.295 " denotes t significant at .01 level. + denotes 1 significant at .05 level. 81 Table 15: Mean Difference Score for Androgynous Husbands and Wives EIIID" Androgynous Bushman Mites Mean Difference radrtronal Wife eminent ” l’OCCI'lCS When to shop 5.0105 -1.3995 * How much to spend 5.0314 -1.7519* Where to shop 4.9319 -1.5814* [Bedroom Furniture When to buL 5.6806 -.9961 * How much to spend 5.8848 .2485 Which style to buy 5.4346 -l.7154* othing for children When to buy 5.0419 -1.2648 * What styles to buy 5.6230 -.6070 * How much to spend_,_,,_ .. . ..,,......... .1—1 ...................... Traditional Husband my}. fij.5“?jji‘5:""";ij'ii .* 1, [Major Appliances When to buy -1.0696 * How much to spend -.3371 + Which brand to buy -1.2743 * Automobile When to buy -.0862 How much to spend .4207 + Make and model 1.3473 * [Family Savings When to save -.6089 * How much to save -.5294 * How to invest -.0986 " denotes mean difference significant at .01 level. + denotes mean difference significant at .05 level. Table 16: Mean Difference Score for Sex-typed Husbands and Wives. 82 Mean When to shop 4.0286 7.6216 -3.5930 * How much to spend 4.0762 7.5367 -3.4605 * Where to shop 3.8667 7.4826 -3.6159 "' room Furniture When to buy 5.2905 6.3089 -1.0184 * How much to spend 5.5000 6.2510 -.7510 * Which style to buy 5.1429 6.7490 -1.6061 * othing for children When to buy 4.2476 6.8842 -2.6366 * What styles to buy 4. 2286 6.9421 -2. 7135 "‘ How much to spend 47000 _._,_..,7.10193 2 3193‘ ajor Appliances When to buy 6.6333 6.0000 .6333 "‘ How much to spend 6.5524 5.7529 .7995 * Which brand to buy 6.9524 6.1622 .7902 * Automobile When to buy 6.8810 5.3591 1.5219 * How much to spend 7.0048 5.3166 1.6487 * Make and model 7.1048 4.8996 2.2052 * amily Savings When to save 6.0143 5.8687 .1456 How much to save 6.2524 5.9382 .3142 + How to invest savings 6.3571 5.6255 .7316 * "' denotes mean difference at significant at .01 level +denotes mean difference at significant at .05 level 83 "how much to spend on the automobile " and "which make and model of automobile to buy, " the mean differences in purchase decision scores between androgynous husbands and androgynous wives are in the opposite direction from those of their sex-typed counterparts. Null Hypothesis 2-2 A one-way analysis of variance with sex-role attitude (flexible/traditional) as the independent variable and purchase decision score for the wife dominant decisions as the dependent variable was performed to determine if husbands with flexible sex-role attitudes, husbands with traditional sex-role attitudes, wives with flexible sex-role attitudes, and Wives with traditional sex-role attitudes differ in their responsibility for making these purchase decisions (Table 17). The significant F ratio indicates that the mean score of at least one of the groups is significantly different from the mean scores of the other groups. As illustrated, the mean purchase decision score for husbands with flexible sex-role attitudes is 49.4699 compared to a mean score of 35.5698 for husbands with traditional sex-role attitudes, indicating that, as predicted, hus- bands with flexible sex-role attitudes assumed more responsibility in those pur- chase decisions traditionally defined as wife dominant than did husbands with traditional sex-role attitudes (comparison 1). Also, as predicted, comparison 2 shows that wives with flexible sex-role attitudes assumed less responsibility (22601218) in purchase decisions traditionally defined as wife dominant than did wives with traditional sex-role attitudes (82636431). Similar results can be seen with respect to the traditionally defined hus- band dominant purchase decisions. Table 18 shows the ANOVA and the multi- ple comparison tests results for sex-role attitude by purchase decision score for 84 Table 17: Wife Dominant Purchase Decisions by Sex—role Attitude. H0: ul2u22u32u4 Ha: At least one of the population means differ ANOVA TABLE IFRW" ..................... "(1724623955 57487. “'52 [328.6320 ”1.00005“ Within Groups 1491 260820.0420 174.9296 Total 1494 433282.4375 , 349 49.4699 T 35.5698 Flexible wives .1218 T 63.6431 ‘PLANNED MULTIPLE COMPARISONS Comparison 1 Flexible husbands/1‘ raditional husbands Comparison 2 Flexible wives/1' raditional wives vaiueSErrorTVaiuesTProol 1;; j;jgg;;;;;_;;;gg; 13.9001 1.0049 13.833 .000 ' “ -3.5212 1.0848 -3.756 .000 j * The PMC procedure used here is the Dunn or Bonfenoni 1 method. 85 Table 18: Husband Dominant Purchase Decisions by Sex-role Attitude. H0: u1=u2=u3=u4 Ha: At least one of the population means difTer ANOVA TABLE ~ Squares Between Groups 3 392602853 13086.7618 4' 799.2819”. Within Groups 1491 196574.4598 131.8407, Total 1494 235834.7452 , 0544 Traditional 61.1 Flexible wives 57.51 T 47.0327 'PLANNED MULTIPLE COMPARISONS Comparison 1 Flexible husbands/Traditional husbands Comparison 2 Flexible wives/Traditional wives ,- . 10,4776 ‘ The PMC procedure used here is the Dunn or Bonfenoni 1 method. 86 the traditionally defined husband dominant purchase decisions. The mean score for husbands with flexible sex-role attitudes is 56.0544 compared to a mean score of 61.1773 for husbands with traditional sex-role attitudes. As predicted, therefore, husbands with flexible sex-role attitudes assumed less responsibility in purchase decisions traditionally defined as husband dominant than did hus- bands with traditional sex-role attitudes (comparison 1). Comparison 2 also indicates that, as predicted, wives with flexible sex-role attitudes assumed more responsibility (x=57.5103) in decisions traditionally defined as husband dominant than did those wives with traditional sex-role attitudes (x=47.0327). The second null hypothesis (the amount of responsibility in household purchase decisions traditionally defined as wife dominant and traditionally defined as hus- band dominant is the same for husbands with flexible sex-role attitudes, hus- bands with traditional sex-role attitudes, wives with flexible sex-role attitudes, and wives with traditional sex-role attitudes) is, therefore, rejected. Item Analysis An analysis of each of the 18 purchase decisions was also performed to determine how the groups varied along each of these dimensions. Results of the analysis for wives are presented in Table 19 while results for the male spouses are presented in Table 20. As illustrated in Table 19, mean differences between the purchase decision scores of wives with flexible sex-role attitudes and wives with traditional sex-role attitudes for 5 of the 9 traditional wife dominant decisions (how much to spend on groceries, when to buy bedroom furniture, how much to spend on bedroom furniture, which style of bedroom furniture to buy, and how much to spend on clothing for the children) failed to reach significance at either the .01 or the .05 level. However, mean Table 19: E1 '11 I 11' I 87 Item Analysis (Purchase Decision by Attitude for Wives). radmonal . . Wife ._ fmm _, amt in” oceries When to shop 6.7179 7.6839 -6.314 " How much to spend 7.1632 7.4360 -1.811* Where to shop 6.8046 7.4687 -4.312"‘ [Bedroom Furniture When to buy 6.6138 6.3488 1.830 How much to spend 6.3103 6.3134 -.021 Which style to buy 6.9425 7.0981 -1.126 othing for children When to buy 6.3632 7.2098 -5.475 “ What styles to buy 6.3632 7. 2098 -5.475 * How much to spend 6:64.14, 1 .4 _ . AA 6901931”, . .. 1506 ajor Appliances When to buy 6.8575 5.4687 9.989 "‘ How much to spend 6.2736 5.3678 6.885 "' Which brand to buy 6.8989 5.9101 6.612 * Automobile When to buy 6.3724 4.8992 9.888 * How much to spend 6.2713 4.9782 8.540 ’ Make and model 5.2276 4.7984 2.828 "' [Family Savian When to save 6.5218 5.2425 11.074 "‘ I How much to save 6.7379 5.2207 12.237 " How to invest savings 6.3494 5.1471 8.693 " denotes t significant at .01 level. + denotes t significant at .05 level. 88 Table 20: Item Analysis (Purchase Decision by Attitude for Husbands). flexible Iranitinnal t I When to shop 4.8424 4.0378 4.903 * I How much to spend 4.8138 4.0581 4.679 * Where to shop 4.7880 3.6366 6.975 * m Furniture When to buy 5.8338 4.5988 7.955 * How much to spend 6.0544 5.0320 6.537 "‘ I Which style to buy 5.8567 4.2762 10.676 * kilothing for children When to buy 5.5186 3.1890 14.054 * What styles to buy 5.9083 2.9942 17.640 * How much to spend M .1 ,V. 152.8539 M _ , . 3.7471. , . _ , H 1 1.821.1*+ ajor Appliances When to buy 6.5501 6.3750 1.175 How much to spend 6.2120 6.6890 -3.382 * Which brand to buy 6.5129 6.8837 -2.313 + Automobile When to buy 6.2006 7.1919 -6.207 * How much to spend 6.5530 7.2907 —4.545 * Make and model 6.5100 7.2500 -4.548 "‘ [Family Sam When to save 5.6074 6.2994 -5.588 * How much to save 5.7622 6.6512 -6.689 * How to invest savings 6.1461 6.565 -2.701 " denotes t significant at .01 level. + denotes t significant at .05 level. 89 differences between the two groups for all 9 of the traditional husband dominant decisions were significant (p.<.01) and in the predicted direction. Thus, it would appear that wives with flexible sex-role attitudes do not neces- sarily assume significantly less responsibility in traditionally defined wife dominant decisions. Most of the difference between these two groups is ac- counted for by the greater amount of responsibility assumed by wives with flexible sex-role attitudes with respect to the traditionally defined husband dominant purchase decisions. Table 20 shows the mean differences between husbands with flexible sex- role attitudes and husbands with traditional sex-role attitudes for each of the 18 purchase decisions. As indicated, the mean differences between the groups for each of the traditional wife dominant purchase decisions are significant (p.<.01) and in the predicted direction, indicating that husbands with flexible sex-role at- titudes do in fact assume more responsibility in purchase decisions traditionally defined as wife dominant than do husbands with traditional sex-role attitudes. Very similar results can be seen with respect to the traditional husband dominant purchase decisions. Here, mean differences between the groups on all but one decision (when to buy major appliances) are significant at either the .01 or .05 level. Thus, husbands with flexible sex-role attitudes are also shown to assume less responsibility than husbands with traditional sex-role attitudes in purchase decisions traditionally defined as husband dominant. 90 HYPOTHESIS 3 The third hypothesis tests race as a determinant of household purchase decision responsibility. The formal hypothesis is as follows: There is a significant relationship bet ween race and household pur- chase decision responsibility such that AngIo-American spouses of' both sexes will assume more responsibility in purchase decisions traditionally defined as gender appropriate and less responsibility in purchase decisions traditionally defined as gender inappropriate than will African-A men'can spouses of' either sex. The following null hypothesis was tested. Null Hypothesis 3: Ho; (11 = u2 = 113 = u The amount of responsibility in household purchase decisions tradi- tional] y defined as wife dominant and traditionally defined as hus- band dominant is the same for Anglo-American husbands, African- American husbands, Anglo-American wives, and African -American wrves. Analysis of Variance A one—way analysis of variance with race/sex (Anglo-American husbands, African-American husbands, Anglo-American wives, African-American wives) as the independent variable and purchase decision score for the wife dominant decisions as the dependent variable was performed to determine if the four groups differ in their responsibility for making these purchase decisions (Table 21). The significant F ratio indicates that the mean score of at least one of the groups is significantly different from the mean scores of the other groups. The mean score for the Anglo-American husbands is 37.1366 com- pared to a mean score of 49.8007 for the African-American husbands. This in- dicates that, as predicted, African-American husbands assumed more responsibility than Anglo-American husbands in purchase decisions traditionally 91 Table 21: Wife Dominant Purchase Decisions by Race/Sex. H0: u1=u2=u3=u4 Ha: At least one of the population means differ ANOVA TABLE ..................... , _, .. , . , . z 7 , . y 1 ~.;,,-:;:j.;:;:53'.g2s;-;;;$::=‘:jEsfg:'-‘;?;53333:?2:gigi:3:353:53532932211:?1? Squares ‘iit}.FiégizfsQ;;E;:;;a:}:§igé2‘iiji5:551,izfgiffiig'i ». x 3.555%???“ Between Groups 3 166388.7367 554629122 315.1960 . Within Groups 1478 260073.6736 175.9632 Total 1481 426462.4103 , ’ p ’ 37.1366 49. 64.0484 9203 *PLANNED MULTIPLE COMPARISONS Comparison 1 Anglo-American/African-American Husbands Comparison 2 Anglo-American/African-American Wives * The PMC procedure used here is the Dunn or Bonfermni tmethod. 92 defined as wife dominant (comparison 1). Also as predicted, comparison 2 shows that African-American wives assumed less responsibility 012589203) in purchase decisions traditionally defined as wife dominant than did Anglo- American wives (i=64.0484). Table 22 shows the results of the ANOVA and the multiple comparison test for the traditionally defined husband dominant decisions by race/sex. The mean purchase decision score for the Anglo-American husbands is 61.242 com- pared to a mean score of 54.8514 for the African-American husbands, indicat- ing that, as predicted, African-American husbands assumed less responsibility than did Anglo-American husbands in traditionally defined husband dominant decisions (comparison 1). Also, as predicted, comparison 2 illustrates that African-American wives assumed more responsibility (‘x=56.2198) in traditionally defined husband dominant purchase decisions than did Anglo-American wives (it-49.7488). Thus, null hypothesis 3 (The amount of responsibility in household purchase decisions traditionally defined as wife dominant and tradi- tional] y defined as husband dominant is the same for Anglo-American hus- bands, Afiican-Amen’can husbands, Anglo-American wives, and African-Amen'can wives) is rejected. Item Analysis Again, an analysis of each of the 18 purchase decisions was performed to illustrate how African-American and Anglo-American subjects vary along each dimension. Results for the Anglo-American and the African-American wives are presented in Table 23, and results for the African-American and the Anglo-American husbands are presented in Table 24. Mean differences between African-American wives and Anglo-American wives are significant (p.<.01) for 93 Table 22: Husband Dominant Purchase Decisions by Race/Sex. Ho: ul— — — —=u3 Ha: At least one of the population means differ ANOVA TABLE Between Groups 3 27353.1493 91 17. 7164 65.5552 Within Groups 1478 205566.9937 139.0846 Total 1481 232920.1430 African-American H *PLANNED MULTIPLE COMPARISONS Comparison 1 Anglo-American/African-American Husbands Comparison 2 Anglo-American/Anglo-American Wives * The PMC procedure used here is the Dunn or Bonferrcni t method. 94 Table 23: Item Analysis (Purchase Decision by Race for Wives). Anglo— African- Ametiean American I om -.; t _ , 7;; ‘g?=;::;:;sr'::tasssg-rtrfirgé-at: A???33?;ii:iiiizgfifi.5:113297325253;5;;5?;;}:§§;§'j'iszfsfiéfzj‘jZf-fif1555515:i{$5532332;???‘5551iii;27%jjijLi-jgiifljr533?;T?Iii:413i:5i}%§3:?:*iiii.3139'5295'32?-??i?r‘?féi*?§3i€:15:51:iiigsgi'iaijiéi roceries When to shop 7.7327 6.4712 8.274 * How much to spend 7.6359 6.8709 5.121 * Where to shop 7.5230 6.6239 5.825 * Bedroom Furniture When to buy 6.4954 6.4918 .025 * How much to spend 6.5461 6.0330 3.519 * Which style to buy 6.8456 7.1951 -2.499 * Clothing for children When to buy 7.1290 6.2885 5.400 “‘ What styles to buy 7.0230 6.6071 2.647 * How much to spend 7.1175 6.3242 4.603 " Traditional Husband [Dominant Decisions ajor Appliances When to buy 5.7765 6.7445 -6.846 * How much to spend 5.7143 6.0137 -2.242 * Which brand to buy 5.9977 6.9973 -6.662 * Automobile When to buy 5.1221 6.3874 -8.439 * How much to spend 5.1452 6.3132 -7.711 * Make and model 4.8525 5.2473 -2.603 * amily SJavings When to save 5.7972 6.0797 -2.370 * I How much to save 5.7995 6.3269 -4.081 * How to invest savings 5.5438 6.1099 -4.050 * " denotes t significant at .01 level. N = 798 (434 Anglo-American; 364 African-American). 95 Table 24: Item Analysis (Purchase Decision by Race for Husbands). Anglo- African- Amedean American I When to shop 4.2655 4.6858 -2.537 * L How much to spend 4.1211 4.8412 -4.439 * Where to shop 3.8608 4.6757 -4.860 * room Furniture When to buy 4.6521 5.9764 -8.431 "‘ How much to spend 5.1289 6.1149 6.228 * Which style to buy 4.5103 5.8041 -8.521 "‘ lothing for children When to buy 3.3918 5.6588 43.412" What styles to buy 3.3041 6.0169 -15.903 * How much to spend . . ._.,,3-?921,, _ . w. . _ l'fmdlmnal H“Shana11::$3..,‘E'ifii'ifffzt‘i'7 “”5v‘:*55?" V 'i‘ni: I-fifeliii::53;:t;::5:s;'5.::?13.39};7 :4-‘2 -. it;-;:.~>r.-.3:;:~;.:. .: : . . .32. 038 1 _.6-02,79 ‘ When to buy How much to spend 6.6340 6.1834 3.113" Which brand to buy 6.7320 6.6250 .656 Automobile When to buy 7.1675 6.0338 6.964 * How much to _spend 7.3299 6.3446 5.991 "‘ Make and model 7.2784 6.3311 5.753 * amily Savings When to save 6.2887 5.4561 6.435 * How much to save 6.6108 5.6284 7.001 "‘ How to invest savings 6.7191 5.8142 5.962 * ‘ denotes t significant at .01 level. N = 684 (388 Anglo-American; 296 African-American). 96 all but 1 (when to buy furniture) of the 18 purchase decisions. One of the mean differences (which style of furniture to buy) between the groups is, how- ever, not in the predicted direction. Results here indicate that, as predicted, African-American wives assume less responsibility in traditionally defined wife dominant purchase decisions and more responsibility in traditionally defined hus- band dominant purchase decisions than do Anglo-American wives. Table 24 shows similar differences between African-American and Anglo- American husbands. Mean differences between the African-American and the Anglo-American groups are significant (p.<.01) and in the predicted direction for each of the traditional wife dominant purchase decisions. Mean differences between the two groups is significant (p.<.01) for all but two (when to buy major appliances and which brand of major appliance to buy) of the traditional husband dominant decisions. All mean differences, here, are in the predicted direction. Thus, as predicted, African-American husbands assume more respon- sibility in traditionally defined wife dominant purchase decisions, and less responsibility in traditionally defined husband dominant decisions than do Anglo-American husbands. As such, hypothesis 3 is supported. HYPOTHESIS 4 Among Anglo-American subjects of both sexes, we expect to find the fol- lowing relationship between sex-role identity and household purchase decision responsibility: masculinity, as measured by the BSRI, should be positively correlated with the husband dominant purchase decisions and negatively corre- lated with the wife dominant purchase decisions; femininity, as measured by the BSRI, should be positively correlated with the wife dominant decisions and negatively correlated with the husband dominant decisions. However, since sex 97 roles in African-American culture are less salient than in Anglo culture, the relationship between masculinity/femininity and household purchase decision responsibility should not be as strong among the African-American subjects. Thus, the following null hypothesis was tested: Null Hypothesis 4—1: H01: 1'] = 17 The correlation bet ween sex-role identity and household purchase decision responsibility is the same among African-American and Anglo-American subjects. Recall that sex-role attitude is operationally defined in terms of a subject’s tendency to engage in sex-role stereotyping. Here, the sex-role attitude (sex stereotyping) scale was divided between traditionally defined masculine ac- tivities and traditionally defined feminine activities to create two separate measures. Activities comprising the masculinity scale are: playing football; repairing telephone wires, making up the budget of a large corporation; work- ing as a brain surgeon; and commanding soldiers on the battlefield. The ac- tivities used to create the femininity scale are: taking care of children; cleaning up the house; teaching elementary school; running a government agency; and making up the family budget. Among male and female subjects of both races, we expect to find the fol- lowing relationship between sex-role attitude and household purchase decision responsibility: the masculine activities will be positively correlated with the hus- band dominant decisions and negatively correlated with the wife dominant decisions; and the feminine activities should be positively correlated with the the wife dominant decisions and negatively correlated with the husband dominant decisions. 98 Thus, the following null hypothesis was tested: Null Hypothesis 4—2: H02: 1'] = r2 The correlation bet ween sex-role attitude and household purchase decision responsibility is not the same among African-American and Anglo-American subjects. Product Moment Correlations The following product moment correlations were calculated between sex- role-identity and household purchase decision responsibility for both Anglo- American and African-American male and female subjects: (1) BSRI masculinity scale with wife dominant decisions; (2) BSRI masculinity scale with husband dominant decisions; (3) BSRI femininity scale with wife dominant decisions; and (4) BSRI femininity scale with husband dominant decisions. Correlations for Anglo-American and African-American males are presented in Table 25 while the correlatations for the Anglo-American and African-American female subjects are shown in Table 26. As illustrated in Table 25, for the Anglo- American males, the correlations between the BSRI masculinity scale and the wife dominant decisions (r=-.3604), the BSRI femininity scale and the wife dominant decisions (r=.4498), and the BSRI femininity scale and the husband dominant decisions (r=-.1630) are all significant at the .001 level, while the cor- relation between the BSRI masculinity scale and the husband dominant decisions (r=.1228) is significant at the .01 level. Moreover, each of these cor- relations is in the predicted direction. However, among African-American males, none of the correlations between sex-role identity and purchase decision respon- sibility (r=.0956 for masculinity with the wife dominant decisions; r=.0228 for masculinity with the husband dominant decisions; r=.0762 for femininity with the wife dominant decisions; r=.0918 for femininity with the husband dominant decisions) are significant at either the .001 or the .01 level. 99 Table 25: Correlations (Sex-roles with Purchase Decision Responsibility) ANGLO-AMERICAN MALES Purchase Decision Responsibility Wife Dominant v n :1 vrvr-w v" . r" x Husband Dominant v! v-y ‘vl x..y. .1 ' y..'--vv-vn‘vr-u Sex-roleidennty _ BSRI masculinity scale - 3604 1'“ A 1228 “M,ch_- BSRI femininity scale 4498 i" if gxérole attitude .................. "15%.: $1.: 2.5-5:5 i5;.é:5iii?:iiiiiiiéii?£22253;5.22;;2-23357522i?;§.:§:{:;J§'[if iii}? «E19:25;ifjiéiiiiiiégiiT9525€5£é§2§2§ Masculine activities -.5985 *“ 2385 *‘nhmmnmm Feminine activities .5756 *"' -.2662 ** AF RICAN-AMERICAN MALES Purchase Decision Responsibility Wife Dominant Husband Dominant Decisions . ssmhascuhnny scale .0965 H L H BSRI {mummy scale .0762 W " -;]j§,;'_;;.g ............................................... ...7.....1...._11..u......1. .. ...wu.z ...n....-.1.r~\. .1..71—\ ..... Masculine aClivitifi -2961" 1550 ** Feminine activities .2114 " -.l483 *" Numbers are product moment correlations " denotes r significant at .001 level. "‘ denotes r significant at .01 level. 100 Table 26 shows that the same relationship exists for the female subjects. As illustrated, for the Anglo-American females, the correlations between the BSRI masculinity scale and the wife dominant decisions (r=-.3647), the BSRI femininity scale and the wife dominant decisions (r=.2618)), the BSRI mas- culinity scale and the husband dominant decisions (.3389) and the BSRI femininity scale and the husband dominant decisions (r=-.2783) are all sig- nificant (p.<.001) and in the predicted direction. However, as with the male subjects, none of the correlations between sex-role identity and purchase decision responsibility (r=.0258 for masculinity with the wife dominant decisions; r=.0574 for masculinity with the husband dominant decisions; r=.0566 for femininity with the wife dominant decisions; r=.0437 for femininity with the husband dominant decissions) are significant for the African-American female subjects. Thus, null hypothesis 4-1 (The correlation between sex-role identity and household purchase decision responsibility is the same among African- Ama'ican and Anglo-American subjects) is rejected. The following product moment correlations were calculated between sex- role-attitude and household purchase decision responsibility for both Anglo- American and African-American male and female subjects: (1) masculine activities with the wife dominant decisions; (2) masculine activities with the husband dominant decisions; (3) feminine activities with the wife dominant decisions; and (4) feminine activities with the husband dominant decisions. Correlations for Anglo-American and African-American males are presented in Table 25 while the correlatations for the Anglo-American and the African- American female subjects are shown in Table 26. As illustrated in Table 25, for the Anglo-American males the correlations between the masculine activities and the wife dominant decisions (r=-.5985), the masculine activities and the hus- band dominant decisions (r=.2385), the feminine activities and the wife 101 Table 25: Correlations (Sex-roles with Purchase Decision Responsibility) ANGLO—AMERICAN FEMALES Purchase Decision Responsibility Wife Dominant Husband Dominant . '.'. .‘."'. .‘. ,r_ v, :33; _-,-_ '. .. 4 . .v.. ., .'-.'. ‘5'”, .r.-.'. u, . Vt71~~r r .. . .31 1 . v . v, r- v: -,- {v3 -M810ns , ., . ,v '3, ,'. , ‘. ' 2' p ,3 J), .I. Mlons '.".'.' .'.'. ' a . .'.".'. '.'. Sex “ ~role1den mil .. __ —- v—rYY BSRI masculinity scale - 3647 is”. A 9 if 3389 es“ ' 351311611111“le scale , 2618 "._ 2783" Sex-role ' ' ' """ Masculineactmtles I 1‘ I V -4378** A "if 3635" Feminine activities .4433 " -.3l97 ** AFRICAN-AMERICAN FEMALES Purchase Decision Responsibility Wife Dominant Husband Dominant ‘.It I, S y’.‘.‘,'1 ,’.' j L.’,‘,')'.'.'.‘A'.' ‘.'J'.' ‘ '.'LL'.‘.' (3.2 L'.‘ ‘. . ..' L'.’. ,‘. .’.’.A‘.‘.’ ‘l':: M3. .'.'V'/ .‘.'.‘. .‘-'."‘.A ' '.'. L‘L- ‘. ' ' ' .'.'. .' ' {.13 r'.'u'.' ‘. .' ‘,‘,‘,‘L‘,‘ ‘ . .‘.'.' K‘.‘._'.'.‘W ;’ '.‘ ._ ssh: masculinity scale 0258 .0574 . BSRI femlnlmtvscale 0566 0437 _ ......1... ......z............ ,.-.r.:.1:: ... ...... ........... Mascufineacnvmes . -3163" 3105 u Feminine activities .1412 "”' —.4410 ** Numbers are product moment correlations " denotes r significant at .001 level. "‘ denotes r significant at .01 level. 102 dominant decisions (r=.5756), and the feminine activities and the husband dominant decisions (r=-.2662) are all significant (p.<.001) and in the predicted direction. Similar results can be seen for the African—American males. Here, each correlation (r=-.2961 for masculine activities with wife dominant decisions; r=.1550 for masculine activities with husband dominant decisions; r=.2114 for feminine activities with wife dominant decisions; r=-.1483 for feminine activities with husband dominant decisions), although not as strong as the correlations for Anglo-American males, is significant (p.<.001) and in the predicted direction. As illustrated in Table 26, for the Anglo-American females, the correla- tions between the masculine activities and the wife dominant decisions (r=-.4378), the masculine activities and the husband dominant decisions (r=.3635), the feminine activities and the wife dominant decisions (r=.4433), and the feminine activities and the husband dominant decisions (r=-.3197) are all sig- nificant (p.<.001) and in the predicted direction. Likewise among African- American females, each correlation (r=-.3l63 for masculine activities with the wife dominant decisions; r=.3105 for masculine activities with the husband dominant decisions; r=.l412 for feminine activities with the wife dominant decisions; r=-.4410 for feminine activities with the husband dominant decisions) is significant (p.<.001) and in the predicted direction. Again, although significant, the correlations between the two variables are not as strong for the African-American females as it is for the Anglo-American females. Based on these results, null hypothesis 4-2 (The correlation between sex- role attitude and household purchase decision responsibflity is not the same among Affican-Amefican and Anglo-American subjects) is rejected. 103 RESEARCH QUESTION Finally, some researchers have suggested that among African-Americans, the order of birth of a child may be a more dominant factor in differentiating role behavior than is gender alone. African-American subjects were placed into one of six groups based on their order of birth as indicated by questionnaire responses. The groupings were as follows: group 1 (lst born), n=264; group 2 (2nd born), n=98; group 3 (3rd born), n=162; group 4 (4th born), n=14l; group 5 (5th born or later), n=83; group 6 (only child), n=46. The mean sex- role identity score for subjects in group 1 is 111.5016; the mean score for group 2 is 110.4914; the mean score for group 3 is 110.4944; the mean score for sub- jects in group 4 is 106.1048; subjects in group 5 registered a mean score of 108.7642; and the mean score for group 6 is 109.0552. Although most of these differences are not significant, they are, generally in the predicted direction, with group 1 (lst born) recording the highest score and groups 4 and 5 register- ing the lowest. Age also appears to be significantly related to sex-role identity among African-American subjects, although more among men than among women. The mean sex-role identity score for African-American males age 39 and under is 112.5876 compared to a mean score of 101.8750 for African-American males age 40 and over (p.<.01). For African-American women, the sex-role identity scores are closer, yet still significant (p.<.05); 111.5676 for African-American women age 39 and under and 113.9336 for women age 40 and older. It should be noted, however, that the direction of the female group difference is counter intuitive. Also, the greater difference between the mean scores of the African- American males is mainly due to the fact that the older men scored lower than the younger group on the scale’s feminine items. These differences apparently 104 do not hold across racial lines, however. The mean sex-role identity score for Anglo-American males age 39 and under is 103.8022 while the mean score for Anglo-American males age 40 and older is 102.5918. Similarly, Anglo- American women age 39 and under registered a mean sex-role identity score of 106.6796 while the mean score for Anglo-American women age 40 and older was 107.8595. Neither of these differences is significant at either the .01 or the .05 level. Although more research needs to be conducted on this topic, data presented here suggests that age may be more of a factor in sex-role differentia- tion among African-Americans than among Anglo-Americans. CHAPTER V CONCLUSION GENERAL DISCUSSION A major thesis presented in the present study is that race is a major determinant of differences in sex-role attitudes and identities. In accordance with this thesis, a theoretical framework was offered to (I) examine among African-Americans and Anglo-Americans the nature and prevalence of sex-role differentiation based upon inherent personality characterizations for men and women; (2) compare the attitudes of African-Americans and Anglo-Americans toward equality between the sexes; and (3) test whether flexible attitudes toward a variety of purchase decisions are related to non-traditional sex-role attitudes, and whether flexibility in purchase decisions is associated with a more flexible definition of oneself in terms of sex-role norms. As predicted, Anglo-American subjects of both sexes who defined their identity in terms of traditional sex-role norms (sex-typed identity) also displayed a tendency to approve of the traditional sex stereotypical standards of behavior for the sexes (traditional sex-role attitudes). The data reveal that these subjects were more likely to evaluate an activity or task in terms of traditional stereotypes which def me the sex appropriateness or inappropriateness of the ac- tivity or task. As such, this lends further support to Bem’s "Gender Schema Meaty" which sees sex-typed individuals as having a general tendency to parti- tion the world into masculine and feminine categories, and to encode and or- ganize information (about themselves and others) in terms of the culture’s definitions of masculinity and femininity. 105 106 Anglo subjects with sex-typed identities and traditional sex-role attitudes were also the most likely to assume responsibility for purchase decisions in ac- cordance with traditional gender based norms which define the sex appropriate- ness or inappropriateness of the individual purchase. This, however, was not the case among androgynous subjects of either sex. Unlike the sex- typed subjects, androgynous subjects showed significantly less of a tendency to engage in sex-role stereotyping. As indicated by Bern, it would appear that these individuals do not evaluate people or situations on the basis of traditional standards of appropriate or inappropriate behavior for men and women. The data also show that both androgynous males and females, in general, are more flexible in assuming responsibility in traditionally defined gender-based household purchase decisions than are sex-typed males and females. This was more the case with the traditionally defined husband dominant purchase decisions, however. The traditionally defined wife dominant decisions regarding "shopping for groceries, " purchasing bedroom fumiture, " and "big/ing clothing for the children, " remain much more within the domain of the wife than the husband, regardless of the sex-role orientation of the wife. Also as predicted, results of this study offer support for the assumption that the sex-role identities of African-American males and females are more androgynous than the sex-role identities of Anglo-American males and females. Data presented in chapter 4 clearly show a significant difference between the sex-role identity scores of African-American and Anglo-American subjects. The most "androgynous" group was the African-American female group (x2113), with Anglo-American males displaying the lowest sex-role identity score (i=102.98, least androgynous group). Sex-role identity scores for the African- American males and the Anglo-American females were nearly equal (x=107 .3730 and 107.3576, respectively). That African-American women and men displayed 107 higher sex-role identity scores than their Anglo-American counterparts is consis- tent with previous research on race and culture (e.g., Erikson, 1968; Ladner, 1972; Lewis, 1975; Kochman, 1981). For example, research indicates that both African-American women as well as African-American men are viewed as asser- tive, aggressive, independent, and self-confident (Lewis, 1975). Moreover, according to Ladner (1972) young African-American women are taught to be strong, and prepared to take on the responsibility of supporting as well as rais- ing their families. It is not surprising, therefore, that most of the variation be- tween the sex-role identity scores of African-American women and Anglo-American women is accounted for by the tendency of African-American women to score significantly higher than Anglo women on the BSRI masculine traits (masculinity score = 54.6884 for African-American women; masculinity score = 48.7214 for Anglo-American women). This does not mean, however, that African-American women are less feminine than Anglo-American women. Scores for the two groups on the BSRI feminine items were about equal (femininity score = 58.3209 for African-American women; femininity score = 58.6362 for Anglo-American women), indicating that African-American women are no more nor less feminine than Anglo-American women. Interestingly, some studies (see e.g., Petigrew, 1964) have chosen to em- phasize the more feminine nature of African-American males as compared to their Anglo-American counterparts. However, data presented here indicate that African-American men are no more feminine than are Anglo-American males. The mean femininity scores for the African-American male (52.5238) and the Anglo-American male (51.3348) subject groups were not significantly different. African-American males did, however, score significantly higher than Anglo- American males on the feminine items, eager to soothe hurt feelings, loves children, sensitive to the needs of others and sympa thetic. These differences are 108 in accordance with the race and culture literature in that studies have found that both African-American men and African-American women place a high value on personal relationships; are nurturant and highly interactive physically with children; are expressive emotionally; and are adept at handling interper- sonal relationships. Still, the greatest discrepancy in sex-role identity scores among the African-American and the Anglo-American males is men in each group’s response to the BSRI masculine items. Here, African-American males scored significantly higher than Anglo-American males (masculinity = 54.6452 for African-American males; masculinity = 51.6452 for Anglo-American males, p.<.01). The only masculine items that did not produce significant differences between the two groups were the items ambitious; independent, and individualis- tic. Logically, one might expect Anglo-American males to score exceptionally high on these traits since they have traditionally been used to represent the very essence of masculinity in Anglo-American society. The mean differences between the groups on each of these items was, nonetheless, in the direction of the African-American males. Moreover, some researchers (e.g., Harris, 1991; Millham and Smith, 1981) have questioned whether or not the concept of androgyny has any concep- tual meaning when applied to the African-American population. Research indi- cates that to the extent to which sex-role differences do exist among African-Americans, they do not include widespread differentiation on the basis of personal qualities and characteristics. Data presented in this study would cer- tainly support this thesis. Furthermore, since androgyny appears to have little or no conceptual meaning among African-Americans, sex role differentiation as a means of predicting sex related behavior among the African-American popula- tion is tenuous. Such was the case in this study when an effort was made to re- late sex-role identity to purchase decision responsibility among 109 African-American subjects. Here, it was found that African-American subjects were less likely than Anglo-American subjects to assume responsibility in pur- chase decisions based on traditional norms governing the sex appropriateness of the decision. It appears that among African-Americans, the influence to make a purchase decision may be determined less by the role position of the in- dividual (husband/wife) and more by other, as yet unexplored, factors. As predicted, both sex-role identity and sex-role attitude were significant- ly correlated with purchase decision responsibility among Anglo-American sub- jects. However, only sex-role attitude was significantly correlated with purchase decision responsibility among the African-American subjects. Sex-role attitudes were, however, much better at predicting purchase decision responsibility for gender ”inappropriate" rather than gender "appropriate" decisions among African-American subjects. For example, knowing that an African-American wife is less traditional (more flexible) in her sex-role attitudes would enable us to predict that she would also be likely to assume greater responsibility in tradi- tionally defined husband dominant purchase decisions. The same knowledge, however, would not enable us to predict that she would necessarily assume less responsibility in purchase decisions traditionally defined as wife dominant. The same holds for African-American husbands as well. This does not appear to be the result of a flaw within the theoretical model, however, since the theoretical framework enables us to make predictions about both gender "inap- propriate" as well as gender "appropriate" decisions among Anglo-Americans. In other words, knowing that an Anglo-American wife is less traditional in her sex-role attitudes enables us to (I) predict that she will be likely to assume greater responsibility in traditionally defined husband dominant purchase decisions, and (2) that she will be likely to assume less responsibility in tradi- tionally defined wife dominant purchase decisions. This, again, reinforces the 110 assumption that sex roles, in general, are much less salient in African-American culture than they are within Anglo-American culture. Again, it may be that fac- tors unrelated to sex roles have a more significant effect on purchase decision responsibility among African-Americans than do sex roles alone. Although they produced only limited effects in this study, variables such as age and order of birth may yet prove to have some usefulness in predicting purchase related behavior among African-Americans. The present research findings as well as other information on African- American sex-role orientation raises an important question. If behavioral expec- tations in African-American culture are comparable for males and females, and if birth order rather gender alone is more important in differentiating behavior, then what factor does contribute to a child’s sexual identity? Lewis suggests that in a society where there is discontinuity between the asexual child and the sexual adult (as in Anglo-American society), the only way a child can assure his/her masculinity/femininity is by displaying traits defined as masculine or feminine. It is, therefore, important that these traits lack ambiguity so that the child is assured of his/her proper sexual identification. In African-American cul- ture, on the other hand, children are considered sexual rather than asexual beings. They learn about sex, menstruation and birth while very young, so that there is continuity between their lives as children and as adults. For the African-American child his sex identity is primarily tied to his definition of him- self as a sexual being, rather than to behavior which has arbitrarily been defined as masculine. As such, in a culture where independence, nurturing, assertiveness, etc., do not distinguish between males and females, and where sexuality is an expected attribute of the person from childhood on, a boy under- stands that he is a male on the basis of his sexuality and success at seduction 111 and a girl realizes her femaleness on the basis of her sexuality and her ability to procreate. This study also reveals several other interesting aspects in relation to ra- cial differences in sex-role attitudes. First, Anglo-American males were found to be more traditional in their sex-role attitudes than any of the other race/sex groups under study. Anglo-American male subjects showed a significantly greater tendency to engage in sex-role stereotyping (x=47.9424) than did either the African-American male subjects 6:50.791), the African-American female subjects (x=53.5442), or the Anglo-American female subjects (x=52.1227). Such findings are consistent with research studies from previous decades (e. g., Crovitz and Steinmann, 1980) which indicate that Anglo-American males have been _ slow to accept women in non-traditional roles. Although the social movements of the 19603 and 19703 may have had a profound impact on the ways in which Americans, in general, view equality between the sexes, results presented here in- dicate that Anglo-American men, in 1991, continue to be fairly reluctant to ac- cept women in non-traditional roles. Secondly, some previous studies have found African-American women to be more traditional in their sex-role attitudes than Anglo-American women. Data from this study, however, presents evidence against this assumption. l-lere, African-American and Anglo-American women, on the average, were both fairly flexible in their sex-role attitudes. The only tasks about which African-American women expressed more traditional attitudes than Anglo- American women were the activities playing football and commanding soldiers on the battlefield; and in actuality each of the race/sex groups showed a tenden- cy to stereotype these activities on the basis of sex. Moreover, that African- American women, in general, would be somewhat flexible in their sex-role attitudes is in accord with much of the race and culture literature. Here, 112 studies have emphasized how independent, assertive, competent, and achieving African-American women have needed to be historically and how those charac- teristics conflict with the traditional female role (e. g., Dill, 1975; Harrison, 1974; Ladner, 1971). For example, as Lewis (1975) points out, the impact of colonialism has a differential effect on role expectations in the oppressed group (African-Americans) such that girls are reared to fulfill adult roles of respon- sibility, while boys lack the necessary training to fill mainstream adult roles; that is to become providers, achievement oriented, etc. Under this system, women continue their traditional functions of childrearing and household care but also are forced to assume the role of provider as well, in order to compen- sate for the ineptitude of the male. Thus, it would seem logical to resolve any cognitive dissonance arising from such conflict by espousing non-traditional norms for the female role, since it could not be resolved by adopting tradition- al norms. STRENGTHS AND LIMITATIONS OF STUDY Although interpretation of null findings must always be tentative, the results of this study illustrate the important influence that sex-role orientation has on household purchase decision responsibility. Unlike most previous studies on sex-role orientation which have generally used relatively small samples selected from student populations, the present study uses a larger sample that is more representative of the populations under study. F urther- more, even national surveys of sex-role orientations have often not included suf- ficient numbers of African-American respondents for comparative analyses. A special effort was made in this study, however, to ensure an ample number of 113 both African-American and Anglo-American subjects for constructive com- parisons. The research design did contain three limitations which may have af- fected the results, however. First, the use of a single spouse to describe the relative influence of husbands and wives in purchase decisions has been the sub- ject of some debate (e.g., Blood and Wolfe, 1960; Davis, 1970). Husbands’ and wives’ perceptions of decision making authority have not always been perfectly congruent. The data in this study present the perceptions of only one spouse and may not be completely accurate with respect to actual family decision making patterns. However, since the purpose of this study was merely to describe the relative influence of husbands and wives in various purchase decisions, interviewing only one spouse should be sufficient (see e.g., Davis, 1970). A second limitation of the study involves the possibility that in many cases decisions concerning the products and services used in the study had not been made in the recent past. Thus, a portion of the sample was probably responding hypothetically with regards to some of the product decisions rather than on the basis of recent experience. Furthermore, the categories of husband and wife included, not only married individuals, but people who were divorced, separated, and widowed as well. One could argue that subjects in the later three categories should not have been included in the study. But again, the ob- jective of the present study was not to examine specifically the purchase decision making process within traditional and non-traditional families, but to investigate the relationship between an individual’s own sex-role orientation and his/her responsibility in various household purchase decisions. 114 ADVERTISING\MARKETING IMPLICATIONS Results from this study indicate that marketers may do well to invest more time and effort into exploring the effects of sex-role orientation on con- sumer decision making. Specifically, the following interrelated research findings should be considered. First, traditionally defined wife dominant decisions regarding shopping for groceries, purchasing bedroom furniture, and buying clothing for the children, remain much more within the domain of the wife than the husband, irrespective of the sex-role orientations of the spouses. Marketers should, there- fore, be advised to continue directing the majority of advertising for items in these product categories to the female spouse. This is not to suggest that hus- bands (especially androgynous husbands) do not assume any responsibility in purchase decisions related to these areas. Rather, that wives, no matter what their sex-role orientation, still remain the primary decision makers for items in these product categories. This coupled with the finding that wives with androgynous sex-role iden- tities and flexible sex-role attitudes often assume a greater amount of respon- sibility in traditionally defined husband dominant decisions may, in fact, support the 'lsuperwoman" (all things to all people) image that has been the focus of much advertising throughout the 19808. Although advertisers have often been criticized for depicting women in this manner, the present study sug- gests that women today may, indeed, be performing a wider range of activities than ever before. Although this may be true, whether women today wish to be portrayed as superwomen is a question that marketers will need to address in future research. For now, one solution is simply to continue portraying women performing a number of activities in a wide variety of situations. 115 Data presented here also seems to suggest that the "sensitive guy of the eighties" masculine image may have been more of an apparition that a reality. For example, although the androgynous male has acquired many of the traits (e. g., affection, compassion, sensitivity) traditionally confined to the female gender, he has not relinquished many of the masculine traits (e. g., ambition, competitiveness, dominance) that combined to produce the contentious macho (Marlboro Man '9 image of previous decades; an image that many had as- sumed was very much on the decline. Marketers may do well, to realize that the "macho man" is still very much alive and well. The demeanor of the macho man of the 19903 has, however, been tempered by the addition of some feminine characteristics that have hopefully produced a more emotionally balanced and secure individual. As with women, therefore, creating a wider variety of characterizations for men should also be the direction of future adver- tising. As such, advertisers might do well to allow the situation, as well as the product, to determine the role behavior of the man or woman depicted in the advertisement. Perhaps the most conspicuous marketing implications of this study are those relating to the amount of responsibility assumed by African-American hus- bands in traditionally defined wife dominant purchase decisions. This is espe- cially true for decisions regarding the purchase of children’s clothing. Here, African-American husbands assumed almost twice as much responsibility as Anglo-American husbands. As such, advertisers and marketers might do well to emphasize the extensive role that many African-American husbands play in the purchase of children’s clothing. Perhaps advertisements directed to both the husband and the wife which depict both partners engaged in the purchase of these products will prove more effective than ads aimed at the African- American wife alone. 116 RESEARCH IMPLICATIONS A number of questions have yet to be answered in relation to sex roles and purchase decision responsibility within the African-American community, and there are issues that require additional examination and testing through fu- ture research. For example, some marketing researchers (e. g., Sexton, 1972; Bauer and Scott, 1970), have argued that many behavioral differences among African-Americans and Anglo-Americans are based on class rather than on culture or ethnicity. The chief reason for this belief is that the black cultural perspective is more prevalent among African-Americans at a lower socio- economic level than among middle or upper income African-Americans (see e.g., Kochman, 1980; Valentine, 1968). Thus, not fully understanding the ef- fects of the acculturation process, or the "myth" surrounding the black past, many researchers see such differences as indicative of class rather than cultural effects. But, as Valentine points out, this is only to recognize that "ethnic iden- tity and subcultural distinctness of all or many minorities are greatest for group members who are poor" (p. 25). Thus, just as poor first generation Irish, Italian, Jewish, or Ukrainian groups are likely to be more "ethnic" than their third generation middle-class counterparts, so would poor African-Americans be more "ethnic" than their middle-class counterparts whose social networks, or level of education, has brought them more within the sphere of influence of dominant Anglo cultural norms and values. That African-Americans, even after several generations, should retain their original ethnic patterns and perspec- tives simply speaks to the extent to which racial segregation has kept the African-American community culturally insular (Kochman, p. 14). This is not to deny that lower income African-Americans will have pat- terns of behavior (consumer as well as other patterns) that are class derived or 117 related. It is simply to say, as Kochman suggests, that "ifa pattern or perspec- ti we is to be called class related, it must be one that arises from a class or colonial situation " (p.14). Kochman provides the following example of a be- havior (consumer) pattern that results from a class rather than a cultural perspective. As a class assignment, Kochman often asked students to go into more expensive department stores in the Chicago area and pretend to buy a product in order to investigate spwch patterns of store employees. In doing so, the students often found themselves receiving a great deal of personal atten- tion from the sales clerks. Most of the African-American and Latino students attributed part of the reason for such attentiveness to the employees’ assump- tion that they were going to steal something. This assumption never occurred to the white middle-class students, who saw such personal attention as simply reflecting the kind of service one gets at the better stores and, of course, the eagerness of the sales clerk to make a sale. The opposing views of the black and the Latino students on the one hand and the white students on the other, are class-related conceptions; African-Americans and Hispanic-Americans have both learned through decades of colonialism that they are "never above suspicion" in the eyes of the colonizer (Anglo-Americans). Thus, class con- sciousness not only affects the way individual’s define their own behavior, but the ways in which they define the behavior of others as well (e. g., how one defines ”friendliness"). white shoppers in this example may have seen the sales clerks as being very friendly, while black shoppers viewed such behavior as menacing. That this perception is class-related is further illustrated by the fact that feeling as if you are never above suspicion is a prevailing view held by newly arrived white ethnic groups as well. However, as these groups (Irish, Italian, Jewish, German, etc.) assimilate into American society (as they become 118 middle-class) their class-related conceptions change to those of mainstream Americans. Such class-related concepts have important implications in the area of African-American shopping behavior. For example, Sexton (1972) studied black shoppers and found that they tend to be more concerned with the store’s ambience (whether or not they were treated with respect by store employees and management) than were white shoppers. The white shoppers were much more concerned with the quality of products than with the personality of the employees. Sexton rightly attributed such differences to class-related variables (i.e., income, education, mobility, etc.), but he failed to understand fully the ef- fect of class consciousness on the shopping behavior of African-Americans. Therefore, he may have wrongly assumed that the chief reason black shoppers placed great emphasis on "store friendliness" was because of the shoppers’ lack of mobility, due to low income, which denied them the privilege of comparison shopping outside their community. And since presumably, all stores within the neighborhood were of equal stature, quality, or the lack of quality, became less of an issue in store differentiation. But the Kochman example suggests that even when mobility due to poverty is not a factor, African-Americans will still place great emphasis on their perceptions regarding how well they are treated by store employees. Classifying the African-American sex-role orientation patterns presented in this study as culture rather than class related is simplified by the study’s re- search methodology. Since the socio-economic and demographic characteristics of both the African-American and the Anglo-American subjects are relatively the same -- those of the middle-class -- behavior differences between the two groups cannot be attributed to differences in class. Moreover, I contend that since poor African-Americans have been found to be more "ethnic" in their 119 behavior than their middle-class counterparts; and since the African-American subjects who participated in this study were primarily middle-class; we can fur- ther assume that the cultural patterns related to sex-role orientation presented in this study would be even more prevalent among African-Americans at a lower socio-economic level. Finally, if marketers are to better understand the African-American con- sumer market it is paramount that they attempt to more fully understand the cultural as well as the sociological factors affecting Black Americans. In rela- tion to African-American sex-role orientation, researchers must begin to distinguish between adaptive patterns which would stem directly from the situa- tion of oppression (class-related) on the one hand, and the unique cultural pat- terns which define early childhood training and which have their roots in African tradition, on the other. Lewis (1975) refers to this as a distinction be- tween micro—structural (internal cultural practices) and macro-structural (pres- sures from the wider society) factors. Both of these factors influence the behavior of the individual and in particular his/her role expectations as an adult. Further, it is probable that in early childhood socialization, when be- havior patterns are probably most unconscious, and at an age when the child is minimally influenced by dominant societal expectations, there is greater cultural influence; while in later socialization, particularly as the child reaches puberty, socialization reflects more closely the structure of expectations and opportunities provided for African-American men and women by the dominant society. At this later stage black socialization is adaptive to macrostructural constraints. Thus, an attempt to analyze African-American family life as an expression of a unique black culture cannot ignore the forces of oppression and racism which also affect family life. Although, an investigation of such magnitude exceeds the scope of the present study, I believe that the material presented here does 120 provide a sufficiently clear delineation of culture based sex role patterns to enable researchers to begin to investigate more fully their effects on a wider range of consumer behavior variables. LIST OF REFERENCES LIST OF REFERENCES Allen, H. (1981) "Black Media’s Low Profile Poses a Problem," Advertising Age, Oct. 19. Almquist, E.M. ( 1975) "Untangling the Effects of Race and Sex: The Disadvantaged Status of Black Women," Social Science Quarterly, 56, 129-142. Antill, J .K. and Cunningham, J .D. (1980) "The Relationship of Masculinity, Femininity, and Androgyny to Self-Esteem," Australian Journal ofPsychol- ogy, 32, 195-207. Antill, J .K., Cunningham, J .D., Russell, G. and Thompson, N.L. (1981) "An Australian Sex-Role Scale," Australian Journal of Psychology, 33, 169-183. Babbie, E. (1989) The Practice of Social Research, 5th ed., Belmont, Ca.: Wadsworth Publishing Co. Bandura, A. (1969) Principles ofBeha vior Modification. New York: Holt, Rinehart and Winston. Bandura, A. (1977) Social Learning Theory. Englewood Cliffs, N.J.: Prentice-Hall. Baratz, J. and Baratz, S. (1972) "Black Culture on Black Terms: A Rejection of the Social Pathology Model," in T. Kochman (ed.) Rappin ’and Stylin’ Out. Communication in Urban Black America. Urbana, Il.: University of Illinois Press. Barry, T.E., Gilly, MC. and Doran, LE. (1985) "Advertising to Women with Dif- ferent Career Orientations," Journal of A dvertising Research, 25, 26-34. Bartos, R. (1980) Yhe Moving Target. New York: The Free Press. Bauer, RA. and Cunningham, SM. (1970) "The Negro Market," Journal ofAdver- tisingResearch, 10, 3-13. Bayer, A.E. (1975) "Sexists Students in American Colleges: A Descriptive Note," Journal ofMam'age and the Family, 37, 391-397. 121 122 Bechtoldt, HP. (1959) "Construct Validity: A Critique,” American Psychologists, 14, 619-629. Bern, S. L. (1974) "The Measurement of Psychological Androgyny," Journal ofCon- sultrng and Clinical Psychology, 42, 155- 162. Bern, S.L. (1975) "Sex Role Adaptability: One Consequence of Psychological Androgyny," Journal of Personality and Social Psychology, 31, 4, 634—643. Bern, S.L. (1977) "On The Utility of Alternative Procedures for Assessing Psychological Androgyny," Journal of Consulting and Clinical Psychology, 45, 196-205. Bern, S.L. (1979) ”Theory and Measurement of Androgyny: A Reply to the Ped- hazur-Tetenbaum and Locksley-Colten Critiques," Journal of Personality and Social Psychology, 37, 1047-1054. Bern, S.L. (1981a) Bem Sex Role Inventory: Professional Manual. Palo Alto, Ca.: Consulting Psychologists Press. Bern, S.L. (1981b) "Gender Schema Theory: A Cognitive Account of Sex-typing," Psychological Review, 88, 354-364. Bern, S.L. (1981c) "The BSRI and Gender Schema Theory: A Reply to Spence and Helmreich," Psychological Review, 88, 369-371. Bern, S.L. (1982) ”Gender Schema Theory and Self-Schema Theory Compared: A Comment on Markus, Crane, Bernstein, and Siladi’s Self-Schemas and Gender," Journal ofPersonality and Social Psychology, 43, 1192-1 194. Bern, S.L. and Lenney, E. (1976) "Sex Typing and the Avoidance of Cross-Sex Be- havior," Journal of Personality and Social Psychology, 33, 1, 48-54. Bennett, L. (1964) The Negro Mood. Chicago: University of Chicago Press. Berkman, H.W. and Gilson, C. (1986) ConsumerBeha vior: Concepts and Strategies, 3rd Ed., Boston: Kent Publishing. Berzins, J .I., Welling, M.A., and Wetter, RE. (1978) ”A New Measure of Psychological Androgyny Based on the Personality Research Form," Journal of Clinical and Consulting Psychology, 46, 126- 1 38. Bettrnan, J .R., Capon, N. and Lutz, R.J. (1975) "Multiattribute Measurement Models and Multiattribute Theory: A Test of Construct Validity," Journal of ConsumerResearch, 1, 1-15. 123 Blood, R.O. and Wolfe, D.M. (1960) Husbands and Wives: The Dynamics ofMar- ried Living. Glencoe, 11.: The Free Press. Bohmstedt, G.W. (1969) "A Quick Method for Determining the Reliability and Validity of Multiple-Item Scales," American Sociological Review, 34, 542-548. Bronfenbrenner, U. (1961) "The Changing American Child - A Speculative Analysis," Merrill Palmer Quarterly, 7, 73-84. Broverman, I.K., Broverman, D., Clarkson, F., Rosenkrantz, PS, and Vogel, SR. (1970) "Sex-Role Stereotypes and Clinical Judgements of Mental Healt ," Journal of Consulting and Clinical Psychology, 34, 1-7. Bullock, H. (1961) "Consumer Motivations in Black and White," Harvard Business Review, 39, 89-104. Cann, A. and Siegfield, W.D. (1987) "Sex Stereotypes and the Leadership Role," Sex Roles, 17, 401-408. Cano, L., Solomon, 8., and Holmes, BS. (1984) "Fear and Success: The Influence of Sex, Sex-Role Identity, and Components of Masculinity," Sex Roles, 10, 341- 346. Carter, D.B., Ed. (1987) Current Conceptions of Sex Roles and Sex Typing: Theory and Research. New York: Praeger. Churchill, G.A. Jr. (1979) "A Paradigm for Developing Better Measures of Market- ing Constructs, " Journal of Marketing Research, 16, 64-73. Cook, ER (1985) Psychological Androgyn y. New York: Pergamon Press. Constantinople, A. (1973) "Masculinity-Femininity: An Exception to a Famous Dic- tum?" Psychological Bulletin, 80, 389-407. Crane, M. and Markus, H. (1982) "Gender Identity: The Benefits of a Self-Schema Approach, " Journal of Personality and Social Psychology, 43, 6, 1 195-1 197. Crovitz, E. and Steinmann, A. (1980) "A Decade Later: Black-White Attitudes Toward Women’s Familiar Role," Psychology of Women Quarterly, 5, 2, 170- 176. Cunningham, I.C.M. and Green, RT. (1974) "Purchasing Roles in the US. Family: 1955-1973," Journal of Marketing, 38, 61-64. Davis, H.L. (1970) "Dimensions of Marital Roles in Consumer Decision Making," Journal of Marketing Research, 7, 168-17 7. 124 Davis, H.L. ( 1971) "Measurement of Husband-Wife Influence in Consumer Pur- chase Decision," Journal of Marketing Research, 2, 305-312. Davis, H.L. and Rigaux, RP. (1974) "Perception of Marital Roles in Decision Processes," Journal of Consumer Research, 1, 51-62. Day, B. (1974) Sexual Life Between Blacks and Whites. London: Collins. Deitrich, F .H. and Kearns, T.J. (1989) Basic Statistics: An Inferential Approach, 3rd ed. Riverside, N.J.: Dellen/Macmillan. DuBois, (1924) The Git? of Black Folks: The Negro in the Maldng of America. Reprint, New York: Washington Square Press (1970). Edwards, AL and Ashworth, CD. (1977) "A Replication Study of Item Selection for the Bern Sex Role Inventory," Applied Psychological Measurement, 1, 501-507. Edwards, K.J. and Norcross, EN. (1980) "A Comparison of Two Sex-Role Androgyny Measures in a Study of Sex-Role Identity for Incarcerated Delin- quent and Nondelinquent F emales," Sex Roles, 6, 859-870. Epstein, CF. (1973) "Positive Effects of the Multiple Negative: Explaining the Suc- cess of Black Professionals," American Journal of Sociology, 78, 912-935. Erkison, EH. (1968) Identity: Youth and Crisis. New York: Norton. Fagot, B. (1978) "The Influence of Sex of Child on Parental Reactions to Toddler Children," Child Development, 49, 2, 459-465. Feldman, S.S., Biringer, Z.C. and Nash, SC. (1981) "Fluctuations of Sex-Related Self-Attributions as a Function of Stage of Life Cycle," Developmental Psychology, 17, 24-35. Fry, J. (1971) "Personality Variables and Cigarette Brand Choice," Journal of Marketing Research, 8, 298-304. Gaudreau, P. (1977) "Factor Analysis of the Bern Sex Role Inventory," Journal of Consulting and Clinical Psychology, 45, 299-302. Gayton, W.F., Havu, G.F., Ozmon, KL. and Tavormina, J. (1977) "A Comparison of the Bern Sex Role Inventory and the PFR ANDRO Scale," Journal ofPer- sonality Assessment, 41, 619-621. Gentry, J .W. and Doering, M. (1977) "Masculinity-Femininity Related to Consumer Choice," in B.A. Greenberg and D.N. Bellinger (Eds) Contemporary Market- ing Thought. Chicago: American Marketing Association, 423-427. 125 Gentry, J .W., Doering, M., and O’Brien, T. (1977) "Masculinity and Femininity Factors in Product Perception and Self Image," in Advances in Consumer Research, 5, ed. H.K. Hunt, Ann Arbor, Mi.: Association for Consumer Re- search, 326-332. Gilligan, C. (1982) In A Dillérent Voice: Psychological Theory and Womens Development. Cambridge, Mass: Harvard University Press. Gitter, A., Block, H. and Mostofsky (1972) "Race and Sex in Perception of Emo- tion," Journal of Social Issues, 28. Glass, G.V. and Hopkins, K.D. (1984) Statistical Methods in Education and Psychology, 2nd ed. Englewood Cliffs, N.J.: Prentice Hall. Glazer, N. and Moynihan, DP (1963) Beyond The Melting Pot. Cambridge: MIT Press. Gonzalez, A. (1982) "Sex Roles of the Traditional Mexican Family: A Comparison of Chicano and Anglo Students’ Attitudes," Journal ofCross-Cultural Psychology, 13, 3, 330-339. Gough, H. (1952) "Identifying Psychological Femininity," Educational and Psychological Measurement, 22, 433-455. Gough, H. (1975) Manual for the California Psychological Inventory. Palo Alto, Ca.: Consulting Psychologists Press, Inc. Green, RT. and Cunningham, I.C.M. (1975) "Feminine Role Perceptions and Farni- ly Purchasing Decisions," Journal of Marketing Research, 12, 325-332. Gump, J .P. (1975) "Comparative Analysis of Black Women’s and White Women’s Sex-Role Attitudes," Journal of Consulting and Clinical Psychology, 43, 6, 858-863. Hacker, HM. (1976) "Class and Race Differences in Gender Roles," in L. Duber- man (ed.) Gender and Sex in Society. New York: Praeger Publishing. Hale, J. (1980) "The Black Woman and Child Rearing." In LaFrances Rodgers-Rose (ed.) The Black Woman. Beverly Hills, Ca.: Sage, 79-87. Harrington, D.M. and Andersen, SM. (1981) "Creativity, Masculinity, Femininity, and Three Models of Psychological Androgyny," Journal of Personality and Social Psychology, 41, 744-757. Harris, AC. (1981) "Methodology for Determining the Effects of Acculturation on Mainstream American Blacks." M.A. Thesis. Chicago: University of Illinois at Chicago, Department of Communication and Theatre. 126 Harris, A.C.(l991) "On Determining the Validity of the Bern Sex Role Inventory," Unpublished Manuscript. Harrison, A0. (1974) "T he Dilemma of Growing Up Black and Female," Journal of the Social and Beha vioral Sciences, 20, 28-40. Heilbrun, AB. (1984) "Sex-Based Models of Androgyny: A Further Cognitive Elaboration of Competence Differences," Journal of Personality and Social Psychology, 46, 216-229. Hershey, MR. (1977) "The Politics of Androgyny? Sex Roles and Attitudes Toward Women in Politics," American Politics Quarterly, 5, 261-287. Hershey, MR. (1978) "Racial Differences in Sex-Role Identities and Sex Stereotyp- ing: Evidence Against A Common Assumption," Social Science Quarterly, 58, 4, 583-596. Herskovits, M.J. (1941) The Myth of The Negro Past. Boston: Beacon. Hochschild, AR. (1973) "A Review of Sex Role Research," American Journal of Sociology, 78, 1011-1029.. J affe, L.J. (1990) "The Effect of Positioning on the Purchase Probability of Financial Services Among Women With Varying Sex-Role Identities," Advances in Consumer Research, Vol. 17 . Jaffe, L.J. and Berger, PD. (1988) "Impact on Purchase Intent of Product Position- ing and Sex-Role Identity," Psychology and Marketing, 5, 259-271. Kagan, J. (1964) "Acquisition and Significance of Sex Typing and Sex-Role Iden- tity," in M.L. Hoffman and L.W. Hoffman (eds) Review of Child Develop- ment Research, 2, 137-167, New York: Russell Sage Foundation. Kaplan, AC. and Bean, J .P. (1976) Beyond Sex-Role Sta'eotypes, (eds) Boston: Lit- tle-Brown. Kelly, J .A., F urman, W. and Young, V. (1978) "Problems Associated With The Typological Measurement of Sex Roles and Androgyny," Journal of Consult- ing and Clinical Psychology. 46, 1574-1576. Kelly, J .A. and Worell, J. (1977) "New Formulations of Sex Roles and Androgyny: A Critical Review," Journal of Consulting and Clinical Psychology, 45 , 1101- 1115. Keppel, G. (1982) Design and Analysis: A Researchers Handbook, 2nd ed. Englewood Cliffs, N.J.: Prentice Hall. 127 Kirk, RE. (1968) Experimental Design: Procedures for the Beha vioral Sciences. Bel- mont, Ca.: Brooks/Cole Publishing Co. Kochman, T. (1971) "Cross Cultural Communication: Contrasting Perspectives, Conflicting Sensibilities," The Florida FL Reporter, 9, 3-16, 53—54 Kochman, T. (1981) Black and White Styles in Conflict. Chicago: University of Chicago Press. Kochman, T. (1989) African-American/A nglo Culturally-Based Patterns ofDrTl férence: Trainer’s Guide. Chicago: Kochman Communication Consultants, Ltd. Kohlberg, LA. (1966) "Cognitive Development Analysis of Children’s Sex-Role Concepts and Attitudes," in BE Maccoby (ed.) The Development of Sex Drflerences. Stanford: Stanford University Press. Ladner, J .A. (1972) Tomorrow’s Tomorrow: The Black Woman. Garden City, NY: Doubleday. Leigh, T.W., Rethans, A.J. and Whitney, TR. (1987) "Role Portrayals of Women in Advertising: Cognitive Response and Advertising Effectiveness," Journal of A d vertising Research, 54-63. Lewis, D.K. ( 1975) "The Black Family: Socialization and Sex Roles," Phylon, 36, 221-237. Lewis, H. (1964) Blackways of Kent. Chapple Hill, NC: University of North Carolina Press. Lewis, 0. (1969) "The Culture of Poverty," Scientific American, 215, 4, 19-25. Lubinski, D., Tellegen, A. and Butcher, J .N. (1983) "Masculinity, Femininity, and Androgyny Viewed as Distinct Concepts," Journal of Personality and Social Psychology, 44, 428-439. Maccoby, E. and J acklin, C.N. (1974) The Psychology of Sex Differences. Stanford: Stanford University Press. Mack, D. (1978) "The Power Relationship in Black Families and White Families," in R. Staples (ed.) The Black Family: Essays and Studies. Belmont, Ca.: Wadsworth. Markus, H., Crane, M., Sernstein, S., and Siladi, M. (1982)"Se1f-Schemas and Gender," Journal of Personality and Social Psychology, 42, 1, 38-50. 128 Marsh, H.W., Antill, J .K. and Cunningham, J .D. (1987) "Masculinity, Femininity, and Androgyny: Relations to Self-Esteem and Social Desirability," Journal of Personality, 55, 661-683. Marsh, H.W. and Richards, GE. (1989) "A Test of Bipolar and Androgyny Perspec- tives of Masculinity and Femininity: The Effect of Participation in an Out- ward Bound Program," Journal of Personality, 57, 115-137. Middletown, R. and Putney, S. (1960) "Dominance in Decisions in the Family: Race and Class Differences," in CV. Willie (ed.) The Family Life of Black People. Columbus, Ohio: Charles E. Merrill. Millham, J. and Smith, LE. (1981) "Sex-Role Differentiation Among Black and White Americans: A Comparative Study," The Journal of Black Psychology, 7, 2, 77-90. Mills, C.J. (1983) "Sex-typing and Self-Schemata Effects on Memory and Response Latency," Journal of Personality and Social Psychology, 45, 1, 163-172. Mischel, W. (1966) "A Social Learning View of Sex Differences in Behavior," in BE. Maccoby (ed.) The Development of Sex Differences. Stanford: Stanford University Press. Mischel, W. (1977) Essentials of Psychology. New York: Random House. Moreland, J .R., Gulanick, N., Montague, E.K., and Harren, V.A. ( 1978) "Some Psychometric Properties of the Bem Sex Role Inventory," Applied Psychologi- cal Measurement, 2, 249-256. Morris, G. and Cundiff, E. (1971) "Acceptance by Males of a Feminine Product," Journal of Marketing Research, 7, 372-374. Mussen, PH. (1969) Child Development and Personality, 3rd ed. New York: Har- per and Row. Myers, A.M. and Gonda, G. (1982) "Empirical Validation of the Bern Sex Role In- ventory," Journal ofPersonality and Social Psychology, 43, 304—318. Myrdal, G. (1944) An American Dilemma: The Negro Problem and Modern Democracy. New York: Harper and Brothers. ‘ Nobles, W.W. (1975) "Africanity in Black Families," The Black Scholar, (June). Orlofsky, J .L. (1981) "Relationship Between Sex Role Attitudes and Personality Traits and The Sex Role Behavior Scale-l: A New Measure of Masculine and Feminine Role Behaviors and Interests," Journal of Personality and Social Psychology, 40, 927-940. 129 Orlofsky, J .L., Ramsden, M.W. and Cohen, RS. (1982) "Development of the Revised Sex Role Behavior Scale," Journal of Personality Assessment, 46, 632-638. Osmond, M.W. and Martin, P.Y. (1975) "Sex and Sexism: A Comparison of Male and Female Sex Role Attitudes," Journal of Marriage and the Family, 37, 744-758. Parker, G.V.C. (1969) "Sex Differences in Self-Description on the Adjective Check List," Educational and Psychological Measruement, 29, 99-113. Pedhazur, E.J . (9182) Multiple Regression in Beha vioral Research: Explanation and Prediction, 2nd ed. New York: Holt, Rinehart, Winston. Pedhazur, E.J. and Tetenbaum, T.J. (1979) "Bern Sex Role Inventory: A Theoretical and Methodological Critique," Journal of Personality and Social Psychology, 37, 996-1016. Pettigrew, T.F. (1964) A Profile of the Negro American. Princeton: Van Nostrand Reinhold. Powell, G.N. and Butterfield, DA. (1979) "The ’Good Manager’: Masculine or Androgynous?" Academy of Management Journal, 22, 235-243. Prakask, V. and Flores, RC. (1989) "A Study of Psychological Gender Differences: Applications for Advertising Format," in Advances in Consumer Research, 13, 231-237. Qualls, W.J. (1987) "Household Decision Behavior: The Impact of Husbands’ and Wives’ Sex Role Orientation," Journal of Consumer Research, 14, 264-279. Qualls, W.J. (1984) "Sex Roles, Husband-Wife Influence, and Family Decision Be- havior," in Advances in Consumer Research, Vol. 11, (ed.) Thomas C. Kin- near, Ann Arbor, Mi.: Association of Consumer Research, 270-275. Ratliff, ES. and Conley, J. (1981) "The Structure of Masculinity-Femininity: Multi- dimensionality and Gender Differences, " Social Beha vior and Personality, 9, 41 -47. Rebecca, M., Hefner, R. and Oleshansky, B. (1976) "A Model of Sex Role Transcen- dence," Journal of Social Issues, 32, 3, 197-206. Roopnarine, J .L. (1986) "Mothers’ and Fathers’ Behaviors Toward the Toy Play of Their Infant Sons and Daughters," Sex Roles, 14, 2, 59-68. Roopnarine, J .L. and Johnson, J .E. (1987) Approaches to Early Childhood Educa- u'on, (ed.) Columbus, Ohio: Merrill Publishing Co. 130 Roopnarine, J .L. and Mounts, NS. (1985) "Mother-Child and Father-Child Play," Early Child Development and Care, 20, 2-3. 157-169. Rosen, D. and Granbois, D. (1983) "Determinants of Role Structure in Family Financial Management," Journal of Consumer Research, 10, 253—258. Rosen, RH. (1977) "Sex Role Attitudes of Black and White Women," Internation- al Journal of Women Is Studies, 1, 6, 544-554. Rosenkrantz, P., Vogel, 8., Bee, H., Broverman, I. and Broverman, D.M. (1968) "Sex-Role Stereotypes and Self-Concepts in College Students, Journal ofCon- sulting and Clinical Psychology, 32, 287-295. Ruch, L. (1984) "Dimensionality of the Bern Sex Role Inventory: A Multidimen- sional Analysis," Sex Roles, 10, 99-117. Scanzoni, J. (1975) "Sex Roles, Economic Factors, and Marital Solidarity in Black and White Marriages," Journal of Marriage and the Family, (Feb.) 130-144. Schmitt, B.H., Leclerc, F. and Dube-Rioux (1988) "Sex-Typing and Consumer Be- havior: A Test of Gender Schema Theory," Journal of Consumer Research, 15, 122-128. Sexton, D.E., Jr. (1972) "Black Buyer Behavior," Journal of Marketing, 36, 36039. Sirgy, M.J. (1982) "Self-Concept in Consumer Behavior," Journal of Consumer Re- search, 9, 287-300. Spence, J .T. and Helmreich, R.L. (1979) "On Assessing ’Androgyny’," Sex Roles. 5, 721-738. Spence, J .T. and Helmreich, R.L. and Stapp, J. (1975) "Ratings of Self and Peers on Sex Role Attributes and Their Relation to Self-Esteem and Conceptions of Masculinity and F emininity," Journal of Personality and Social Psychology, 32, 29-39. Smith, PK, and Daglish, L. (1977) ”Sex Differences in Parent and Infant Behavior in the Home," Child Development, 48, 4, 1250-1254. Snow, M.E., Jacklin, C.N. and Maccoby, E. (1983) "Sex of Child Differences in Father-Child Interactions at One year of Age, " Child Development, 54, 227- 232. Stewart, W.A. (1974) "Acculturative Processes and the Language of the American Negro," in W.W. Gage (ed.) Language in its Social Setting. Washington, DC: Anthropological Society of Washington. 131 Truriel, D. (1984) "Sex Role Typing and Ego Identity in Israeli, Oriental, and Western Adolescents," Journal of Personality and Social Psychology, 46, 440- 457. Tunnel], G. (1981) "Sex Role and Cognitive Schemata: Person Perception in Feminine and Androgynous Women," Journal of Personality and Social Psychology, 40, 6, 1126-1136. Valentine, CA. (1968) Culture and Poverty. Chicago: University of Chicago Press. Venkatesh, A. (1980) "Changing Roles of Women-«A Life-Style Analysis," Journal of Consumer Research, 7, 189-197. Venkatesh, A. and Tankersley, CB. (1979) "Magazine Readership by Female Seg- ments," Journal of Advertising Research, 19, 31-38. Wall, K.A. (1979) "Trying to Reach Blacks? Beware of Marketing Myopia, " Adver- tising Age, May 21. Weiers, RM. (1984) Marketing Research. New York: Prentice Hall. Whipple, T.W. and Courtney, A.E. (1985) "Female Role Portrayals in Advertising and Communication Effectiveness: A Review," Journal of A dvertising, 14. Willie, CV. (1981) "Dominance in the Family: Black and White Experience," The Journal of Black Psychology, 7, 2, 91-97. Wilson, RR. and Cook, ER (1984) "Construct Validity of Four Androgyny Instru- ments," Sex Roles, 11, 813-837. Woodson, CB. (1936) The African Background Outlined. New York: New American Library. Worth, L.T., Smith, J. and Mackie, D.M. (1992) "Gender Schematicity and Preference for Gender-Typed Products, " Psychology and Marketing. New York: John Wiley & Sons, Vol. 9(1): 17-30. Young, V.H. (1970) "Family and Childhood in a Southern Negro Community," American Anthropologist, 72, 269-288. APPENDIX iAPPEmUJDK Sample Questionnaire 1. Following is a list of personality characteristics. Please rate each of the characteristics as it applies to you, with 1 meaning "never or almost never true," and 7 meaning "always or almost always true." Never Always Acts as a leader 1 2 3 4 5 6 7 Affectionate 1 2 3 4 5 6 7 Ambitious 1 2 3 4 5 6 7 Assertive 1 2 3 4 5 6 7 Athletic 1 2 3 4 5 6 7 Compassionate 1 2 3 4 5 6 7 Competitive 1 2 3 4 5 6 7 Dominant 1 2 3 4 5 6 7 Eager to soothe hurt feelings 1 2 3 4 5 6 7 Gentle 1 2 3 4 5 6 7 Has leadership abilities 1 2 3 4 5 6 7 Independent 1 2 3 4 5 6 7 Individualistic 1 2 3 4 5 6 7 Loves children 1 2 3 4 5 6 7 Sensitive to the needs of others 1 2 3 4 5 6 7 Sympathetic 1 2 3 4 5 6 7 Tender 1 2 3 4 5 6 7 Understanding 1 2 3 4 5 6 7 Warm 1 2 3 4 5 6 7 Willing to take a stand 1 2 3 4 5 6 7 132 (cont’d) A. 133 Following is a list of products and services along with several decisions that need to be made when purchasing each product or in performing each Please indicate, from 0%-100%, the degree service. of responsibility you have, relative to your spouse, for making each decision. Groceries When to shop How much to spend Where to shop Bedroom Furniture When to buy How much to spend Which style to buy Major Appliances 50% (e.g., air conditioner, furnace, hot water tank) When to buy How much to spend Which brand to buy . Automobile When to buy How much to spend Make and Model Family Savings When to save How much to save How to invest savings Clothing for Children When to buy What styles to buy How much to spend 100% (cont’d) 3. Following is a list of jobs. 134 Please rate your belief regarding the appropriateness of each job for members of either sex, with 1 meaning "much more appropriate for one sex than the other," and 7 meaning "appropriate for either sex." Much more appropriate for one sex than the other Playing football 1 2 Taking care of children 1 2 Cleaning up the house 1 2 Commanding soldiers on the battlefield 1 2 Teaching elementary school 1 2 Making out the family budget 1 2 Repairing Telephone Wires 1 2 Running a government agency 1 2 Making up the budget of a large corporation 1 2 Working as a brain surgeon 1 2 Appropriate for either sex 5 6 7 5 6 7 5 6 7 5 6 7 5 6 7 5 6 7 5 6 7 5 6 7 5 6 7 5 6 7 135 (cont’d) (CIRCLE 93;) 4. Please record your sex: Male Female 0.00.0... 1 5. Please record your race: White .................. ....... 1 Black ......................... 2 Asian/Pacific Islander ........ 3 Other ......................... 4 6. Please record the category that includes your age: Under 30 years of age 30-39 years of age 40-49 years of age 50-59 years of age 60 years of age and over ...... (II-DUMP 7. Please record the category that includes your marital status: Married Widowed Divorced ...................... Separated ..................... Single/never married .......... U'I-hUNH 8. Please record the category that includes your 1990 total income: Less than $5,000 $5,000 to $7,499 $7,500 to $9,999 $30,000 to $32,499 $32,500 to $34,999 $35,000 to $39,999 $10,000 to $12,499 $12,500 to $14,999 $45,000 to $49,999 $15,000 to $17,499 $50,000 to $59,999 $17,500 to $19,999 $60,000 to $74,999 $20,000 to $22,499 $75,000 to $99,999 $22,500 to $24,999 $100,000 to $124,999 $25,000 to $27,499 $125,000 and over $27,500 to $29,999 $40,000 to $44,999 136 (cont’d) 9. Please record the category that includes your highest level of education: Grade School Less than 8 years .......... 1 8 years .................... 2 High School 1-3 years .................. 3 4 years .................... 4 College 1-3 years .................. 5 4 years .................... 6 5-8 years ...... ............ 7 10. Please indicate your occupation: 11. Please indicate your birth order in relation to other siblings. lst born ..... ................ 1 2nd born 0 O O I O O O O O O O O O O O O ..... 2 3rd born .. ...... ... .......... 3 4th born 0 I O O O O I O O O O O O O O O O O O O O 4 5th born or latter ........... 5 Only child ...... ............. 6 IGRN STRTE UNI I.V 111111111121111111011 11111111151111112111111118