Felt gender typicality and sex-typing: Examining felt gender typicality, sex-typing, and their relation to adjustment
Table of Contents Dedication iii Acknowledgments iv Abstract v List of Figures x List of Tables xiii List of Appendices xiv Chapter 1 1 1.1 Overview 1 1.2 Question 1: The basis of felt gender typicality 4 1.3 Question 2: Felt gender typicality, sex-typing and adjustment 5 1.4 Question 3: The role of centrality, evaluation, and felt pressure 6 1.5 Background 7 1.6 The present study 22 Chapter 2 24 2.1 Participants 24 2.2. Procedure 25 2.3 Measures 25 Chapter 3 32 3.1 Introduction to results 32
3.2 Correlations among variables 35 3.3 Age and sex differences 38 Chapter 4 45 4.1 Overview 45 4.2 Does children’s liking for same-sex activities relate to felt gender 45 typicality? 4.3 Does the flexibility of children’s gender beliefs moderate the 47 relation between sex-typing and felt gender typicality? 4.4 Does the degree of liking for feminine and masculine sex-typed 49 interests, relative to each other, influence felt gender typicality? Chapter 5 51 5.1 Overview 51 5.2 Do felt gender typicality and sex-typing have the same relation to 51 adjustment? 5.3 Does sex-typing moderate the relation between felt gender 58 typicality and adjustment? Chapter 6 63 6.1 Overview 63 6.2 Does the extent to which children experience their gender as 63 important (centrality), valued (evaluation), and constraining (felt pressure) impact felt gender typicality and sex-typing’s
associations with adjustment? Chapter 7 71 7.1 Overview 71 7.2 The basis of felt gender typicality 71 7.3 Felt gender typicality, sex-typing, and adjustment 73 7.4 Sex-typing as a moderator of the relation between felt gender 76 typicality and adjustment 7.5 Unexpected findings/limitations 77 7.6 Future directions 88 7.7 Summary and conclusions 92 Appendices 125 References 140
List of Figures Figure 1 Gender stereotype flexibility as a potential moderator of the 94 relation between sex-typed preferences and felt gender typicality. Figure 2a Potential relations among felt gender typicality, sex-typed 95 preferences and adjustment. Figure 2b Sex-typed preferences as a moderator of the relation between 96 felt gender typicality and adjustment. Figure 3a Centrality, evaluation, and felt pressure as potential moderators 97 of the relation between felt gender typicality and adjustment. Figure 3b Centrality, evaluation, and felt pressure as potential moderators 98 of the relation between sex-typed preferences and adjustment. Figure 4 Predicted values for felt gender typicality as a function of 99 participant sex and feminine interests. Figure 5 Predicted values for felt gender typicality as a function of 100 participant sex and masculine interests. Figure 6 Predicted values for felt gender typicality as a function of 101 masculine interests and gender stereotype flexibility. Girls only. Figure 7 Predicted values for felt gender typicality as a function of age, 102
feminine interests, and masculine interests. Girls only. Figure 8 Predicted values for self-worth (Rosenberg) as a function of 103 age and felt gender typicality. Figure 9 Predicted values for self-worth (Harter) as a function of age, 104 sex, and felt gender typicality. Figure 10 Predicted values for social anxiety as a function of sex and 105 feminine interests. Figure 11 Predicted values for self-worth (Rosenberg) as a function of sex 106 and masculine interests. Figure 12 Predicted values for social anxiety as a function of age, sex, and 107 masculine interests. Figure 13 Predicted values for self-worth (Harter) as a function of age, 108 feminine interests and masculine interests. Girls only. Figure 14 Predicted values for self-worth (Rosenberg) as a function of age, 109 felt gender typicality, and feminine interests. Girls only. Figure 15 Predicted values for self-worth (Rosenberg) as a function of age, 110 felt gender typicality, and feminine interests. Boys only. Figure 16 Predicted values for peer competence as a function of age, felt 111 gender typicality, and feminine interests. Girls only. Figure 17 Predicted values for peer competence as a function of age, felt 112 gender typicality, and masculine interests. Girls only.
Figure 18 Predicted values for self-worth (Rosenberg) as a function of 113 feminine interests and centrality/evaluation. Boys only. Figure 19 Predicted values for self-worth (Rosenberg) as a function of 114 age, feminine interests, and centrality/evaluation. Girls only. Figure 20 Predicted values for self-worth (Harter) as a function of 115 masculine interests and centrality/evaluation. Girls only. Figure 21 Predicted values for social anxiety as a function of age, felt 116 gender typicality, and felt pressure. Girls only. Figure 22 Predicted values for self-worth (Harter) as a function of age, 117 feminine interests, and felt pressure. Boys only.
List of Tables Table 1 Correlations among sex-typed preferences, gender stereotype 118 flexibility, gender identity, and adjustment variables. Table 2 Means (SDs) for sex-typed preferences, gender stereotype 119 flexibility, gender identity, and adjustment variables by sex and cohort. Table 3 Sex and sex-typed preferences predicting felt gender typicality. 120 Table 4 Age as a moderator of the relation between felt gender 121 typicality and self-worth (Rosenberg). Sex as a moderator of the relation between sex-typed preferences and self-worth (Rosenberg) and social anxiety. Table 5 Sex-typed preferences as a moderator of the relation between 122 felt gender typicality and self-worth (Rosenberg) and peer competence. Table 6 Centrality/evaluation as a moderator of the relation between 123 sex-typed preferences and self-worth. Results for the Rosenberg are for boys only. Results for the Harter are for girls only. Table 7 Summary of major findings. 124
List of Appendices
Appendix A 125 Analyses examining differences between siblings vs. non-siblings and repeat vs. non-repeat participants. Appendix B 127 Lit of all measures administered. Appendix C 132 Full description of measures used in the present study. Appendix D 137 Factor loadings for the Sex-Typed Preferences measure based on an exploratory factor analysis with promax rotation. Appendix E 138 Initial regression analyses conducted with the masculine non-athletic and masculine athletic subscales.
Chapter 1 Introduction 1.1. Overview Boys and girls differ in a number of ways. From toy preferences to academic interests to occupational aspirations (see Ruble, Martin, & Berenbaum, 2006), sex typing manifests itself across a variety of domains. Because sex typing permeates many aspects of an individual’s life, the consequences and implications of sex-typing have been an area of considerable interest to psychologists, with researchers examining everything from clothing choice to playmate preference as an outcome of sex-typing. One major area of interest concerns the relation between sex-typing and psychological adjustment. Research in this area has generally been conducted from one of two perspectives, with one approach characterizing sex-typing as good for psychological adjustment (Kagan, 1964) and the other approach characterizing sex- typing as bad for psychological adjustment (Bem, 1981). Although both perspectives have been the subjects of numerous critiques (e.g., Bem, 1974; Spence & Helmreich, 1980, 1981), much contemporary research remains heavily influenced by the latter perspective. This is perhaps most evident in the large number of studies concerned with documenting the ways in which gender may limit children’s opportunities to
develop valuable skills or competencies in certain areas (see Ruble et al., 2006). While the value of this type of research is indisputable, recent research suggests that the relation between sex-typing and psychological well-being may be more complex than has been previously acknowledged. It is widely recognized among gender development researchers that gender identity is multidimensional (Ruble et al., 2006); and prior research (see Ruble et al., 2004, 2006) points to four dimensions of gender identity—felt gender typicality (feeling similar to same-sex others), centrality (the importance of gender as an identity), evaluation (the feelings individuals ascribe to their gender identities), and felt pressure (pressure felt from parents, peers, and the self to behave in a sex typed manner)—likely to be relevant to the relation between sex typing and adjustment. Of these dimensions of gender identity, felt gender typicality and its relation to adjustment is perhaps the most contentious. Specifically, although several recent studies have shown that felt gender typicality is a positive predictor of self-worth and peer competence for children (Carver, Yunger, & Perry, 2003; Corby, Hodges, & Perry, 2007; Egan & Perry, 2001; Smith & Leaper, 2005), the notion that felt gender typicality may promote positive adjustment is provocative. This is largely due to the wide acceptance of Bem’s (1981) idea that being highly sex-typed contributes to maladjustment. That perspective coupled with the assumption that in order to feel gender typical one must also be sex-
typed has rendered findings showing positive associations between felt gender typicality and adjustment controversial. Yet, theoretical perspectives differ as to the extent to which gender identity is expected to relate to sex typing and little is known about the relation between felt gender typicality and sex-typing. Although measures of sex-typing have been used to infer individuals’ gender identities in the past, researchers argue that this practice is problematic (Tobin, Menon, Menon, Spatta, Hodges, & Perry, 2010). Moreover, even if felt gender typicality and sex-typing are related, they represent different constructs. That is, felt gender typicality represents individuals’ self-judgments of similarity to same-sex others whereas sex-typing represents the extent to which individuals possess attributes that are considered to be more prevalent among same-sex than other-sex members. Thus, a primary goal of the present research is to investigate the extent to which felt gender typicality is associated with sex-typing, and whether they have the same consequences and implications for children’s well-being. In addition, in spite of wide acceptance of the idea that gender identity is multidimensional, few researchers have examined how other dimensions of gender identity may qualify felt gender typicality and sex-typing’s relations with adjustment. That is, felt gender typicality may be positively associated with adjustment, but only among individuals who consider gender a central and positive aspect of their self- concept. In addition, having same-sex interests may not be harmful for adjustment
for individuals who do not feel pressured to behave in a sex-typed way. Thus, a second goal of the present research is to examine how different dimensions of gender identity influence felt gender typicality and sex-typing’s relations to adjustment. The present research extends prior work concerning the implications of sex- typing for adjustment by addressing three primary research questions: (1) What is the relation between felt gender typicality and sex-typing? (2) Do felt gender typicality and sex-typing have similar consequences and implications for adjustment? and (3) How do different dimensions of gender identity felt gender typicality and sex-typing’s relations to adjustment? 1.2 Question 1: The basis of felt gender typicality Our first question concerns the basis of felt gender typicality. Ideas regarding the basis of felt gender typicality can be traced back to Spence (1984), who believed that self-judgments of gender typicality resulted from a complex mental calculus in which individuals considered 1) the extent to which they possessed same-sex characteristics and 2) how central or definitive of each gender they considered such characteristics to be (see Tobin et al., 2010). Thus, we wanted to examine 1) whether children’s reported liking for same-sex activities predicted felt gender typicality and 2) whether children’s beliefs about those attributes, in this case the flexibility of children’s gender stereotypes, moderated the relation between felt gender typicality and sex-typing. That is, while girls who possess more masculine interests may also
report lower felt gender typicality, we were interested in whether these girls would feel more gender typical if they held flexible beliefs regarding who should engage in male sex-typed activities. In addition, we wanted to explore whether felt gender typicality was necessarily at odds with possessing other-sex interests. That is, concern over the positive associations between felt gender typicality and adjustment stem not only from the assumption that high levels of felt gender typicality may be commensurate with high levels of sex-typing, but also from the idea that if individuals have high levels of same-sex interests, that they also have low levels of other-sex interests, and therefore possess a limited behavioral repertoire. However, Bem (1974, 1981) has argued that individuals can possess high levels of same- and other-sex interests, and that people with this constellation of interests (e.g, androgyny) have the best adjustment outcomes. Thus, we explored how relative levels of same- and other-sex interests, within an individual, influenced feelings of gender typicality. 1.3 Question 2: Felt gender typicality, sex-typing, and adjustment A second goal of the present research is to examine whether felt gender typicality and sex-typing have similar consequences and implications for adjustment. While felt gender typicality has consistently been shown to be positively associated with adjustment (see Section 1.1), predictions regarding sex-typing are less clear. When sex-typing has been linked to adjustment, measures of sex-typing have
generally been based on participants’ self-endorsements of gendered personality traits. However, measuring sex-typing in this way is considered to be problematic for a variety of reasons (Spence, 1984). In addition, given that being sex-typed in one domain (i.e., personality) only moderately predicts being sex-typed in another domain (i.e., behavior) (Ruble et al., 2006), Bem’s (1981) concerns about the behavioral inflexibility of sex-typed individuals may be unfounded. That is, even if individuals are sex-typed, they may not be as rigid, and therefore as prone to maladjustment, as Bem suggested. In addition to examining the implications of felt gender typicality and sex- typing for adjustment separately, we were also interested in their joint implications for adjustment. Specifically, although several recent studies clearly show that felt gender typicality is positively related to adjustment, we were interested in whether this relation would vary with level of sex-typing. We anticipated, for example, that felt gender typicality may have more impact for children showing low levels of same-sex interests or high levels of other-sex interests. 1.4 Question 3: The role of centrality, evaluation, and felt pressure The final question in the present research concerns whether different aspects of gender identity—centrality, evaluation, and felt pressure—impact felt gender typicality and sex-typing’s relations with adjustment. Studies suggest that how children think about their group identity is critical for understanding identity relevant
outcomes (Ruble et al., 2004). Yet, research examining felt gender typicality and sex- typing and their relations to adjustment has given little consideration to how children’s thoughts about gender identity may impact these relations. However, research in the social identity literature show that different dimensions of identity are informative regarding adjustment outcomes (e.g., Rowley, Sellers, Chavous, & Smith, 1998). Thus, we were interested in whether these three dimensions of identity would moderate felt gender typicality and sex-typing’s associations with adjustment. For example, although we anticipated that felt gender typicality would be beneficial for adjustment, we thought it was possible that this could be limited to children who consider gender to be both an important and positive part of the self. Likewise, we thought that any beneficial or benign implications of sex-typing for adjustment could be limited to children who do not feel pressured to behave in a sex-typed way. 1.5 Background In this section, we review the literature most relevant to the present study. First, we present an historical overview concerning psychologists’ interest in and thinking about the relation between sex-typing and adjustment, highlighting the theoretical shifts that have occurred over the last 40 years. Following that, we discuss more recent perspectives on gender identity and their implications for how we should examine the relation between sex-typing and felt gender typicality. Next, we briefly review the literature concerning felt gender typicality and sex-typing and their
relations to adjustment. Finally, we discuss the multidimensional nature of gender identity and its relevance to the present research. 1.5.1 Historical overview. In spite of the controversy concerning findings showing positive relations between felt gender typicality and adjustment, the idea that felt gender typicality may be beneficial for adjustment is not new. Since Freud, researchers have suggested that feeling gender atypical may adversely affect adjustment (Perry, 2004). In fact, the contrasting notion that sex-typing may contribute to maladjustment is a relatively more recent theoretical development. This can be attributed, in large part, to Bem’s (e.g., 1974, 1981) pioneering work suggesting that internalization of gender norms may limit people’s ability to react adaptively across situations, thereby resulting in maladjustment. Bem’s ideas proved to be enormously influential and since that time, androgyny, rather than gender typicality, has come to be viewed as the optimal standard (Ruble & Martin, 1998). Initially, it was thought that identification with a same-sex parent would lead to the adoption of sex-typed attributes, allowing children to perceive themselves as similar to same-sex others. This was believed to result in conceptions of the self as masculine/feminine. In addition, it was argued that should things go awry such that children identify with the opposite-sex parent, they would fail to adopt sex-typed attributes. Failure to behave in a sex-typed manner was believed to result in children feeling different from same-sex others and regarding themselves as less masculine/feminine, resulting in significant maladjustment. Thus behaving in a sex-
typical manner was considered optimal because of its theorized role in promoting a sense of self as male or female (Kagan, 1964). However, during the mid 1970s, the field of gender development experienced an important theoretical shift, with androgyny rather than gender typicality, becoming the admired standard (Ruble & Martin, 1998). In a radical departure from previous theories of sex-typing, Bem (1974) convincingly argued that rather than exemplifying healthy adjustment, sex-typed individuals may be at risk for significant maladjustment. Specifically, Bem proposed that internalization of gender norms could limit people’s ability to react adaptively across situations. Supporting her claim, data showed that when given the choice of performing a “sex-appropriate” versus “sex- inappropriate” task, sex-typed individuals opted to perform more “sex-appropriate” than “sex-inappropriate” behaviors, even though they would be paid more money to do the “sex-inappropriate” activities. Furthermore, when sex-typed participants were forced to perform a “sex-inappropriate” activity, they reported higher feelings of psychological discomfort compared to androgynous individuals. Thus, it was concluded that because sex typed individuals were reticent to engage in opposite-sex typed behavior, even when it was more advantageous for them to do so, sex-typing promoted rigid and potentially dysfunctional behavior, not healthy psychological adjustment (Bem & Lenney, 1976). While Bem’s (1981, 1984) ideas that sex-typing may result in maladjustment were revolutionary, like those theorists before her, she assumed that endorsement or
enactment of sex-typed characteristics were directly linked to a person’s sense of self as a male or a female. Thus, the notion that in order for people to feel gender typical they must be gender typical remains with us today, rendering the idea that feeling gender typical may promote positive adjustment a controversial issue among researchers. Indeed, suggesting that feeling gender typical may be good for adjustment is at odds with Bem’s arguments that we should promote androgynous, rather than sex-typical behavior, in children, but only if one subscribes to the notion that feelings of gender typicality are necessarily commensurate with the number of same-sex interests an individual possesses. Unlike Bem (1981, 1984), who believed that sex typing was indicative of individuals’ sense of themselves as male or female, Spence (1984) argued that sex typing was less directly connected to gender identification. In support of this idea, Spence (1993) noted that the global items “masculine” and “feminine” on the Bem Sex Role Inventory (BSRI) formed a separate bipolar factor. Spence believed that this suggested that participants’ self-ratings of these items represented something different from their self-ratings of the other trait-based masculine/feminine items (see also Pedhazur & Tetenbaum, 1979). Furthermore, Spence (1984) speculated that people needed to possess only a certain number of gender-congruent characteristics to feel masculine/feminine, implying that people need not be rigidly sex-typed to feel typical for their gender. Spence theorized that people periodically compute summary judgments of themselves
as gender typical by weighing more heavily those gender congruent qualities they posses and discounting those they do not. Spence used homosexuals’ and heterosexuals’ conceptions of homosexuals’ masculinity/femininity as an example. Specifically, Spence (1984) argued that homosexuals give less weight to sexual preference and more weight to other sex-typed attributes when making summary judgments of their masculinity/femininity. In contrast heterosexuals, weigh sexual orientation more heavily when evaluating themselves and others as masculine/feminine. In essence Spence’s ideas suggest that individuals may use different criteria, in a potentially self-serving way, when evaluating their sense of maleness or femaleness. This implies that while felt gender typicality and sex-typing are undoubtedly related, their relation to each other is likely to be influenced by individuals’ thoughts or feelings about gender, particularly their ideas of what aspects of sex-typing are most definitive of their gender category. 1.5.2 Current perspectives. Like Spence (1984), Tobin and colleagues (2010) argue that it is dubious to infer an individual’s gender identity directly from his/her level of sex-typing. Expanding on Spence’s (1984) ideas, they suggest that when examining self-judgments of gender typicality it is important to assess both children’s gender self-attributions (e.g., “I like trucks”) as well as the ways in which they view gender-attribute associations (e.g., “Boys like trucks”) in relation to their gender group.
Relevant to this idea are suggestions that children’s personal interest or participation in a particular activity may influence their perception of whether an activity is masculine or feminine, which in turn may influence the relation between sex-typing and felt gender typicality. For example, Liben and Bigler (2002) argue that despite knowing that ballet is a female stereotyped activity, a boy who enjoys ballet may nevertheless change his opinion that ballet is primarily for girls. Thus, a boy who enjoys ballet may still feel gender typical, provided that he is flexible in his beliefs as to who should do ballet. In the present study we examined the relation between felt gender typicality and sex-typing, taking into consideration that children may have unique views regarding which attributes they associate with their gender and that the flexibility of their gender beliefs in particular may moderate this relation. Specifically, we asked children to consider not only whether they liked a particular sex-typed activity, but also whether they believed a specific activity was primarily for girls, boys, or both. In this way, we were able to account for children’s personal beliefs about the nature of gender-stereotyped activities when examining the relation between felt gender typicality and sex-typing. In addition, in light of the highly influential nature of Bem’s (1981) ideas concerning the benefits of androgyny for adjustment, we explored whether felt gender typicality, even if it was associated with having greater same-sex interests, necessarily
implies the absence of having other-sex interests. Given that Bem was primarily concerned with the relative levels of same- and other-sex-typed interests within an individual, it seemed important to examine the joint impact of same- and other-sex interests on felt gender typicality. 1.5.3 Felt gender typicality, sex-typing, and adjustment. Felt gender typicality. felt gender typicality refers to children’s feelings of similarity to same-sex others. Measures in current use ask children to consider their similarity to same-sex others across a variety of domains and are often more concrete and specific than parallel measures used with adults (e.g., Spears, Doosje, & Ellemers, 1997). For example, children may be asked to consider whether they have the same interests as same-sex others or whether their personality is similar to same-sex others’ personalities (e.g., Egan & Perry, 2001). Thus, felt gender typicality is really more of a direct reflection of perceived similarity to same-sex others than an explicit evaluation of whether one is typical of or normal for their gender group, although there is likely a great deal of overlap with these latter judgments. Studies with adults have assessed feelings of typicality or similarity to group members to approximate levels of group identification (e.g., Hogg & Hardie, 1991; Hogg, et al., 1993). However, researchers have begun to recognize that, although group identification and self-typicality may be part of the same underlying construct, self-judgments of typicality are important in their own right and empirically distinct
from group identification (Kashima, Kashima, & Hardie, 2000; Jetten, Spears, & Manstead, 1997; Spears, et al., 1997; Verkuyten & Nekuee, 1999; Egan & Perry, 2001). In a recent review, Ashmore, Deaux, and McLaughlin-Volpe (2004) concluded that perceived goodness of fit with a social group (i.e., self-typicality) should be measured separately from other measures of identification, in part, because of its association with specific identity behaviors. Research with children has recognized that felt gender typicality constitutes an important component of gender identity (see Egan & Perry, 2001; Carver, Yunger, & Perry, 2003), and several studies have shown that feeling similar to same-sex others is associated with better psychological adjustment in middle childhood (see section 1.1.1). Ruble et al., (2004) also suggest that a sense of feeling connected to others may be a key component underlying the relation between group identity and self- esteem. That is, given the notion that humans have a universal need to feel connected to and form relationships with others (Baumeister & Leary, 1995), strongly identifying with or feeling connected to members of a social group may have significant psychological benefits. Moreover, a recent study with college students revealed a significant, positive correlation between felt gender typicality and feelings of being a typical college student, suggesting that felt gender typicality may, at least in part, reflect a sense of belongingness or connectedness to others (DiDonato, Fesi, Backer, Ravago, Ndimbie, Roberts, & Berenbaum, 2006). Thus, we predicted that felt gender typicality would be positively associated with adjustment.
Sex-typing. Confusion regarding both the conception and measurement of sex- typing can be traced back to when Constantinople (1973) declared the terms masculinity and femininity “to be among the muddiest concepts in the psychologist’s vocabulary” (p. 390). As already noted, one of Bem’s (1974) major contributions was to measure masculinity and femininity separately in individuals. Unfortunately Bem focused primarily on individuals’ personal endorsements of a set of masculine and feminine personality traits, which later proved to be highly problematic, particularly when thinking about relations between scores on the BSRI and adjustment outcomes (e.g., Spence, 1984). Owing to the fact that children do not reliably understand traits until well into middle-childhood (Ruble & Dweck, 1995), assessments of sex-typing in children have relied less on measuring personal endorsements of masculine and feminine personality traits, instead utilizing both parent and observational reports of, for example, children’s behaviors or peer choices. However, these measures have also been subject to a number of limitations, such as not allowing masculinity and femininity to remain orthogonal (See Ruble & Martin, 1998 for a more detailed discussion). Thus, in the present study, we sought to avoid the pitfalls of earlier research with adults by measuring sex-typing as the extent to which children reported liking same- and other-sex activities. In addition, children’s liking for same- and other-sex activities was grouped into separate subscales, thus allowing masculine and feminine interests to vary independently.