For decades, as I taught classes in the psychology of women and gender, a pivotal moment occurred early in the semes-ter. It happened when I asked my students to imagine themselves as participants in the world describe by Ursula Le Guin (1969) in her science fiction novel, The Left Hand of Darkness. In that tale, the inhabitants of Winter, Le Guin’s imaginary planet, are not distinguished by sex. Not divided into sex categories in their daily lives, they become female or male for only a few days each month, during which they engage in (hetero) sex and may initiate a pregnancy. In Le Guin’s scenario, individuals could not predict whether they would become female or male in any given month, so a person might end up as the father of several children and the mother of several others.

My students were often unnerved, sometimes even distraught, by this fantasy. In heated discussions, they noted that, although LeGuin’s described arrangement would pull the rug out from under sex discrimination, deeply change parenting arrangements, and allow people to determine their career directions without concerns about tokenism and minority status, it would also make human relationships boring and life uninteresting. It would, they thought, make it more difficult to come up with an identity and with a set of personal guidelines for how to act, dress, and connect with other people. Perhaps, they thought, planet Winter would be a more just place than planet Earth, but it would be dull.

The reactions to this fantasy call into question any notion that consideration of the meaning of femininity and masculinity is merely an arcane academic exercise. Femininity and masculinity may be slippery, abstract concepts, impossible to define, and subject to cultural and temporal modification—but their impact is undeniably deeply personal and powerful. When Sandra Bem took on these concepts as a focus of inquiry, she zeroed in on a core aspect of our psychology—one that had remained virtually unexamined until she challenged some of the most basic assumptions that had been made about them. The articles in this issue speak to and extend the numerous ways in which she interrogated, challenged, and aimed to disrupt notions of gender categorization. A reading of them underscores the lasting impact of her thinking on our field.

Themes in Bem’s Work Identified in the Special Issue

A reading of the articles in this issue suggests that Sandra Bem’s conceptual and empirical work made its mark by making the invisible visible. From her classic article peeling back the assumptions built into the “nonconscious ideology” of sexism (Bem and Bem 1976), to her development of a method to see (and to measure) femininity and masculinity in a new way rather than as opposing poles on a single continuum (Bem 1974), to her highlighting and querying of the process by which children are trained to categorize so much of their environment unnecessarily according to gender (Bem 1981, 1993), her ability to get all of us to look at gender differently, to try out new perspectives, is evident. Her adoption of the notion of The Lenses of Gender to title her major theoretical book (Bem 1993) was thus particularly apt.

What is primarily made visible through Bem’s work is the impact of societal context on the experience and expression of gender categories. As noted in several of the articles in this special issue (Keener and Mehta 2016), Bem argued that society, by emphasizing gender categories inordinately, encourages children (and adults) to consider gender as a highly salient category, to process information in terms of gender categories, and to adopt and enact such categories more broadly and more rigidly than necessary. Liben and Bigler (2015) build on this idea by showing that gendered language applied to activities and roles affects children’s preferences for these activities and roles. As noted in the following, Mehta and Dementieva (2016) and Keener and Strough (2016) investigate ways in which social context can affect the performance of gender—again underscoring the notion that gender is not simply a trait, but is developed and constructed by individuals in interaction with the environment.

If the shape and importance of gender categories are socially constructed, then it stands to reason that they could be re-imagined, re-shaped, and perhaps de-emphasized through changes in society. One of the things about Sandra Bem that many feminist psychologists seem to appreciate is her strong commitment to use her conceptual and academic work to drive social change. From her early work with Daryl Bem on the impact of sex-discriminatory job advertisements (Bem and Bem 1973) to her commitment to use her understanding of gender schemas and sexism schemas to enact gender–neutral parenting (Bem 1998), she demonstrated a fierce commitment to the practical, personal politics of gender. As noted in her memoir (Bem 1998), she did what she could to minimize the gratuitous use of gender categories to label people, activities, and careers and to channel them in gendered directions. At the most personal level, she devoted a major effort to raising her own children in a gender-neutral way and to sharing her methods of doing that. Indeed her story of her son Jeremy’s adventure in wearing a barrette to school remains a tale told in many university classrooms—a concrete and still relevant illustration of the power of gender polarization.

Bem deliberately and persistently applied her work to shap-ing the social environment, but, in turn, that environment also affected and shaped her work and its impact. As Golden and McHugh (2016) and Liben and Bigler (2015) note, she hated to be categorized, and this resistance to external attempts to pigeonhole her no doubt fueled her efforts at disrupting society’s gender categories. Starr and Zurbriggen (2016) also note that her scholarly style—more focused on elucidating big-picture concepts than on testing the detailed workings of those concepts—may have been partly driven by her joint position in Women’s Studies, where a “lone-scholar” model and interdisciplinary, conceptual, activism-oriented work tends to be supported and encouraged more than in psychology departments. Golden and McHugh (2016) suggest that sexism, in terms of the academy’s sometimes dismissive judgements about the importance of scholarly work on women and gender, may well have dampened the recognition given to her work and her persistence in testing certain aspects of her theories.

The conclusion driven by the articles in this issue is that Bem’s theoretical insights and conceptual work have inspired a staggering amount of research, driving hundreds of studies of androgyny (Donnelly & Twenge, this issue; Martin, Cook, & Andrews, this issue) and used in wide-ranging ways in investigations of the cognitive underpinnings of gender (Starr & Zurbriggen, this issue). There is agreement, however, that there is a great deal of work yet to be done in the empirical exploration of her theories. As the researchers represented here have extended and built upon her work, they have attempted to resolve contradictions, identify and answer questions and refine concepts in Bem’s treatment of gender categorization, androgyny, and gender schema theory—and they have also applied her work to new practical questions.

Themes in the Examination and Extension of Bem’s Work

Among the most dominant (and sometimes overlapping) themes in these papers building on Bem’s work are interrogation of the concept of identity, exploration of the impact of social context on the expression/performance of gender, attention to developmental issues, and refining/improving the measurement of identification with gender categories and of gender schematicity.

Identity

With the Bem Sex Role Inventory (BSRI), Bem (1974) attempted to measure femininity, masculinity, and androgyny by examining respondents’ endorsement of adjectives shown in pilot studies to be differentially desirable for women and men. Although a leap forward from the traditional unipolar and unidimensional notion that femininity and masculinity were defined by agreement with forced-choice items that maximally differentiated females and males, this approach has been criticized as far from perfect. It is not clear, for example (e.g., Spence 1983), that respondents who strongly endorse most of the masculine items and few of the feminine items on the BSRI or other measures of masculinity/femininity actually think of themselves as masculine—something I have been reminded of on a yearly basis when giving undergraduate students the opportunity to complete and score the BSRI.

In some articles in this issue, the question of identity is refined and some of its complexities unpacked. For example, Menon (2016) examines a model of gender identification that includes dimensions of felt typicality, felt contentedness, and felt pressure. Her study shows that all three dimensions of gender identification are related to gender-typicality of relationship styles and that, furthermore, adolescent girls and boys whose relationship styles were gender-atypical decreased in felt typicality and felt contentedness over several months. The findings suggest that identity shifts in response to self-observation, and they support Bem’s idea that individuals assess their behavior through the lens of a gender schema. Martin et al. (2016) take another approach to identity, describing an “identity androgyny” perspective that considers gender identification as reflecting how the self identifies with and feels a connection with both one’s own and the other gender. They describe research findings supportive of the idea that children who expressed high similarity to both gender groups experienced benefits in terms of peer relationships. Bem’s elaboration of the concept of androgyny and her BSRI provided a place to start in the quest to understand the complexities of gender identification and its consequences. As these more recent approaches show, however, there is more to gender identification than endorsing a series of stereotypically feminine or masculine traits.

Social Context and the Performance of Gender

One of the major themes in Sandra Bem’s body of work is that gender, as well as the way we think about it and act it out, is culturally constructed. Gender categories, she argued (e.g., Bem 1981, 1993), were built from societal expectations, from social arrangements that channeled individuals into gender-segregated groups, from an accumulation of situations, and from informational assertions that made salient the distinction between female and male. However, she did not emphasize the ways in which, having been the recipient of all these cultural messages, an individual’s expression or performance of gender might also be affected by context on a situation-by-situation basis. Two papers in this issue advance our understanding of the importance of immediate context in gendered behavior. Mehta and Dementieva (2016) explore the idea that gender has both trait- and state-like qualities. By showing that femininity and masculinity, as assessed by a short, momentary version of the BSRI, apparently change in step with the gender of the peers in U.S. college students’ immediate environment, they provide support for the idea that gender is, to some extent, not simply a trait, but a state that can be affected in noticeable ways by the interpersonal context. Keener and Strough (2016), also demonstrate the influence of interpersonal context on the performance of gender, showing that U.S. women’s and men’s preferred agentic and communal strategies for resolving conflicts vary in response to the gender of the person with whom they anticipate a conflict. They illustrate the point that gender is not simply something individuals have, as an aspect of personality, but something they do or perform in reacting to a particular context.

Keener and Strough (2016) found that women and men do not respond as mirror images of each other in reacting to the gender composition of social contexts, and they suggest that a change in such contexts may have different consequences for women and men. Mehta and Dementieva (2016) and Menon (2016) also found that women’s and men’s responses were not neatly parallel. Well-documented in U.S. studies, such differences may be rooted in the differences and in the status and value attached to feminine and masculine behavior (Cikara and Fiske 2009; West et al. 2012), in the relative strength and negativity of the messages females and males receive that they should not cross gender boundaries (Croft, Schmader, and Block 2015; Hort, Fagot, and Leinbach 1990; Reigeluth and Addis 2016), and/or to other uninvestigated variables. The persistent tendency for girls and boys, women and men, to not simply respond as opposites, to not just exchange places in response to contextual shifts, underlines Bem’s (1974) original point that femininity and masculinity cannot be understood only as opposing poles on the same continuum.

Gender Development

If society constructs gender and individuals internalize it, this does not happen all at once; there must be developmental processes involved. As the articles in this issue illustrate, Bem’s account of the development of gendered identification through interactions between cultural processes of gender polarization and cognitive processes of self-socialization via gender schemas helped to generate considerable research attention on gender development. Her analysis of how children learn to think about and identify with gender categories, may, however, have had its strongest impact via her description of her own attempts to raise children in a gender-neutral way. She tried to raise her children with both relatively narrow gender schemas (focusing on the reproductive organs as the only critical difference between females and males) and relatively well-developed sexism schemas. Interestingly, whereas gender schema theory has generated a great deal of research, few researchers have investigated children’s acquisition of sexism schemas or the impact of such schemas. Liben and Bigler (2015) describe two studies in which children in the United States were successfully taught to notice and critique sexism in peers’ remarks and in the media. The long term impact of teaching children to recognize sexism calls for more investigation.

With respect to the development of sex-typing and gender schemas, much research attention has been focused on children. Authors in this issue have argued for more attention to other age groups, such as adolescents (Menon, this issue), young adults (Keener & Strough, this issue; Mehta &, Dementieva, this issue) and for a lifespan perspective (Starr & Zurbriggen, this issue). Because societal pressures toward gender role conformity do not end with childhood, but rather continue through social interactions and situational cues throughout the lifespan (Deaux and Major 1987), a focus that reaches beyond childhood should provide greater understanding of whether, how, and how much individuals modify and adapt their gender-related self-views over their lifetimes. Additionally, as noted by Donnelly and Twenge (2016), individuals who grow up in different historical periods are exposed to different informational and normative pressures with respect to gender roles and identity—making it important to consider that different cohorts (e.g., women who formed their identities during second-wave or third-wave feminism or during periods when there was lesser or greater societal support for women in the labor force) may experience different pressures with respect to gender. There remains considerable scope for researchers to explore the ways in which historical/cultural change affects gender roles and identities in ways that may be cohort-specific.

Measuring the Concepts

Even though the BSRI has been used widely and persistently since its inception, it has never been without controversy in terms of the adequacy with which it measured identification with gender roles or sex-typing, as noted by Golden and McHugh (2016) and Martin et al. (2016). Perhaps because it represented a conceptual breakthrough, because it provided a clear improvement over the traditional M-F measures, and because it was easy to use, its imperfections were often glossed over. However, it is arguable that a better measure would have promoted more and better research into the antecedents, correlates, and consequences of androgyny.

Starr and Zurbriggen (2016) also cite measurement issues to explain why Bem’s gender schema theory, although broadly generative in terms of conceptual reach, did not always generate the focused, programmatic research that might be expected from such a potentially important and practical theory. As they note, Bem used the BSRI as a way of measuring the likelihood that individuals would be gender-schematic in their information processing, arguing that someone who was gender-typed should be more likely to rely on gender-schematic processing. Starr and Zurbriggen comment that there is “challenging slippage between the constructs of sex-typed and gender-schematic” (p. 9), that may have made it difficult to test and advance the theory. I think they are correct on this point. Researchers have found the proposed link between gender-typing and gender schematicity to be problematic (Deaux, Kite, and Lewis 1985; Spence 1991), and many years ago, at least one study demonstrated, among a sample of U.S. college women, that gender schematicity as measured by response latency to gender-stereotyped traits was uncorrelated with gender-typing (Freedman and Lips 1996).

Another measurement issue is identified by Martin et al. (2016) in their discussion of the complexities of testing Bem’s proposal that androgyny is linked to behavioral flexibility and thus to better mental health. They note that, whereas Bem (Bem 1974, 1975; Bem and Lenney 1976) demonstrated empirically that androgynous individuals were more likely in some situations to cross, and to feel comfortable crossing, the boundaries of gender roles, other researchers have not always found the predicted link between androgyny and adjustment or mental health. They argue that one source of this apparent inconsistency may be that the BSRI measures individuals’ views of their typical selves, but not of their felt potential to behave differently under different conditions. They report on research that assessed self-perceived capability of enacting various interpersonal behaviors, which illustrated that feeling broadly capable over a range of behaviors that tapped both stereotypical femininity and stereotypical masculinity was linked to self-esteem and other adjustment-related variables. Other research included in their discussion indicated that, rather than a simple lack of rigidity, the strategy that is positively related to adjustment is flexibility appropriate to specific situations, or functional flexibility. Their use of a dynamical systems approach to assess adaptive flexibility represents a creative advance in tackling one of Bem’s most interesting hypotheses: that the flexibility inherent in androgyny is related to adaptability and good mental health.

Practical Questions for Future Research

A reading of the articles in this special issue suggests that there are many intriguing questions stemming from Bem’s initial work on androgyny, gender schema theory, and the impact of gender polarization that have yet to be explored. Liben and Bigler (2015) comment that we do not yet know enough about how easily gender as a category can be ignored or minimized. Nor, they argue, do we know very much about the extent to which children’s gender-typing of the self and of others are related, or about how the content of children’s gender schemata may be changing in response to cultural changes in women’s and men’s role. Furthermore, it appears that gender fluidity has become more accepted in the years since Bem’s initial work, and little is known about the impact on gender schema development of that increased acceptance. Starr and Zurbriggen (2016) and Mehta and Dementieva (2016) suggest that research on gender schema theory should branch out to include more studies of queer and genderqueer populations in order to investigate the role gender schemas may play in these groups’ subversion of gender roles. They argue, too, that gender schema theory could be profitably combined with other theoretical approaches, such as objectification theory, to enrich both approaches and provide a better understanding of behavior. Martin et al. (2016) advocate for more research on questions about how adaptive flexibility may act protectively to buffer negative consequences of crossing gender boundaries. Finally, although, as noted by Starr and Zurbriggen (2016), Bem’s theories and measures have been used in research in a number of different countries, most of the research has clearly been based in Western cultures. Keener and Strough (2016) call for an expansion of the research to a variety of cultures, Donnelly and Twenge (2016) illustrate the potential importance of examining gender stereotypes across temporal contexts, and several of the authors represented in this issue note the necessity of re-examining the content of stereotypical femininity and masculinity as measured by the BSRI in light of both historical change and potential applicability in other cultural contexts.

The BSRI and Changes in Gender Attitudes

One arguable benefit of the long-term and widespread use of the BSRI, imperfect though it may be, is the potential to use changes in patterns of responses to assess historical changes in gender-related attitudes. Donnelly and Twenge (2016) address this possibility in their discussion of two meta-analyses that, in combination, reveal shifts in U.S. college students’ endorsement of items on the femininity (F) and masculinity (M) scales of the BSRI between 1974 and 2012. They report that, over that time, women’s M scores increased significantly, whereas their F scores, along with men’s M and F scores, showed no change. However, in the years between 1993 and 2012, women’s F scores decreased significantly, whereas their M scores did not change, and both scores remained stable for men. One pattern documented by these findings is that women and men have not changed in step with each other—at least in terms of adjusting their self-endorsed qualities. Women, not men, are the ones who have been changing in the way they view themselves.

These findings may show that women have become more masculine and less feminine over the years, if we accept the early 1970s definitions on which the scales were based. However, as the authors note, the scale items may not reflect modern gender stereotypes. Thus, the changes in women’s response patterns may say little about their identification with femininity or masculinity. Rather, it may say something about cultural changes in the acceptability of certain qualities in women; in other words, they may tell us that cultural notions of femininity have changed.

The scale items on the BSRI that purport to measure femininity and masculinity may be in need of revision, but the traditional assumption that femininity and masculinity are opposites may still be widely accepted. A recent U.S. study suggests that gender stereotypes, in terms of the traits, role behaviors, occupations, and physical characteristics attributed to women and men, remain as entrenched and dichotomized now as they were during the 1980s (Haines, Deaux, and Lofaro 2016). More than 40 years after the publication of Bem’s (1974) original article detailing the development of the BSRI, many of my students still look as though a light has dawned or a curtain has been pulled away when they are presented with the idea that femininity and masculinity do not have to be conceptualized as opposites. Along with the many ways in which the authors in this issue have demonstrated the continuing usefulness of her ideas as springboards for new understandings of gender, the startled expressions on my students’ faces remind me that Sandra Bem’s work is not simply an important piece of feminist history, but has continuing relevance in the “real” world of people’s experience. I suspect this state of affairs would disappoint, but not surprise her.