Consumer researchers have been examining the impact of gender identity—the degree to which an individual identifies with masculine and feminine personality traits—on various consumer variables for nearly four decades. However, significant gender identity findings in consumer research have been rare, perhaps because of (1) operationalization problems (Palan, Kiecker, and Areni 1999), (2) inappropriate interpretation and application of gender identity to consumer variables (Gould 1996), or (3) blurring gender categories (Firat 1993).
This paper presents a thorough review, grounded in theoretical models of gender identity, of consumer behavior studies in the marketing literature that have examined gender identity. Based on the literature review, the paper evaluates whether gender identity research is still warranted, and proposes specific research questions to guide future research. Terminology in Gender Identity Research Several different terms have been used over the course of gender identity research to signify gender identity. This creates confusion about what is being studied.
This review only includes studies that have specifically examined the degree to which an individual identifies him- or herself with masculine and feminine personality traits. The term “gender identity” is used throughout the paper to refer to this definition, regardless of what it might have been called in previous studies. Theoretical Background of Gender Identity Gender identity is considered to be a two-dimensional model, with masculine traits comprising one dimension, and feminine traits the other. Psychologists believe that varying degrees of these traits coexist within an individual (Gill et al. 1987).
Two instruments, created in the mid-1970s, have dominated gender identity research, the Bem Sex Role Inventory (BSRI) and the Personal Attributes Questionnaire (PAQ). With either instrument, individuals can be categorized according to which set of gendered traits they primarily identify with. There are two competing gender identity theories impacting the conceptualization of consumer research. Bem’s (1981a) gender schema theory posits that individuals acquire and display traits, attitudes, and behaviors consistent with their gender identity, so gender identity is predictive of broad gender-related constructs.
Spence (1984) posits that gender identity is multifactorial with gendered traits constituting but one factor; therefore, gender identity is predictive only of situations in which gendered traits are likely to have impact. If one wants to understand other gender constructs, e. g. , gender role attitudes, a measure specific to that construct would have to be used. Related to these theories is gender salience, the idea that a gender-related self-concept has to be activated in order for gender identity to be meaningful in a particular context.
Gender Identity Studies in Consumer Behavior Thirty-one studies, dating from 1963-2001 are included in the literature review. Gender identity has been significantly linked to several different consumer variables (e. g. , leisure activities, sex-role portrayals, shopping behavior), but biological sex was often significant when gender identity was not. Gender identity was more likely to be significant when studies were carefully conceptualized. Theoretical Foundations of Gender Identity Studies in Consumer Behavior Studies were categorized with respect to three theoretical issues: (1) the consistency of gender identity conceptualization/operationalization with either gender schema theory (GST) or multifactorial gender identity theory (MGIT); (2) the relevance of gendered traits to study; and (3) the ecognition that gender identity is variable, dependent on salience. A majority of the studies was consistent with MGIT; further, studies consistent with MGIT more often produced significant gender identity findings than those consistent with GST. Studies where gender identity was judged to be irrelevant to the other study variables reported biological sex to be a more significant predictor than gender identity. Only three studies recognized that identification of gendered traits could vary in different situations. Enthusiasm or Skepticism for Gender Identity?
Despite the fact that many studies failed to find significant relationships, when studies carefully conceptualize gender identity (i. e. , as consistent with MGIT and as being relevant), the construct does provide meaningful insights into consumer behavior. Consequently, the paper concludes that enthusiasm for gender identity research is warranted. FUTURE RESEARCH DIRECTIONS AND CONCLUSION Even though the review suggests that gender identity research should be continued, there are some issues that future researchers need to consider.
Three research questions are presented and discussed to guide future research: 1. Which gender variable(s) is most relevant to the study under consideration? 2. Are there additional personality traits or subdimensions of instrumental and expressive traits that are associated with gendered responses? 3. What is the relationship between gendered personality traits and gender salience in different contexts? This paper takes the position that future research should recognize gender identity as a multifactorial construct.
Consumer researchers have been measuring only one aspect of gender identity, gendered personality traits, and sometimes inappropriately using this construct to predict other gender variables (Fischer and Arnold 1996). Therefore, researchers need to consider carefully which gender variable is most important to include in any given study; researchers also need to use consistent terminology to avoid conceptual ambiguity. Moreover, it may be advantageous to measure simultaneously multiple facets of gender identity (e. g. , roles, attitudes, traits) for richer analysis.
Researchers also need to address the operationalization problems evident in the BSRI and PAQ, and consider developing new instruments. The use of qualitative methods in this line of inquiry may be very fruitful, especially since gender categories are blurring. Finally, researchers need to better understand how gender is activated and made salient in different contexts. Understanding this question will enable researchers to understand in what situations gendered traits are likely to be influential, and to understand the interplay between gendered traits and gender salience.
It also comes as no surprise that much is known about sex and gender and how they impact buying and consuming activities. Yet there is one gender-related variable, gender identity1, that has both intrigued and perplexed consumer behavior researchers for over four decades. Gender identity, sometimes referred to as an individual’s psychological sex, has been defined as the “fundamental, existential sense of one’s maleness or femaleness” (Spence 1984, p. 83).
Since gender is culturally derived, gender identity is similarly rooted in cultural understandings of what it means to be masculine or feminine (Firat 1991; Lerner 1986). For many years, sex and gender were thought to be inseparable—that is, men were masculine and women were feminine. But what consumer behavior researchers, among others, recognized long ago was that some men were more feminine than masculine while some women were more masculine than feminine. In the postmodern culture in which we now live, this separation of gender from sex is even more apparent.
Untangling the intricate threads of masculinity and femininity began in the 1930s when the first assessment of gender identity was attempted (Terman and Miles 1936). It was not until the 1960s, however, that gender identity made its first appearance in consumer-related studies (Aiken 1963; Vitz and Johnston 1965), and with the emergence of new conceptualizations of gender identity in the mid-1970s (e. g. , Bem 1974; Spence, Helmreich, and Stapp 1975), the study of gender identity in consumer research intensified, continuing into the 1980s and 1990s.
Even when criticism of the most frequently used gender identity measures, the Bem Sex Role Inventory and the Personal Attributes Questionnaire, surfaced in the 1980s (see, e. g. , Gill et al. , 1987; Marsh and Myers 1986), inclusion of gender identity in consumer research continued unabated. However, significant gender identity findings in consumer research have been rare, causing some as early as the mid-80s to suggest that the inclusion of gender identity in consumer research is unproductive and should be abandoned (Roberts 1984).
Several reasons have emerged to explain the lack of significant findings. For example, some have suggested that the operationalization of gender identity has stymied research (Palan, Areni, and Kiecker 1999); contributing to operationalization problems may be the inappropriate use of terms resulting in conceptual ambiguity. Gould (1996) suggested that the inappropriate interpretation/application of gender identity to consumer research variables produced contradictory findings.
In more recent literature, the potential contribution of gender identity research to the understanding of consumer behavior has been questioned since the meanings of masculinity and femininity have blurred (Firat 1993). Given these concerns, the purpose of this paper is to present a thorough review of consumer behavior studies in the marketing literature that have examined gender identity2; it should be noted that the review has a Western (i. e. , North 1 The term “gender identity” was not the original term used to describe the examination of gendered personality traits.
Rather, the term has emerged over the past 40 years of consumer behavior research. As will be discussed later in this review, the psychological literature has used a variety of terms to signify what we now typically label gender identity. Some of these terms include sex-typing (Bem 1974), and sex-role stereotypes (Spence, Helmreich, and Stapp 1975). Terminology has also varied in consumer behavior literature. Some studies have referred to gender identity as sex-role identity (Kahle and Homer 1985); sex-role self-concept (Stern 1988), or sex-role orientation (Gentry and Doering 1979).
Still others have not specifically labeled the concept, but instead refer to measuring masculine and feminine personality traits (Fry 1971; Worth, Smith and Mackie 1992) or socialization traits (Aiken 1963). The tendency in the last decade by consumer behavior researchers has been to use the term gender identity (Fischer and Arnold 1994; Gainer 1993; Kempf, Palan, and Laczniak 1997; Palan, Areni, and Kiecker 1999, 2001). This literature review adopts the current practice in consumer behavior research by using the term “gender identity. Nonetheless, the implications of consistently using terminology and the relationship of terminology choice to gender identity theories are serious and will also be discussed in this review. The scope of this literature review is to examine the usefulness of gender identity as an explanatory variable of consumer behavior in the marketing context.
Therefore, this review has confined itself to examining gender identity studies published only Palan / Gender Identity in Consumer Behavior Research 2 American) bias, since the studies have all been based in Western societies. Despite a previous review of gender identity research in consumer behavior (Stern 1988), an updated review is necessary, not only to include consumer behavior studies that have been published since 1988, but also to examine the merits of gender identity research given the scholarship advances that have been made in this area in the past decade.
The review is grounded in the theoretical models of gender identity with the specific goal of providing a framework to address the question: “To what extent are differences in gender identity useful in explaining meaningful variations among consumers? ” Addressing this question will help clarify whether enthusiasm or skepticism about gender identity research in consumer behavior is warranted, and, if warranted, what direction(s) future research should take. TERMINOLOGY IN GENDER IDENTITY RESEARCH
As with any carefully conceived research project, a literature review requires a clear and unambiguous definition of the subject under analysis. With respect to this particular literature review, this requires not only understanding what gender identity is, but also understanding how it is related to, yet different, from other sex- and gender-related terminology. This is especially important because several different terms, signifying essentially the same thing, have been used over the course of gender identity research in consumer behavior.
For example, sex-role identity (Kahle and Homer 1985), sex-role orientation (Gentry and Doering 1979), and sex-role self-concept (Stern 1988) are all terms referring to the measurement of gendered personality traits in men and women that have been used in consumer research. It was not until the 1990s that the term gender identity saw more consistent use (Fischer and Arnold 1994; Gainer 1993; Kempf, Palan, and Laczniak 1997; Palan, Areni, and Kiecker 1999, 2001). Contributing to the inconsistent use of terminology is the fact that the term “gender” is often treated in both academic discourse and in the media as interchangeable with “sex. Deaux (1985) acknowledged that this debate was still brewing in the field of psychology, though it was becoming more standard to use “sex” to refer to an individual’s biological sex and “gender” to refer to psychological features associated with biological sex that are socially constructed. That is the convention adopted in this literature review. Thus, “sex” refers to an individual’s biological sex, whether one is a woman (female) or a man (male). Most human beings are born as either male or female3 and by about two or three years of age, children become aware of their biological sex (Money and Ehrhardt 1972).
In contrast, “gender is the cultural definition of behavior defined as appropriate to the sexes in a given society at a given time. Gender is a set of cultural roles” (Lerner 1986). About the same time that children become cognizant of their biological sex, they also have an awareness of culturally-derived gender norms; for example, children know positive stereotypes of their own sex and negative stereotypes about the other sex (Kuhn, Nash, and Brucken 1978). Thus, at a very early age, children begin to develop a belief system with respect to cultural gender roles.
According to Spence (1985), “gender is one of the earliest and most central components of the self-concept and serves as an organizing principle through which many experiences and perceptions of self and other are filtered” (p. 64). in marketing-related journals and or conference proceedings. Further, this review only includes studies that have examined gender identity as the degree to which an individual identifies with masculine and/or feminine personality traits. Consequently, studies that have examined gender differences, sex or gender roles, attitudes toward gendered images, etc. may not be included in this literature review if gender identity was not also examined (see pages 3-4 for a more detailed discussion on differentiating gender identity from sex- and gender-related concepts). Because gender identity is strictly defined as identification with personality traits, it is typically associated with empirical measurement; qualitative studies (in marketing-related literature) examining the degree to which an individual identifies him- or herself as masculine or feminine have not been done, or at least were not found for inclusion in this review. Some babies are born as “intersexed. ” These infants have ambiguous genitalia making sex assignment problematic. Even though diagnostic tests are performed to better determine the true biological sex of the infant, the common practice is to pronounce the infant as male or female depending on its physical features, which is a decision distinctly imbued with cultural meaning (Kessler 1990).
3 As children continue to become culturally socialized, they add to their belief system regarding gender, developing cognitive networks of associations to biological sex. One type of association that children learn is the culturally defined personality traits linked to being male (masculine traits) or female (feminine traits). The extent to which an individual identifies or thinks of him- or herself as masculine or feminine is what we now refer to, in consumer behavior research, as gender identity (Fischer and Arnold 1994).
Two other concepts, gender role and gender role attitudes, have at various times been mistaken to be synonymous with gender identity in consumer behavior studies (for a review, see Fischer and Arnold 1994), when, in fact, they are not. Gender role refers to the culturally-derived behaviors and activities associated with masculinity or femininity that individuals choose to adopt. Gender role attitudes refers to an individual’s beliefs about the roles, rights, and responsibilities of men and women.
While gender identity is certainly related to both gender roles and gender role attitudes, it does not necessarily have to be congruent with gender role or gender role attitudes (Deaux 1985; Fischer and Arnold 1994). For example, it would not be unusual for a woman to report a more favorable attitude toward egalitarian gender roles in the home than toward traditional gender roles, but at the same time be engaging in behaviors associated with a traditional feminine gender role; empirical measurement of that same woman’s gender identity might show that she identified equally with masculine and feminine personality traits.
This literature review only includes studies that have specifically examined the degree to which an individual identifies him- or herself with masculine and feminine personality traits, regardless of the term that has been used to signify the concept. However, to avoid confusion, all of the studies will be referred to as examining gender identity. Implications of using this designation will be discussed further in the final sections of the paper.
THEORETICAL BACKGROUND OF GENDER IDENTITY Masculine and feminine personality traits, upon which gender identity is based, are associated with instrumental/agentic and communal/expressive tendencies, respectively (Parsons and Shils 1952), in Western societies. Instrumental/agentic tendencies are defined as “concern with the attainment of goals external to the interaction process” (Gill et al. 1987, p. 379).
Personality traits such as independence, assertiveness, reason, rationality, competitiveness, and focus on individual goals are the hallmarks of masculinity (Cross and Markus 1993; Easlea 1986; Keller 1983; Meyers-Levy 1988; Weinreich-Haste 1986). A proclivity toward communal/expressive tendencies, however, “gives primacy to facilitating the interaction process itself” (Gill et al. 1987, p. 380). Expressiveness involves understanding and dealing with emotions in self and others, although it is not “being emotional”; rather, it concerns ersonality traits focused on being actively interdependent and relational. Understanding, caring, nurturance, responsibility, considerateness, sensitivity, intuition, passion, and focus on communal goals are traits associated with femininity (Cross and Markus 1993; Easlea 1986; Keller 1983; Meyers-Levy 1988; Weinreich-Haste 1986). Early gender identity research hypothesized a single bipolar dimension of masculinity/femininity; that is, masculinity and femininity were opposites on one continuum (Terman and Miles 1936).
Further, gender identity was believed to be correlated with biological sex and constrained by societal stereotypes of appropriate masculine and feminine behaviors (Constantinople 1973). As societal stereotypes changed, however, the assumptions of the unidimensional model were challenged. This led to the development of a two-dimensional gender identity model, in which masculinity and femininity were conceptualized as two separate, orthogonal dimensions, coexisting in varying degrees within an individual (Gill et al. 1987); this conceptualization of masculinity and femininity remains the accepted standard today.
MEASUREMENT OF MASCULINE AND FEMININE PERSONALITY TRAITS Two instruments incorporating the multidimensional conceptualization of masculinity and femininity have dominated gender identity research (Deaux 1985)—the Bem Sex Role Inventory (BSRI) and the Personal Attributes Questionnaire (PAQ). The development of these instruments preceded gender identity theory development, yet, because the researchers associated with gender identity theory were also the same researchers who developed the BSRI and the PAQ
4 masculine and feminine personality traits came to be synonymous (either rightly or wrongly) with measuring gender identity. The BSRI and the PAQ were developed in the mid-1970s as indexes of self-reported instrumental/agentic and communal/expressive traits (Bem 1974; Spence et al. 1975).
Both instruments categorize individuals, using median splits on masculinity and femininity scores, as either being sex-typed (males report themselves as identifying primarily with masculine traits, females report themselves as identifying primarily with feminine traits), cross sex-typed (males report themselves as identifying primarily with feminine traits, females report themselves as identifying primarily with masculine traits), androgynous (either males or females who report themselves as high on both masculine and feminine traits), or undifferentiated (either males or females who report themselves low on both masculine and feminine traits). There are variations on this categorization system; for example, some researchers prefer to categorize individuals as either being masculine schematic (either males or females who report themselves as primarily masculine) or feminine schematic (either males or females who report themselves as primarily feminine) (Markus et al. 1982). The original development of both the BSRI and the PAQ was not based on the theoretical distinctions between masculinity and femininity; instead, reliance was given to college students’ assessments of stereotypically desirable masculine and feminine personality traits (Bem 1974; Spence et al. 1975).
Thus, for the BSRI, Bem (1974) classified 20 traits judged to be significantly more desirable for men than for women (using t-tests) as reflecting masculinity. Similarly, 20 traits were chosen to reflect femininity. The BSRI also contains 20 traits judged to be neutral, i. e. , neither more desirable nor undesirable for men or women. Spence et al. (1975), in a slightly different manner, developed the PAQ by classifying personality traits judged as socially desirable for both sexes but believed to occur to a greater degree in males as masculine, whereas traits deemed to be socially desirable for both sexes but more frequently occurring in females were classified as feminine.
In addition, the PAQ includes a third scale composed of traits deemed to be socially desirable for one sex group, but not the other. Justification of item classification on the PAQ was based by inspecting item content (Spence and Helmreich 1978). Use of the BSRI and PAQ in consumer behavior research has often resulted in nonsignificant findings (see, e. g. , Roberts 1984; Palan et al. 1999). A possible explanation for these findings is the criticism that the masculine and feminine scales of the BSRI and the PAQ are not internally consistent (Collins, Waters, and Waters 1979; Feather 1978; Gaa, Liberman, and Edwards 1979), suggesting the presence of dimensions other than instrumentality and expressiveness.
Indeed, a study conducted by Myers and Gonda (1982) asked almost a thousand subjects to define masculine and feminine with open-ended responses; over 86% of the responses were not represented on the BSRI. This is consistent with research by Palan et al. (1999) that noted the presence of four dimensions in addition to expressiveness and instrumentality in the PAQ. These findings may be related to concerns that the gender identity measurement scales were linked to theory post-hoc, giving rise to the possibility that neither instrument accurately measures instrumental/agentic and communal/expressive traits (Gill et al. 1987; Myers and Gonda 1982; Palan et al. 1999).
For example, although expressiveness is associated with being actively interdependent, the BSRI includes dependent as a femininity trait; likewise, the PAQ includes independent as a masculine trait, yet instrumentality is more about objectivity and manipulation than independence (Gill et al. 1987). GENDER IDENTITY THEORIES As previously stated, gender identity theories were developed and connected to the measurement of gendered personality traits post-hoc. Two competing gender identity theories, gender schema theory and multifactorial gender identity theory, have dominated the psychology literature. Both of these theories, discussed below, have impacted the conceptualization of consumer behavior studies. Gender Schema Theory
Bem’s (1981a) gender schema theory suggests that individuals acquire and display traits, attitudes, and behaviors consistent with their gender identity. Moreover, according to Bem (1981a), gender identity serves as an organizing principle through which individuals process information about themselves and the world around them, although the ability of gender identity to have such an effect varies depending on whether or not an individual is sex-typed. Thus, a sex-typed (or gender schematic) male or female is more likely to be influenced by his or her gender identity than are non-sex-typed men and women (or gender aschematics). A notable aspect of Bem’s work is that her development of the BSRI was the first to treat masculinity and femininity as independent dimensions, thus allowing for an individual to be androgynous, someone who is characterized by high degrees of both masculinity and femininity. Sex typing with the BSRI requires the researcher to calculate an androgyny score, the difference between an individual’s masculinity and femininity scores (Bem 1974).
Although Bem had not yet put forth her gender schema theory when the BSRI was introduced, she later posited that the androgyny score, as a representation of an individual’s total sex role, was diagnostic and predictive of broad gender-related constructs—all kinds of attributes, attitudes, and behaviors that society prescribes for each sex are tied to masculinity/femininity, gender identity, and gender schema (Spence 1985). Using the BSRI as diagnostic of a wide variety of gender-related constructs, however, has been criticized by those who favor the multifactorial gender identity theory. Specifically, many researchers believe that masculine and feminine personality traits have a much narrower diagnostic capability than that purported by Bem, and if used to indicate broader concepts may result in meaningless data. This criticism is revisited when individual consumer research articles are examined later in this literature review. Multifactorial Gender Identity Theory
Contradicting Bem’s theories is the belief that gender-related phenomena are multifactorial in nature with desirable gendered personality traits constituting but one factor (Feather 1984; Spence 1984, 1991; Taylor and Hall 1982). Accordingly, masculinity and femininity are conceptually distinct from gender role expectations, attitudes, preferences, and behaviors, and the use of the BSRI and the PAQ, which measure masculine and feminine traits, should be limited to studies measuring variables relevant to instrumentality and expressiveness. This is consistent with Spence who maintains that neither the PAQ (1984, 1991, 1993) nor the BSRI (1993) is appropriate as an indicator of global gender-related constructs. The nderlying assumption of multifactorial gender identity theory is that gender identity is a combination of gender-related phenomena, associated in varying degrees with each other, such as gender-related attitudes, interests, and role behaviors, and gendered personality traits. Not only is gender identity multifactorial, but each gender-differentiating factor has a different developmental history that varies across individuals because the factors are impacted by multiple variables that are not necessarily gender-related (Spence 1993). Consequently, the specific array of gender-congruent qualities that people display can be quite variable within each sex (Spence 1993), although both men and women do develop gender identities and a sense of belongingness to their sex that is maintained throughout the life span (Green 1974). Not all researchers agree on what specific factors comprise gender identity.
For example, four critical domains of gender-related phenomena have been identified by Spence and Sawin (1985) including: (1) an individual’s global self-concept of his/her masculinity or femininity; (2) gendered personality traits; (3) gender-related interests, role behaviors, and attitudes; and (4) sexual orientation. Ashmore (1990), on the other hand, maintains that the multiple facets of gender identity include two general constructs, sex stereotypes and gender attitudes, and five areas of self-content (i. e. , where the self is the referent): (1) personal-social attributes; (2) social relationships; (3) interests and abilities; (4) symbolic and stylistic behaviors; and (5) biological and physical attributes.
Regardless of which specific facets or factors one associates with the multifactorial concept of gender, it is important to note that gender identity, as defined in multifactorial theory, is a term that encompasses several different aspects of gender, of which gendered personality traits is just one aspect. Thus, the differentiating feature of multifactorial gender identity theory from gender schema theory is that, in multifactorial theory, the measurement of gender identity requires measuring several different factors (i. e. , personality traits, gender attitudes, gender role behaviors, etc. ); if only one factor, such as gendered traits, is measured, then the applicability of that factor is limited to situations where only that one factor is relevant.
Gender schema theory, on the other hand, maintains that it is the measurement of only one facet, masculine and feminine personality traits, that is needed to indicate multiple gender-related concepts. As will be discussed later, many consumer behavior studies have embraced gender schema theory. Others have acknowledged the need to limit the use of gender identity measures to situations in which masculinity and femininity are relevant, consistent with multifactorial gender identity theory.
6 GENDER SALIENCE
Given that the distinction between gender schema theory and multifactorial gender identity theory impacts conceptualized relationships between gendered personality traits, additional gender constructs, and other consumer variables, another important theoretical issue that emerges in gender identity research is gender salience. Indeed, another possible explanation for the plethora of disappointing findings with respect to gender identity in consumer behavior research may be that the extent to which an individual’s gender schema is activated, and, thus, the extent to which individuals engage in gender-related behaviors is likely dependent on how important gender is in a given context. That is, individuals have several different possible schemata that can be activated and influential in a given situation—a working self-concept (Deaux and Major 1987; Markus and Kunda 1986).
While an individual is not likely to change the essence of his/her gender identity, in a situation where a gender-related working self-concept is activated, beliefs and behaviors will be gender-based because of the salience of gender in that situation. Consequently, the gender-relevant characteristics that men and women possess and display and the gender roles they choose to occupy may vary in different contexts (Spence and Sawin 1985). Moreover, the measurement of gender identity may produce different results in a situation where an individual’s gender-related working self-concept has been activated (making gender more salient in the situation) relative to a situation where the same individual’s gender-related working self-concept has not been activated (making gender less salient in the situation) (Gould 1996). Deaux and Major (1987, p. 75-376) suggest that a gender-related working self-concept is likely to be activated when (1) “gender is a central, well-differentiated component of the self-concept,” (2) a gender-related working self-concept has been recently or frequently activated, (3) “immediate situational cues make gender schemata salient” (e. g. , the gender of a word-of-mouth communicator is incongruent with the situation), or (4) another individual’s actions make gender schemata salient (e. g. , an individual overhears a conversation where a man asks a woman for her opinion about a product). Though gender salience research related to consumer behavior has been very limited, significant results have been reported in three studies when situational cues related to gender (e. g. male/female make-up of groups) have been present (Abrams, Thomas, and Hogg 1990; Considine and Gould 1991; Gould and Weil 1991). GENDER IDENTITY STUDIES IN CONSUMER BEHAVIOR In order to understand whether or not further gender identity research in consumer behavior studies is fruitful, it is important to first understand what contributions have or have not been made by previous studies. And, more importantly, it is vital to understand why significant findings have or have not been found. Consequently, after a brief description of gender identity studies in consumer behavior, problems with the theoretical foundations of the studies are examined to identify possible explanations for previous findings.
DESCRIPTION OF GENDER IDENTITY STUDIES Frequency of Studies As can be seen in Table 1, 31 consumer behavior studies4 empirically examining gender identity have been conducted and reported in the past 38 years, beginning in 1963. Other than the two studies conducted in the 1960s and the one study reported to date in the 2000s, each decade has seen an almost equal share of studies, with eight conducted in the 1970s, 11 4 Two studies are included in Table 1 because they are frequently cited in consumer behavior gender identity studies and have been included in a previous literature review (Stern 1988), although they were not published in marketing-oriented journals.
It should be noted that other consumer behavior literature has been published that examines conceptual issues regarding gender identity, including Gould (1996), Prakash and Flores (1984), Roberts (1984), and Stern (1988). Because these articles do not report new findings based on data collection and analysis, they have not been included in this literature review. The absence from this table of any empirical or qualitative gender identity studies, which report new findings, is unintentional.
7 during the 1980s, and nine during the 1990s. The pattern of these studies suggests a steady and persistent interest in gender identity, despite the criticisms of gender identity instruments and despite the often nonsignificant outcomes of gender identity research. TABLE 1 Summary of Consumer Behavior Gender Identity Studies Study Gender Identity Measure Sample Findings Theory* Relevance Variability Aiken 1963 CPI-FE 300 F Significant positive correlation between femininity traits and decoration, interest, and conformity dress “clusters. ” MGIT No No Vitz & Johnston 1965 CPI-FE MMPI 97 F 97 M
Among smokers, significant positive correlation between masculine cigarette brand image and (1) masculine personality traits; and (2) being male. MGIT Yes No Fry 1971 CPI-FE 216 M & F Non-students Feminine males and females prefer cigarette brands with feminine images, although stronger effect when also have high self-confidence. MGIT Yes No Morris & Cundiff 1971 CPI-FE 223 M High feminine/high anxiety males have more unfavorable attitudes toward feminine hair spray product than do low-or-medium feminine males. MGIT Yes No Tucker 1976 PAQ 13 M 11 F Traditional sex roles are blurring relative to gender identification. MGIT Yes No Burns 1977 PAQ 81 F Non-students
Masculinity is significant determining factor in wives’ decision making power. GST Yes No Gentry & Doering 1977 CPI-FE PAQ 100 M 100 F Gender identity is strong predictor of attitudes toward leisure activities, but poor predictor of attitudes about products, brands, and media. Biological sex is better predictor than gender identity for both attitudes and usage differences in all categories. GST Marginal No Gentry, Doering, & O’Brien 1978 CPI-FE PAQ 100 M 100 F Biological sex accounts for more variability than gender identity with respect to perceptions and use of products and leisure activities. GST Marginal No Gentry & Doering 1979 CPI-FE PAQ 00 M 100 F Biological sex is more strongly related to attitudes and usage of leisure activities than is gender identity. GST Marginal No Golden, Allison, & Clee 1979 BSRI (long) 307 M & F Biological sex is significantly related to product sex-typing, while gender identity is not. GST No No Allison, Golden, Mullet, & Coogan 1980 BSRI (long) 307 M & F Biological sex is a better predictor of differences in product sex-typing than is gender identity. GST No No
8 TABLE 1 (CONT) Summary Of Consumer Behavior Gender Identity Studies Martin & Roberts 1983 BSRI (long) 125 M & F Gender identity is significantly related to performance expectations of women entrepreneurs; sex-role attitudes significantly related to expectations of proven individuals regardless of their sex. MGIT Yes No Gentry & Haley 1984 PAQ 86 M 82 F Biological sex may be a better predictor of ad recall and ease of recall than is gender identity. Gender identity differences within-sex more interesting than between-sex differences. GST Marginal No Coughlin & O’Connor 1985 BSRI (long) 420 M & F Non-students
Masculine gender identity explains more differences in purchase intention as a reaction to female role portrayals in ads than does biological sex. MGIT Yes No Kahle & Homer 1985 BSRI (long) 84 M 55 F Biological sex is better predictor of food preferences than is gender identity. GST No No Barak & Stern 1986 BSRI (short)a 614 F Masculinity scale seems to be self-assurance index, which may be more important to consumer behavior than the femininity scale. MGIT Yes No Qualls 1987 BSRI (long) 89 M 89 F BSRI, as a measure of sex-role orientation, is positively related to household influence. GST Yes No Stern, Barak, & Gould 1987 BSRI (short)a SIS 380 M 380 F Non-students
Men identify with the BSRI masculine scale, while women identify with the feminine scale, but both sexes seem to be confused about opposite sex definitions. SIS is highly correlated to biological sex. MGIT NA No Jaffee & Berger 1988 BSRI (short)b 100 F Gender identity is significantly related to preferences for sex role positioning in advertising, but the relationship differs by product category. MGIT Yes No Schmitt, LeClerc, & Dube-Rioux 1988 BSRI (long) 111 M 120 F (total sample for 3 tests) Biological sex is better explanatory variable than gender identity with respect to recall, choice, and memory tasks. GST Yes No Gould & Stern 1989 BSRI (long) SIS 65 M 70 F
Biological sex is better predictor of fashion attitudes than gender identity, but gender identity research may be more important when examining within-sex vs. between-sex differences. GST Marginal Yes Fischer & Arnold 1990 BSRI (long) 299 M & F Non-students Feminine gender identity positively related to involvement in Christmas gift shopping for both men and women; gender role attitudes moderate involvement. MGIT Yes No Gould & Weil 1991 BSRI (long) 59 M 68 F Biological sex is better predictor than gender identity of self-descriptions, feelings, attitudes, and gift choice. Gender identity useful in explaining within-group differences in specific contexts. GST & MGIT Yes Yes9 TABLE 1 (cont) Summary of Consumer Behavior Gender Identity Studies Jaffee 1991 BSRI (short)b 200 F Non-students Modern portrayal of sex roles in advertising improves advertising response to financial services; positioning more essential for high masculine females than for lesser masculine women. GST Yes No Worth, Smith, & Mackie 1992 M/F traitsc 40 M 72 F Preference for gendered images of beer and jeans consistent with gender identity self-ratings. GST Marginal No Gainer 1993 BSRI (short)a 147 M 210 F
Both biological sex and feminine gender identity positively affect attendance at the arts indirectly through involvement; further, feminine gender identity directly affects involvement, while biological sex indirectly affects involvement as a result of childhood experience in the arts. MGIT Yes No Gould & Stern 1993 BSRI (long) SIS 135 M & F Females are more privately gender-conscious than males; androgyny and undifferentiated individuals have higher private gender consciousness. GST NA Yes Fischer & Arnold 1994 BSRI (long) 299 M & F Non-students Both gender identity and gender role attitudes explain more differences in Christmas shopping than does biological sex. MGIT Yes
No Kempf, Palan, & Laczniak 1997 PAQ 105 M & F Masculine gender identity and androgyny positively correlated to general information processing confidence. MGIT Yes No Palan, Areni, & Kiecker 1999 BSRI (short)a PAQ SIS 64 M 51 F Feminine and masculine scales of BSRI and PAQ are not internally consistent. SIS highly correlated to biological sex. MGIT NA No Palan, Areni, & Kiecker 2001 BSRI (short)a 64 M 51 F Masculine males more likely than feminine males to recall gift giving experiences; feminine individuals (both males and females) were person-focused while masculine individuals (both males and females) were object-focused. MGIT Yes No NOTE. Unless otherwise noted, the sampling population for the study was college students. * GST=Gender Schema Theory; MGIT=Multifactorial Gender Identity Theory a) Source for BSRI short scale is Barak and Stern (1986). b) Source for BSRI short scale is Bem (1981b). c) Gender identity is measured with three feminine and three masculine traits; all of the items are included on either the BSRI or the PAQ. Gender Identity Instruments Used Three gender identity instruments have dominated consumer behavior studies in a cyclic pattern—the CPI-FE, the PAQ, and the BSRI. The 1960 and early 1970s studies relied primarily on the California Psychological Inventory—Femininity Scale (CPI-FE) to measure gender identity. The CPI-FE Gough 1952), a subscale of the California Psychological Inventory (Gough 1957), was based on the unidimensional gender identity model, conceptualizing masculinity and femininity as bipolar opposites on the same continuum. One of the early studies (Vitz and Johnston 1965) used a scale from the Minnesota Multiphasic Personality Inventory (MMPI) in addition to the CPI-FE. As the conceptualization of gender identity from a unidimensional model to a two- dimensional model emerged in the mid-1970s, use of gender identity instruments based on the two-dimensional model replaced the CPI-FE.
10 multidimensional gender identity instruments, the PAQ, became the gender identity instrument of choice for consumer researchers from 1971 to 1979, being used in five studies. However, the CPI-FE was used in addition to the PAQ in three of these studies (Gentry and Doering 1977, 1979; Gentry, Doering and O’Brien 1978). Since 1979, the PAQ has only been used in three other consumer behavior studies. By far, however, the gender identity instrument that has been used most frequently in consumer behavior research is the BSRI—since 1979, the BSRI has been used in 20 consumer behavior studies5.
However, complicating the interpretation of results using the BSRI is the fact that several different forms of the BSRI have been used in consumer behavior studies. The original long version of the BSRI, which is comprised of 20 masculine items, 20 feminine items, and 20 neutral items, was used in 13 of the studies. The other studies used one of two shortened versions of the BSRI—Bem’s short scale, consisting of 10 masculine and 10 feminine items (Bem 1981b), was used in three of the studies, while a short scale developed by Stern, Barak, and Gould (1987) was used in two studies. Yet another study, Worth, Smith, and Mackie (1992), used only three masculine and three feminine items, five of which are included on the BSRI, in their study.
Study Findings – CPI-FE The first four studies using the CPI-FE examined the relationship of gender identity to product attitudes/image in three product categories—clothing, cigarettes, and personal hygiene. In three of the four studies, consumers’ beliefs and/or attitudes were found to be consistent with their gender identity (Aiken 1963; Fry 1971; Vitz and Johnston 1965). For example, smokers’ perceptions of cigarette brands as being masculine was significantly correlated to a masculine gender identity (Vitz and Johnston 1965), while femininity was significantly correlated to perceptions of cigarettes as having feminine brand images (Fry 1971).
Only one study (Morris and Cundiff 1971) reported an inconsistency between gender identity and consumer behavior— feminine males, who also reported a high level of anxiety, had unfavorable attitudes toward a feminine hair spray product; the lower their anxiety level, however, the less unfavorable were their attitudes. Study Findings – PAQ The earliest of the PAQ studies classified men and women as being either primarily masculine or primarily feminine (Tucker 1976); based on his results and the changing roles of women, Tucker suggested that traditional gender roles were blurring and would continue to blur. So, not all women would have feminine gender identities, nor would all men have masculine gender identities. This, it was posited, would have implications for consumer behavior. For example, Burns (1977) found evidence of shifting decision making patterns between husbands and wives related to gender identity.
Specifically, he found that highly masculine wives had more influence in purchase decisions in some product areas than did more feminine wives. The three remaining studies were related to the relationship between gender identity and attitudes/usage of products, brands, media, and leisure activities, finding that gender identity was most predictive of consumer participation in leisure activities (Gentry and Doering 1977, 1979; Gentry et al. 1978); gender identity, however, was a poor predictor of attitudes about products, brands, and media. Moreover, more significant relationships were reported when using the CPI-FE than when using the PAQ. Two studies have used the PAQ to examine information processing differences related to gender identity.
Gentry and Haley (1984) examined the relationship between gender schemas and ad recall, while Kempf, Palan, and Laczniak (1997) studied the relationship between gender identity and general information processing confidence (GIPC), finding a positive relationship between masculinity and GIPC. The only other attention the PAQ has received in consumer research is a study comparing gender identity scales used in consumer research (Palan et al. 1999); the findings of this study suggest four dimensions of personality traits (vulnerability, emotionality, composure, and autonomy) in addition to expressiveness (femininity) and instrumentality (masculinity). Study Findings – BSRI
Several different issues have been examined in the BSRI studies, including the relationship between gender identity and (1) consumers’ perceptions of masculinity and femininity in products (Allison et al. 1980; Golden, Allison, and Clee 5 One of the studies categorized with the BSRI studies used a gender identity measure composed of three masculine and three feminine traits, five of which are on the BSRI and one that is included on the PAQ (Worth, Smith, and Mackie 1992). Because the measure is predominantly a BSRI measure, it is classified with these studies for ease in discussion. (2) sex-role attitudes toward women business owners (Martin and Roberts 1983); (3) female role portrayal in advertising (Coughlin and O’Connor 1985; Jaffee 1991; Jaffee and Berger 1988); (4) food/gendered product preference (Kahle and Homer 1985; Worth et al. 1992); (5) information processing (Schmitt, LeClerc, and Dube-Rioux 1988); (6) gift shopping/gift choice/gift exchange (Fischer and Arnold 1990, 1994; Gould and Weil 1991; Palan, Areni, and Kiecker 2001); (7) attendance at the arts (Gainer 1993); and (8) gender consciousness (Gould and Stern 1993). In some of these studies, gender identity was found to play a significant role (e. g. , Coughlin and O’Connor 1985; Fischer and Arnold 1990; Jaffee 1991; Jaffee and Berger 1988; Martin and Roberts 1983). Within these studies, however, are conflicting results with respect to the relative importance of asculinity and femininity in explaining findings. For instance, while Coughlin and O’Connor (1985) found masculinity to be a significant predictor of consumer purchase intentions, Fischer and Arnold (1990) found femininity to be more important than masculinity in relationship to Christmas gift shopping. Several studies reported no significant findings related to gender identity, and, in fact, found more significant relationships related to biological sex than to gender identity (Allison et al. 1980; Golden, Allison, and Clee 1979; Gould and Stern 1989; Kahle and Homer 1985; Schmitt et al. 1988). One study reported significant results only in limited contexts (Gould and Weil 1991).
Another study used the BSRI as a measure of sex-role orientation rather than as a measure of gender identity (Qualls 1987); a positive relationship between gender identity and household influence was found, though a different measure of sex-role orientation was ultimately used in the study (Qualls 1987). Not surprisingly, throughout many of the BSRI studies, there is an echo of concern about the conceptualization and operationalization of the BSRI. In response to these concerns, three consumer behavior studies have examined these issues. Barak and Stern (1986), examining the relationship of demographic, psychographic, and behavioral variables to women’s gender identity, concluded that the masculinity scale of the BSRI seemed to be a self-assurance index, and, as such, was more relevant to consumer behavior than was femininity.
In 1987, Stern, Barak and Gould found that men identified with the BSRI masculine scale and women identified with the feminine scale, but both men and women were confused as to how to define the opposite sex—the outcome of this study was the development of a new measure, the Sexual Identity Scale (SIS), an undisguised measure of sexual identity that was hoped would provide a different way to assess gender identity than existing measures. However, subsequent research has found that the SIS is strongly associated with biological sex (Palan et al. 1999), and, consequently, adds little meaningful information about gender identity. Finally, the internal consistency of the BSRI has been examined (Palan 1998; Palan et al. 999), finding some inconsistent scale loadings with both the masculine and feminine scales. THEORETICAL FOUNDATIONS OF GENDER IDENTITY STUDIES IN CONSUMER BEHAVIOR Just as there has been a “litter of failed BSRI and PAQ studies” outside of marketing (Frable 1989, p. 106), so does the above discussion on the findings of gender identity studies in consumer behavior translate to the inability to make any definite conclusions about how gender identity impacts consumer attitudes and behaviors. Some consumer behavior researchers have interpreted the lack of significant findings and inability to duplicate findings as the result of poorly operationalized gender identity measures (Palan et al. 1999; Stern et al. 1987).
While operationalization problems have certainly contributed to disappointing findings, another factor may be a lack of understanding of the theoretical implications with respect to the conceptualization of gender identity. That is, researchers may not always have been aware of the theoretical differences between gender schema theory and multifactorial gender identity theory. Moreover, prior to Gould and Weil (1991), consumer behavior researchers may not have been aware that the measurement of gender identity could vary from one context to the next. And, given the inconsistent use of sex- and gender-related terms (Fischer and Arnold 1994), some researchers may well have conceptualized gender identity as salient to a particular study, when, in fact, another concept (e. g. , gender roles or gender attitudes) was more salient to the study6.
Consequently, many studies may already have been ill-fated at the conceptualization stage.
6 It should be noted that the inappropriate selection of sex- and gender-related terms can be linked to which gender identity theory a researcher believes to be most true—gender schema theory or multifactorial theory. For example, a researcher who conceptualizes a research study using gender schema theory would believe that a measure of gender identity (based on identification with Palan / Gender Identity in Consumer Behavior Research 12
The theoretical foundations of the gender identity studies were examined for conceptualization problems related to three issues: (1) whether or not the study conceptualized and operationalized gender identity as representative of broader gender variables (i. e. , consistent with gender schema theory) or as a construct representing only masculine and feminine personality traits (i. e. , consistent with multifactorial gender identity theory); (2) whether or not the inclusion of instrumental (masculinity) and expressive (femininity) personality traits was relevant to other study variables; and (3) whether or not any of the studies conceptualized gender identity as variable, subject to changing contexts and gender salient conditions. The findings of this examination are summarized in the last three columns of Table 1.
Because most of the studies do not explicitly categorize themselves into the distinctions represented by the three columns, categorization criteria were developed and used by the author to content analyze each article in Table 1. The categorization criteria are specified in the following paragraphs. Gender Schema Theory vs. Multifactorial Gender Identity Theory Studies were categorized as being consistent with gender schema theory (GST) when the article: (1) explicitly stated that the study was a test of gender schema theory; or (2) did not explicitly describe the study as a test of gender schema theory, but used the measurement of gender identity as an indication of some other sex- or gender-related concept.
In contrast, studies were categorized as being consistent with multifactorial gender identity theory (MGIT) when the article: (1) explicitly stated that the study was grounded in multifactorial gender identity theory; or (2) did not explicitly describe the study as grounded in multifactorial theory, but clearly defined gender identity7 as an individual’s identification with masculine and feminine personality traits and limited the interpretation of results to this definition; or (3) included other gender-related variables, in addition to the measurement of gendered personality traits, which was considered to be an indication that the researchers understood that gendered personality traits was just one gender-related factor, consistent with multifactorial gender identity theory. Relevance of Gender Identity to Other Study Variables Studies were categorized as relevant (yes), not relevant (no), or marginally relevant (marginal) depending on the degree to which identification with masculine and feminine personality traits might logically be associated with or predictive of other study variables (e. g. , products, brands, shopping ehaviors, consumption behaviors). This determination was based on indications provided in the literature review or theoretical background sections of the papers that supported the potential explanatory power of masculine and feminine personality traits to the relationships being studied; moreover, the potential explanatory power had to be recognized as not necessarily being the same as being male or female. Studies were considered to be relevant when evidence from previous research or logical arguments were presented to support specific aspects of instrumentality or communion (e. g. , autonomy, independence, nurturing, gentleness) as being linked to other study variables.
When studies did not provide arguments that specifically discussed why masculine and feminine personality traits were expected to be relevant, but instead rationalized that gender identity was relevant because biological sex was relevant, then studies were classified as not relevant. A few studies were classified as marginally relevant because they partially, but not fully, distinguished the reasoning for including gender identity as being separate from biological sex. masculine and feminine personality traits) is an accurate indicator of gender role attitudes. Qualls (1987) is an example of a study where gender identity is used to indicate sex role orientation; according to gender schema theory, this is perfectly correct procedure, but multifactorial theory (and the stance of this literature review) advocates that the concept most salient to the study is sex role orientation, not gender identity.
Hence, the study should measure sex role orientation, not gender identity. Of course, the inappropriate use of sex- and gender-related terms can also be linked to researcher carelessness. 7 Note that the use of gender identity, as used in consumer behavior research, is at odds with the use of the term in multifactorial gender identity theory. That is, in consumer behavior and in this literature review, gender identity refers to the degree to which an individual identifies with masculine and/or feminine personality traits. But in multifactorial gender identity theory, gender identity is a broader concept that includes gendered personality traits as just one factor.
When possible, this review will try to distinguish between these two terms when discussing the measurement of gendered personality traits in conjunction with multifactorial gender identity theory, however, the reader is reminded of this difference for those instances were the distinction is not clear.
13 Variability of Gender Identity Determination of whether or not a study recognized that individuals’ identification with masculine and feminine personality traits could vary when measured in different contexts or when gender identity was more or less salient to the other variables being examined, was based on explicit mention of this understanding in the article.
That is, when studies specifically discussed the changeability of gender identity in different contexts, they were classified as yes; if studies did not specifically discuss gender identity variability, they were classified as no. GENDER SCHEMA THEORY VS. MULTIFACTORIAL GENDER IDENTITY THEORY As can be seen in Table 1, the theoretical foundations of half of the studies are consistent with gender schema theory, and, therefore, are based on the assumption that gender identity, as measured with the CPI-FE, BSRI or the PAQ, is representative of broader gender concepts than just the gendered personality traits of instrumentality and expressiveness.
For example, Burns (1977) used the PAQ to sex-type respondents, and interpreted the sex-typing to be indicative of the respondents’ sex roles. Kahle and Homer (1985) also used gender identity as indicative of sex roles, and Qualls (1987) used the BSRI as a measure of sex role orientation. Several studies were conceptualized on the assumption that gender identity would be predictive of differences in information processing based on gender schema (e. g. , Gentry and Haley 1984; Jaffe 1991; Schmitt et al. 1988). The 17 studies that conceptualized gender identity consistent with multifactorial gender identity theory, for the most part, limited the interpretation of gender identity results (i. e. identification with masculine and
