Femininity and masculinity, or one’s gender identity (Burke et al. 1988; Spence 1985), refer to the degree to which persons see themselves as masculine or feminine given what it means to be a man or woman in society. Femininity and masculinity are rooted in the social (one’s gender) rather than the biological (one’s sex). Societal members decide what being male or female means (e.g., dominant or passive, brave or emotional), and males will generally respond by defining themselves as masculine while females will generally define themselves as feminine. Because these are social definitions, however, it is possible for a person to be female and see herself as masculine or male and see himself as feminine.

It is important to distinguish gender identity, as presented above, from other gender-related concepts such as gender roles, which are shared expectations of behavior given one’s gender. For example, gender roles might include women investing in the domestic role and men investing in the worker role (Eagly 1987). The concept of gender identity is also different from gender stereotypes, which are shared views of personality traits often tied to one’s gender, such as instrumentality in men and expressiveness in women (Spence and Helmreich 1978). And gender identity is different from gender attitudes, which are the views of others or situations commonly associated with one’s gender, such as men thinking in terms of justice and women thinking in terms of care (Gilligan 1982). Although gender roles, gender stereotypes, and gender attitudes influence one’s gender identity, they are not the same as gender identity (Katz 1986; Spence and Sawin 1985).

From a sociological perspective, gender identity involves all the meanings that are applied to oneself on the basis of one’s gender identification. In turn, these self-meanings are a source of motivation for gender-related behavior (Burke 1980). A person with a more masculine identity should act more masculine, that is, engage in behaviors whose meanings are more masculine such as behaving in a more dominant, competitive, and autonomous manner (Ashmore et al. 1986). It is not the behaviors themselves that are important, but the meanings implied by those behaviors.

Beginning at birth, self-meanings regarding one’s gender are formed in social situations, stemming from ongoing interaction with significant others such as parents, peers, and educators (Katz 1986). Although individuals draw upon the shared cultural conceptions of what it means to be male or female that are transmitted through institutions such as religion or the educational system, they may come to see themselves as departing from the masculine or feminine cultural model. A person may label herself female, but instead of seeing herself in a stereotypical female manner, such as being expressive, warm, and submissive (Ashmore et al. 1986), she may view herself in a somewhat stereotypically masculine manner, such as being somewhat instrumental, rational, and dominant. The point is that people view themselves along a feminine–masculine continuum, some seeing themselves more feminine, some as more masculine, and some as a mixture of the two. It is this selfperception along the feminine–masculine continuum that is their gender identity, and it is this that guides their behavior.

The roots of feminity/masculinity

In Western culture, men are stereotypically seen as being aggressive, competitive, and instrumentally oriented while women are seen as being passive, cooperative, and expressive. Early thinking often assumed that this division was based on underlying innate differences in traits, characteristics, and temperaments of males and females. In this older context, measures of femininity/masculinity were often used to diagnose what were understood as problems of basic gender identification, for example, feminine males or masculine females (cf. Terman and Miles 1936).

We now understand that femininity and masculinity are not innate but are based upon social and cultural conditions. Anthropologist Margaret Mead addressed the issue of differences in temperament of males and females in Sex and Temperament in Three Primitive Societies (1935). This early study concluded that there are no necessary differences in traits or temperaments between the sexes. Observed differences in temperament between men and women are not a function of their biological differences. Rather, they result from differences in socialization and the cultural expectations held for each sex.

Mead came to this conclusion because the three societies showed patterns of temperament that varied greatly with our own. Among the Arapesh, both males and females displayed what we would consider a ‘‘feminine’’ temperament (passive, cooperative, and expressive). Among the Mundugamor, both males and females displayed what we would consider a ‘‘masculine’’ temperament (active, competitive, and instrumental). Finally, among the Tchambuli, men and women displayed temperaments that were different from each other, but opposite to our own pattern. In that society, men were emotional and expressive while women were active and instrumental.

Mead’s study caused people to rethink the nature of femininity/masculinity. Different gender- related traits, temperaments, roles, and identities could no longer be inextricably tied to biological sex. Since Mead’s study, the nature–nurture issue has been examined extensively, and with much controversy, but no firm conclusions are yet clear (Maccoby and Jacklin 1974). While there may be small sex differences in temperament at birth (and the evidence on this is not consistent), there is far more variability within each sex group (Spence and Helmreich 1978). Further, the pressures of socialization and learning far outweigh the impact of possible innate sex differences in temperament. We examine this next.

The development of feminity and masculinity

There are at least three major theories that explain the development of femininity and masculinity: psychoanalytic theory (Freud 1927), cognitive-developmental theory (Kohlberg 1966), and learning theories that emphasize direct reinforcement (Weitzman 1979) and modeling (Mischel 1966). In all of these theories, a two-part process is involved. In the first part, the child comes to know that she or he is female or male. In the second part, the child comes to know what being female or male means in terms of femininity or masculinity.

According to psychoanalytic theory, one’s gender identity develops through identification with the same-sex parent. This identification emerges out of the conflict inherent in the oedipal stage of psychosexual development. By about age 3, a child develops a strong sexual attachment to the opposite- sex parent. Simultaneously, negative feelings emerge for the same-sex parent that are rooted in resentment and jealousy. By age 6, the child resolves the psychic conflict by relinquishing desires for the opposite-sex parent and identifying with the same-sex parent. Thus, boys come to learn masculinity from their fathers and girls to learn femininity from their mothers.

A later formulation of psychoanalytic theory suggests that mothers play an important role in gender-identity development. According to Chodorow (1978), mothers are more likely to relate to their sons as different and separate because they are not of the same sex. At the same time, they experience a sense of oneness and continuity with their daughters because they are of the same sex. As a consequence, mothers will bond with their daughters, thereby fostering femininity in girls. Simultaneously, mothers distance themselves from their sons, who respond by shifting their attention away from their mother and toward their father. Through identification with their father, boys learn masculinity.

Cognitive-developmental theory is another psychological theory of gender-identity development (Kohlberg 1966). Like psychoanalytic theory, it suggests that certain critical events have a lasting effect on gender-identity development, but these events are seen as cognitive rather than psychosexual in origin. Unlike psychoanalytic theory and learning theory (which is discussed next), cognitive- development theory sees the development of a gender identity as preceding rather than following from identification with the same-sex parent. Once a child’s gender identity becomes established, the self is then motivated to display gendercongruent attitudes and behaviors, well before same-sex modeling takes hold. Same-sex modeling simply moves the process along.

Kohlberg identifies two crucial stages of gender identity development:

(1) acquiring a fixed gender identity and

(2) establishing gender-identity constancy

The first stage begins with the child’s identification as male or female when hearing the labels ‘‘boy’’ or ‘‘girl’’ applied to the self. By about age 3, the child can apply the appropriate gender label to the self. This is when gender identity becomes fixed. By about age 4, these gender labels are appropriately applied to others. Within another year or two, the child reaches the second critical phase, that of gender constancy. This is the child’s recognition that his or her gender will not change despite changes in age or outward appearance.

The most social of the theories of genderidentity development are the learning theories. In these theories, it is the social environment of the child, such as parents and teachers, that shapes the child’s gender identity. Here, parents and teachers instruct the child on femininity and masculinity, either directly through rewards and punishments or indirectly through acting as models to be imitated.

Direct rewards or punishments are often given for outward appearance, such as what to wear (girls in dresses and boys in pants); object choice, such as toy preferences (dolls for girl and trucks for boys); and behavior (passivity and dependence in girls and aggressiveness and independence in boys). Through rewards and punishments, children learn appropriate appearance and behavior. Indirect learning of one’s gender identity emerges from modeling same-sex parents, teachers, peers, or models in the media. Children imitate rewarded models’ thoughts, feelings, or behavior because they anticipate that they will receive the same rewards that the models received.

Measuring masculinity and feminity: A Psychological View

Conceptualizing masculinity and femininity and measuring these orientations in men and women originated in the work of Lewis Terman and Catherine Cox Miles (1936), who created a 455-item test to detect masculinity and femininity. They called it the Attitude Interest Analysis Test (AIAT) to conceal its purpose from subjects. The test included such things as word associations, inkblot associations, interest items, and introversion–extroversion items. For example, on the interest items, persons got femininity points for liking (and masculinity points for disliking) ‘‘nursing,’’ ‘‘babies,’’ and ‘‘charades.’’ Individuals received masculinity points for liking (and femininity points for disliking) ‘‘people with loud voices’’ and ‘‘hunting.’’ On the introversion–extroversion items, persons got femininity points for agreeing (and masculinity points for disagreeing) that they ‘‘always prefer someone else to take the lead’’ and that they are ‘‘often afraid of the dark.’’ And they got masculinity points for agreeing (and femininity points for denying) that ‘‘as children [they were] extremely disobedient’’ and that they can ‘‘stand as much pain as others.’’ The responses did discriminate between the sexes, with men displaying higher masculinity and women displaying higher femininity.

Terman and Miles’s masculinity–femininity (M– F) scale became a model for M–F scales for more than three decades (see Morawski 1985 for a review). The M–F scales that followed shared four assumptions with the scale created by Terman and Miles. These included the assumptions that masculinity and femininity were:

(1) deep-seated, enduring characteristics of people

(2) not readily apparent in overt behavior

(3) linked to mental health (an incongruence in sex and masculinity and femininity signaled problems in psychological adjustment), and

(4) opposite ends of a continuum (Morawski 1987).

By the 1970s, researchers had become disenchanted with M–F scales, a timing that coincided with the reemergence of the women’s movement. Three criticisms had developed:

(1) The early M–F scales fostered research that exaggerated the differences between men and women

(2) the feminine characteristics in M–F scales often carried negative connotations, and

(3) the bipolar conception of masculinity–femininity was seen as problematic, that is, one could be masculine or feminine but not both (Morawski 1987).

From the third criticism arose the concept of androgyny and scales to assess it (see Morawski 1987 for a review of the problems surrounding the concept of androgyny).

Androgyny is a combination or balance of masculinity and femininity, allowing for the possibility that individuals can express both. Instead of conceptualizing masculinity and femininity as opposite ends of a continuum, where masculinity on one end precludes one from being feminine on the other end, in androgyny, masculinity and femininity are separate dimensions that can be combined. People can be masculine, feminine, or both (androgynous). Two of the more famous inventories that emerged from the impetus to measure masculinity and femininity on separate, independent dimensions were in psychology: the Bem Sex Role Inventory (BSRI) (Bem 1974) and the Personal Attributes Questionnaire (PAQ) (Spence and Helmreich 1978).

Both the BSRI and PAQ listed attributes that are positively valued for both sexes but are more normative for either males or females to endorse. These are known as the masculine scale and feminine scale, respectively. In the BSRI, respondents indicate the degree to which a series of descriptions are true about them. Examples of descriptions for the masculine scale include ‘‘acts as a leader,’’ ‘‘makes decisions easily,’’ and ‘‘willing to take risks.’’ Examples of descriptions for the feminine scale include being ‘‘affectionate,’’ ‘‘gentle,’’ and ‘‘sensitive to the needs of others.’’

In the PAQ, respondents rate themselves on a series of bipolar items. For the masculine scale, the items range from masculine to not masculine, while the items for the feminine scale range from feminine to not feminine. Examples of items from the masculine scale include ‘‘very independent’’ (vs. ‘‘not at all independent’’) and ‘‘can make decisions easily’’ (vs. ‘‘has difficulty making decisions’’). The feminine scale includes bipolar items such as ‘‘very emotional’’ (vs. ‘‘not at all emotional’’) and ‘‘very helpful to others’’ (vs. ‘‘not at all helpful to others’’). In addition to the masculinity and femininity scales, the PAQ has a third scale, labeled masculinity–femininity, that is in keeping with the bipolar M–F measurement tradition. The bipolar items for this scale are culturally appropriate for males on one end, and culturally appropriate for females on the other. Typical items include ‘‘very dominant’’ (vs. ‘‘very submissive’’) and ‘‘feelings not easily hurt’’ (vs. ‘‘feelings easily hurt’’).

With separate measures of masculinity and femininity, it is possible to ask about the relationship between the measure of masculinity and the measure of femininity. When this relationship was examined, it was found that the two scales were not strongly negatively related, as would be expected if masculinity were the opposite of femininity. Instead, using either the BSRI or the PAQ, the two ratings were relatively unrelated; a person’s score on one scale did not predict that person’s score on the other scale (Bem 1974; Spence and Helmreich 1978). People had all combinations of scores. Initially, people who combined high scores on masculinity with high scores on femininity were labeled as androgynous (Bem 1977; Spence and Helmreich 1978). Later, androgyny was indicated by a small difference between masculinity and femininity scores, representing balanced levels of these two characteristics. The other classifications were masculine (high M and low F scores), feminine (high F and low M scores), and undifferentiated (low F and low M scores).

The BSRI and the PAQ are embedded in very different theories about how gender-related characteristics are organized. For Bem (1981, 1993), scores on the BSRI measure not only the different dimensions of masculinity and femininity; more importantly, the scores measure an underlying unidimensional construct known as gender schematization. Gender schematization is an internalized tendency to see the world in gendered terms. One who is gender-schematic classifies stimuli as male or female rather than other according to other dimensions that could equally be used. Those who score high on masculinity or high on femininity are gender-schematic because they tend to organize information along gender lines. Androgynous people are gender-aschematic.

Spence (1984, 1993), on the other hand, suggests that gender phenomena are multifactorial. In this view, there are numerous attributes, attitudes, and behaviors that culturally distinguish between men and women, but these are not bound together as a single underlying property such as gender schematization. For Spence (1985; Spence and Sawin 1985) the important underlying construct is gender identity, or one’s sense of being masculine or feminine. Culturally defined personality traits, physical attributes, abilities, and occupational preferences, among other things, all contribute to one’s gender identity in unique and individualized combinations. Individuals draw upon these gender characteristics and choose those qualities that are compatible for them as they define themselves as masculine or feminine and ignore other gender qualities. Thus, while societal members may agree on the representation of masculinity and femininity, one’s own masculinity and femininity tends to be more variable and idiosyncratic in nature.

Rather than conceptualizing the items on the PAQ and the BSRI as referring to the broad categories of masculinity and femininity, Spence maintains that these items tap into socially desirable instrumental and expressive traits in men and women, respectively (Spence and Helmreich 1978, 1980). While these traits are related to masculinity and femininity, they do not define one’s overall gender identity. They are simply one of the set of contributors to one’s gender-based self-image. This is supported by the fact that scores on the PAQ and BSRI are not strongly relate to scores on other measures of gender attitudes, attributes, and behaviors (Spence 1993; Spence and Sawin 1985).

Measuring masculinity and feminity: A Sociological View

In sociology, the symbolic interactionist view of masculinity–femininity (Burke 1989; Burke et al. 1988; Burke and Tully 1977) shares much in common with the view held by the psychologist Spence and her colleagues. For symbolic interactionists, gender identity is understood in the context of a body of research known as identity theory (Stryker 1980). According to identity theory, the self is an organized collection of hierarchically arranged identities (self-meanings) that serve as a source of motivation for our behavior (Burke 1980). Identities are organized as control systems that act to maintain congruency between the internalized selfmeanings (one’s identity standard) and perceptions of the meaning of the self in ongoing social situations (Burke 1991). The key in this is one’s self-meanings (Osgood et al. 1957).

Individuals’ gender identity as masculine or feminine is based on the meanings they have internalized from their association with the role of male or female, respectively, in society. Since these are self-meanings, they cannot be directly observed; they must be inferred from behaviors and expressions in which people engage. Gender identity is one of many role identities people hold. In sociology, we assume that roles do not stand in isolation but presuppose and are related to counterroles (Lindesmith and Strauss 1956). For example, the role of teacher takes on meaning in connection with the role of student, the role of mother takes on meaning in relation to the role of child, and so on. The same is true of identities.

Just as the meaning of student (the student identity) is understood in relation to that of teacher (the teacher identity), so too is the meaning of male (masculinity) relative to that of female (femininity). The meanings of masculine and feminine are necessarily contrastive. To be male (masculine) is to be not female (feminine) and vice versa (Storms 1979). Gender meanings thus relate to one another as opposite ends of a single continuum, returning to the bipolar conceptualization of masculinity and femininity. Indeed, masculinity and femininity are negatively related when individuals are asked to judged themselves based on the self-descriptors ‘‘masculine’’ and ‘‘feminine’’ (Spence 1993). Interestingly, young children initially do not see masculine and feminine characteristics as opposites, but as they get older, their views of the genders become increasingly bipolar (Biernat 1991). This contrasting of masculinity and femininity in self-meanings does not necessarily hold for behaviors, since one can engage in both masculine and feminine behaviors.

The procedure that symbolic interactionists use to measure gender identity is based on the method devised by Burke and Tully to measure self-meanings (1977). In an analysis of middle school children’s gender identities, Burke and Tully first collected sets of adjectives from children that the children themselves used to describe the images of boys and girls. These adjectives, together with their opposites, were used as adjective pairs to form a semantic differential scale to measure the meanings of the male and female roles (Osgood et al. 1957). The stem for the semantic differential was ‘‘Usually [boys, girls] are . . . ‘‘Then, through a statistical procedure known as discriminant function analysis, the researchers selected those items that best discriminated between the meanings of boys and girls. Examples of items that best distinguished ‘‘girl’’ meanings from ‘‘boy’’ meanings for these children included ‘‘soft’’ (vs. ‘‘hard’’), ‘‘weak’’ (vs. ‘‘strong’’), and ‘‘emotional’’ (vs. ‘‘not emotional’’). After selecting the most discriminating items, children’s self-descriptions (‘‘As a [boy, girl] I usually am . . .’’) were then summed to form a scale of gender identity.

A gender identity scale constructed along the lines described above has certain properties. First, the scale evolves out of the meanings of maleness in relation to femaleness that actually are held in the population from which the sample is drawn. This procedure contrasts with much research that uses attributes that are assumed to carry meanings of masculinity and femininity with no attempt to check whether they have these meanings for respondents. Second, the measure outlined above incorporates the assumption that meaning is contrastive. The meaning of female is in contrast to the meaning of male and vice versa. Third, by focusing upon self-meanings, we separate issues of who one is (gender identity) from what one does (gender roles) or what one believes (gender attitudes and stereotypes). From this perspective, androgyny may be thought of not as combining masculine and feminine meanings, but as being flexible in the kinds of behaviors in which one engages (sometimes more masculine in meaning, sometimes more feminine in meaning). We now review some of the important research on gender identity that has emerged out of the Burke–Tully method, for which the symbolic interactionist perspective has served as the backdrop.

Research on feminity and masculinity

The symbolic interactionist perspective suggests that the self is defined through interactions with others. Burke and Tully’s (1977) work found that children with cross-sex identities (boys who thought of themselves in ways similar to the way most girls thought of themselves and vice versa) were more likely than children with gender-appropriate identities:

(1) to have engaged in gender-inappropriate behavior

(2) to have been warned about engaging in gender-inappropriate behavior, and

(3) to have been called names like ‘‘tomboy,’’ ‘‘sissy,’’ or ‘‘homo.’’

Not surprisingly, boys and girls with cross-sex gender identities were more likely to have low self-esteem.

Another symbolic interactionist tenet is that people will choose behaviors that are similar in meaning to the meanings of their identities (Burke and Reitzes 1981). Burke (1989) found that among middle school children, boys and girls with a more feminine gender identity earned higher grades than those with a more masculine gender identity. This was true independent of the child’s sex, race, or grade, the subject area, or the sex of the teacher. Since the early years of schooling are more likely to be ‘‘feminized’’ because there are more female than male teachers (Lipman-Blumen 1984), children with a more feminine identity will likely perform better in a ‘‘feminine’’ institution. Among college students, research has shown that males and females with a more feminine gender identity are more likely to inflict and sustain both physical and sexual abuse in dating relationships (Burke et al. 1988). People with more feminine gender identities are likely to be more emotionally expressive and relationship-oriented. Aggression may be used as a last resort to attain a closer relationship.

Within the symbolic interactionist tradition, research demonstrates that the meanings that people attribute to themselves as masculine or feminine (their gender identity) are sometimes more important in predicting how they will behave than is their gender (male or female). For example, early research on conversational behavior reported that males were more likely than females to use more dominant and assertive speech patterns in interaction, such as interrupting (West and Zimmerman 1983) and talking more (Aries 1976). However, a later review of the many empirical studies on interruptions and time spent talking showed that gender had inconsistent effects (James and Clarke 1993; James and Drakich 1993). This inconsistency might be explained through an analysis of gender identity rather than gender. For example, research shows that persons with a more masculine gender identity, irrespective of their gender as male or female, are more likely to use overlaps, interruptions, and challenging statements in a conversation (Drass 1986; Spencer and Drass 1989).

While gender identity may sometimes be more important than gender in determining outcomes, it is also possible for one’s gender (male or female) and one’s gender identity (masculine or feminine) to each result in different displays of behavior. For example, in an analysis of problem-solving discussions between newly married spouses, females and those with a more masculine gender identity were more likely to express negative, oppositional, and dominating behavior, such as complaining, criticizing, or putting down their spouse (Stets and Burke 1996). While masculinity more than femininity should increase dominating behavior (as discussed above), it was surprising that females engaged in more dominating behavior than males. It was discovered that this dominating behavior emerged especially from females who were viewed by their spouses as being in a weak, subordinate position in our society. These women were apparently using coercive communication to counteract the (subordinate) view of them and gain some control. The problem is that in compensating for their weaker status by behaving in a dominant fashion, women may unwittingly be reminding men of their weak position.

While one’s gender identity is generally stable over time, it sometimes changes in response to different experiences. To examine stability and change in gender identity, Burke and Cast (1997) examined the gender identities of newly married couples over the first three years of marriage, finding that the year-to-year stability in gender identity was moderately high. This means that while the gender identities of the respondents did change over this period of time, they did not change markedly. Looking from month to month or week to week, there was almost no observable change.

Identity theory suggests that identities are most likely to change in the face of persistent changes in the environment. The birth of a first child represents a dramatic and persistent change in the environment that confers femininity on women and masculinity on men. Burke and Cast showed that when a couple had their first child, women’s gender identities became more feminine and men’s gender identities become more masculine. Social psychological processes may also modify one’s gender identity. Burke and Cast also found that the more a spouse took the perspective of the other in the marriage, the more the spouse shifted his or her gender identity in the direction of the other’s gender identity. The other’s gender identity was thus verified and supported by the spouse, which may act to minimize marital conflict.

The Future

We mention below several avenues of possible future work in femininity/masculinity. Many more avenues could be identified, for this is an area rich for continuing investigation, but our space is limited. First, we are only beginning to understand issues of stability and change in one’s gender identity. How might gender identities be modified through participation in societal institutions such as the economy, religion, and politics? For example, to what extent and in what ways might employers socialize employees into particular views of being masculine or feminine in order to maintain a smooth flow of work and profit? Are some people more resistant to this socialization than others? Is this issue tied to how relevant gender identity is to individuals?

Related to the above is a second avenue of research, that is, the salience of gender identity across individuals, groups, even cultures. Salience refers to the probability that a particular identity will be invoked in a situation (Stryker 1980). This will vary by situation, but it also varies across individuals. For some, gender is not very relevant, and for others gender is almost always relevant. This returns us to Bem’s notion of gender schematization, or the tendency to see the world in gendered terms. What makes gender identity more or less salient for people, and what are the consequences of that?

Third, we know very little about subcultural, cultural, and cross-cultural differences in the meanings that are attached to femininity and masculinity. Most of what we know concerns Western cultures, yet as Margaret Mead discovered long ago, these patterns are not universal. We need to investigate the variation in the meanings of being masculine and feminine. Such studies may help us understand a society’s division of labor, its differential power and status structure, and how its privileges and responsibilities are allocated. To modify the social system may mean first modifying individual beliefs about masculinity and femininity.

Related Links

This Aricle was Written by

This Article was Published in
Second Edition
A Book by

University of Washington, Seattle


Managing Editor
University of Kansas, Lawrence


edu.learnsoc.org Copyright 2010 - 2012 © All Rights Reserved
  Home | About | Contact | Links