Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 May 10.
Published in final edited form as: Gender Issues. 2019 Oct 22;37(1):1–24. doi: 10.1007/s12147-019-09246-y

An Inductive Analysis of Young Adults’ Conceptions of Femininity and Masculinity and Comparison to Established Gender Inventories

Mike C Parent 1, Laurel R Davis-Delano 2, Elizabeth M Morgan 3, Nathaniel W Woznicki 1, Alisha Denson 1
PMCID: PMC11086686  NIHMSID: NIHMS1913572  PMID: 38736455

Abstract

This study provides an analysis of the content of feminine and masculine characteristics/behaviors described in writing by 366 young women and 289 young men from the U.S. Emergent characteristics/behaviors were placed into domains. For both femininity and masculinity, the domains of “physical differences related to sex” and “emphasized physical differences” emerged. For masculinity, additional domains were: “activities and interests focused on the body,” “powerful or oriented toward power,” and “emotion-control or emotionally-limited.” For femininity, additional domains were “lacking power,” “orientation to other people,” and “emotional.” We then compared the characteristics/behaviors and domains we discovered to gender inventories that are commonly used in the contemporary period. The masculine domains focused on physical differences, activities, and interests that emerged from the present study are mostly absent from these masculinity inventories. The domains focused on power and restricted emotion are evident in these inventories, but these inventories do not cover all of the characteristics within our domains. The feminine domains that emerged from the present study are more often covered in these inventories, but some of the specific feminine characteristics we found are not evident in these inventories. Results are discussed in terms of gender role theory, gender inequality, and potential application for qualitative and quantitative inquiries into the construction of gender.

Keywords: Sex roles, Gender differences, Measurement, Scaling

Scholars of gender have long been interested in the social construction of gender, including the measurement of gender role conformities and ideologies. The measurement of gender role conformities and ideologies has seen advancements in the delineation of taxonomies of femininities and masculinities. Consistent with recommendations for measurement development [18], the process of inventory development for gender role ideologies or conformities usually relies on the use of focus groups in an initial step, so as to generate content for a traditional scale-type measure (e.g., [32, 33, 57]). Qualitative methods that are used to generate items and domains for these inventories are rarely used beyond this purpose, and, as far as we know, have not been used as a means to assess the continued relevance of the content of these inventories. Yet, the richness of qualitative data itself can generate important insights, and may be used to further explore and refine quantitative and qualitative work on the construction of gender (e.g., [6]). Further, because feminist movements and other social forces generate changes in societal gender roles and ideologies over time, including substantial changes in the lives of women in the United States since the 1970s, periodic assessment of existing gender inventories is necessary. Related to these goals, in the present study we undertook inductive analysis of qualitative data to reveal common conceptions of femininity and masculinity in the U.S. and we explored interrelationships between these conceptions.

The Construction of Masculinities and Femininities

As new theories of gender have been developed, new instruments have been designed to assess gender conformities and ideologies. Contemporary theory focuses on gender not as a nomothetic construct but as numerous constructs (e.g., masculinities and femininities; [12]), which vary among individuals, between social contexts, and across different time periods. Concurrently, operationalization of gender roles has transformed from assessments of singular aspects of gender (e.g., expressiveness and agency as in the Bem Sex Role Inventory or the Personality Attributes Questionnaire; [2, 23]) toward multifaceted assessments of gender role ideologies and conformities, including the Conformity to Masculine Norms Inventory (CMNI/CMNI-46; [32, 43]), the Male Role Norms Inventory (MRNI; [28, 30]), the Femininity Ideology Scale (FIS; [29]), and the Conformity to Feminine Norms Inventory (CFNI/CFNI-45; [33, 44]), among many others.

These measures of gender roles have tended to reflect either gender role ideology theory (i.e., the characteristics that individuals believe men and women should possess related to their gender) or gender role conformity theory (i.e., personal adherence to the characteristics that are proscribed for their gender). Examples of the former category include the FIS and MRNI, and of the latter include the CFNI and CMNI. Both of these theories tend to emphasize intra-psychic and interpersonal beliefs, rather than performative or physical, aspects of gender. In general, men and women tend to report moderate levels of traditional gender role ideology, moderate conformity to gender roles for their own sex, and low conformity to cross-sex gender roles.

Quantitative studies designed to explore the construction of gender have suggested multiple feminine and masculine domains in U.S. society. A comprehensive review of the domains of these measures is beyond the scope of this manuscript, though Levant [27], and more recently Thompson and Bennett [55], provide overviews of masculinity inventories. Fundamentally, overlap between measures of masculinity suggest common themes of (1) an emphasis on sexuality; (2) avoidance of femininity, especially concern that one may be perceived to be gay; (3) self-reliance; (4) social dominance; and (4) restricted emotional expression [28, 30, 32, 45].

Fewer measures exist by which to assess feminine gender role ideology and conformity, though two of the relatively more widely used measures are the FIS and the CFNI. The original FIS was composed of five domains: stereotypic image and activities, dependence/deference, caretaking, purity, and emotionality. In the revision of this measure, only the factors of dependence/deference, purity, and emotionality/traditional roles persisted. The CFNI, and its revision the CFNI-45, delineated the factors of thinness, investment in appearance, domestic, involvement with children, modesty, relational, sexual fidelity, romantic relationship, and sweet and nice. Consistent across these measures are emphases on (1) the importance of relationships, and specifically sexual purity or fidelity and (2) a focus on physical appearance and conformity to traditional standards of femininity.

Whereas quantitative measurement of femininities and masculinities is critical to furthering research on the topic and integrating the science of gender studies into other fields, the reduction of qualitative data into survey items, and at best correlational matrices, can obscure facets of femininity and masculinity expressed by participants in qualitative research as well as analysis of the interrelationships among these facets. To our knowledge, qualitative data have never been used to test the content of gender roles present within the existing measures. We argue that returning to qualitative analysis of responses to open-ended questions about gender constructs on a regular basis is important to detect both continuity and change in societal gender constructs. This is necessary because gender roles and ideologies change over time.

Despite this reality, only a few scholars have studied historical changes in scores on commonly utilized gender inventories [3, 19, 53, 58] and historical changes in scores on particular domains/subscales and specific characteristics/items included in these inventories [1, 16, 17, 22]. Although such studies are critical, because these studies use existing measures they are unable to discover new domains/subscales and specific characteristics/items that should be added to existing measures. An inductive approach serves this purpose. Recently, Berger and Krahe [3] used an inductive survey approach to generate items for a new gender inventory, in this case to measure gender conceptions in Germany [3].

Present Study

Our primary goal was to utilize inductive qualitative analysis to examine what contemporary young men and women in the U.S. describe as characteristics of femininity and masculinity. Rather than rely on existing measures of gender roles, we sought to explore the topography of self-generated written responses to prompts about participants’ understanding of femininity and masculinity. When doing so, we examined similarities and differences between our men and women participants. After securing inductive results, we compared our findings to commonly used existing inventories assessing femininity and masculinity to corroborate, as well as to identify potential gaps and other limitations in, these quantitative measures. Heterosexual young adults were selected as participants for this study because heterosexual (not sexual minority) people have the most impact on dominant gender ideals in society. Also, gendered expectations are especially salient during emerging adulthood [40] and reflect the most current zeitgeists of gendered thinking [5]. Our findings may be useful to scholars who construct and modify quantitative gender inventories, and—in a more general sense—our findings may further inform understandings of underlying conceptions of gender in U.S. society.

Method

Procedure

IRB approval was secured for this project from the second and third authors’ institution. Participants were drawn from two sources to complete an online questionnaire about “beliefs and behaviors related to gender and sexual orientation”: Mechanical Turk (mturk) and the introductory psychology participant pool at a college in northeastern U.S. Scholars have found that recruiting research participants from the U.S. via mturk generates samples that are equally or more representative of the U.S. population as recruiting from other commonly utilized sources (e.g., [41]) and that mturk samples are generally valid and reliable (e.g., [8]). Because we sought heterosexual-identified male or female participants who were age 18–25 who both resided and grew up in the U.S., a short demographic questionnaire was given to potential mturk participants to pre-screen for these characteristics. The main questionnaire was only offered to potential participants who met these criteria. These same eligibility criteria were double-checked via questions on the main questionnaire, and main questionnaires with answers that were not aligned with these criteria were removed from the study. All participants completed the questionnaire via a secure online survey platform. Participants from mturk were given four dollars for questionnaire completion; introductory psychology pool participants were given course credit. Only participants who answered the questions pertaining to femininity and masculinity were included in the analyses for the present study.

Participants

The final sample included 655 U.S. participants (366 women and 289 men), 356 from mturk and 299 from the college introductory psychology pool. The mturk sample had proportionately more men than the college sample (179 versus 110), χ2 (1) = 12.00, p < .001. Participants ranged in age from 18 to 25 (Mwomen = 20.84, SDwomen = 2.49; Mmen = 21.32, SDmen = 2.33). Mturk participants were, on average, older than the college participants, Levene’s test F = 115.88, p < .001; equal variances not assumed t (573.39) = 28.39, d = 2.18. Regarding race/ethnicity, women participants identified as White (81%), multiracial (6%), Black (5%), Asian (5%), Hispanic (3%), and Native American, Middle Eastern, or another group (less than 1% each). Among men, participants identified as White (77%), Asian (10%), multiracial (7%), Hispanic (4%), Black (2%), and Middle Eastern or another group (less than 1% each). In the mturk sample, participants identified as White (70%), Asian (13%), multiracial (8%), Black (4%), Hispanic (4%), or another group (1%); in the college sample, participants identified as White (90%), multiracial (4%), Black (3%), Hispanic (2%), Asian (less than 1%), or another group (less than 1%). The mturk sample contained proportionately more Asian participants, χ2 (1) = 38.03, p < .001, while the college sample contained proportionately more White participants, χ2 (1) = 38.37, p < .001. Among mturk participants, 48% were currently in college full or part time; of those in college, 79% were completing an undergraduate degree, 21% were in graduate programs, and less than 1% were completing a professional degree. Of those not in college, less than 1% had less than a high school diploma, 17% had a high school degree, 23% completed some college or an associate’s degree, 58% had a bachelor’s degree, 2% had a master’s degree, and less than 1% had a professional degree.

Measures

Following a set of basic demographic questions, participants responded to two prompts, one focused on masculinity and one focused on femininity. Each prompt inquired, “Among people your age, what are the characteristics/behaviors that are considered [masculine/feminine]? (Please list at least 5 characteristics/behaviors.)” This prompt was chosen to (1) ask about “people your age” rather than the participants’ own beliefs, (2) directly address masculinity and femininity, rather than men/women or males/females, and (3) open the potential responses to a wide range of characteristics rather than limiting them to cognitive phenomena as has been the case in the development of some measures (e.g., [32, 33]). This openness related to the prompts is consistent with the goals of inductive research, which allows participants to define targets themselves rather than the researchers imposing rigid parameters on responses. Participants were provided ample text box space to type in their responses. The names of the second and third authors were in the announcement about the survey opportunity and on the consent form, potentially revealing the gender of the researchers to the participants; however, we do not think it is likely that participants took note of this information or that it influenced their responses on the questionnaire.

Analysis

The primary analysis of the data included identifying categories of characteristics/behaviors that emerged in response to the open-ended prompts about femininity and masculinity. Using inductive analysis that resembles grounded theory [52], the second author used open and axial manual coding to identify themes and sub-themes according to literal content. Then, the prevalence of themes was determined via content analysis [48]. The second author developed a codebook that was used to re-code the data by the 4th and 5th authors. The first author independently coded 10% of the responses. Interrater agreement [26] between the first author and the 4th & 5th authors was high (Cohen’s kappa = .81 for masculinity codes and .75 for femininity codes, indicating near-perfect and substantial concordance). We retained codes that were mentioned by 5% or more of the sample, i.e., were mentioned by 32 or more participants.

To compare similarities and differences in our sample, we undertook two analyses. To assess for equivalencies between proportions of responses between groups, we conducted equivalence testing. Equivalence testing relies on establishing an effect size at which a difference in proportions is so small that the groups are functionally equivalent [51]. To do so, we converted the z-tests to Cohen’s d values, and used values of d = .10 and d = .05 to represent close and equivalent proportions, respectively. We assessed differences between groups (total sample men versus total sample women, college women versus college men, mturk women versus mturk men, college men versus mturk men, college women versus mturk women) using a series of z-tests for differences between proportions, with a Bonferroni correction applied such that the critical value for a difference test was z = 3.09.

For the final phase of our analysis, the second author examined all previously identified common codes of characteristics/behaviors that emerged from analysis of the open-ended responses to the prompts, and placed these codes into larger domain categories that were based on common content among these codes. The first author then reviewed the resulting larger domain categories developed by the second author, and then the two authors refined these categories in collaboration. We placed a few codes in more than one domain, as these codes can be interpreted in more than one way (e.g., “strength” can be perceived as physical or emotional).

Results

Tables 1 and 2 display the frequency counts for the total sample, and for the samples divided by gender and data-source. Tables 3 and 4 feature the larger domain categories that we generated as well as the codes that we placed into these domain categories. Figures 1 and 2 display visual images of the main findings featured in Tables 3 and 4. Supplemental Tables 1 through 4 report test statistics for difference and equivalence tests.

Table 1.

Frequency Counts for Participant Descriptions of Masculinity

Total n in category Total Sample Total Sample
 Women
 Men
Women Men College Mturk College Mturk
655 366 289 189 177 110 179
Code
Strong (e.g., strong, can lift heavy things) 313 (48%) 170 (46%) 143 (49%) 91 (48%) 79 (45%) 44 (40%) 99 (55%)
Muscles (e.g., muscular) 211 (32%) 102 (28%) 109 (38%) 51 (27%) 51 (29%) 38 (35%) 71 (40%)
Athletic/Plays Sports (e.g., plays sport, jock) 162 (25%) 64 (17%) 98 (34%) 32 (17%)a 32 (18%)b 38 (35%)a 60 (34%)b
Hairy (e.g., hairy, hairy legs) 142 (22%) 72 (20%) 70 (24%) 35 (19%) 37 (21%) 31 (28%) 39 (22%)
Emotionally-Limited (e.g., not emotional) 125 (19%) 71 (19%) 54 (19%) 38 (20%) 33 (19%) 21 (19%) 33 (18%)
Confident (e.g., sure of oneself) 125 (19%) 80 (22%) 45 (16%) 44 (23%) 36 (20%) 23 (21%) 22 (13%)
Tough (e.g., rugged) 118 (18%) 54 (15%) 64 (22%) 31 (16%) 23 (13%) 27 (25%) 37 (21%)
Works Out/Physically Fit (e.g., goes to the gym) 109 (17%) 57 (15%) 52 (18%) 27 (14%) 30 (17%) 20 (18%) 32 (18%)
Aggressive (e.g., aggressive) 104 (16%) 61 (17%) 43 (15%) 31 (16%) 30 (17%) 16 (15%) 27 (15%)
Sport-Related (e.g., follows sport, likes sport) 96 (15%) 68 (19%) 28 (10%) 36 (19%) 32 (18%) 14 (13%) 14 (8%)
Deep Voice (e.g., deep voice) 91 (14%) 34 (9%) 57 (20%) 10 (5%) 24 (14%) 14 (13%) 43 (24%)
Brave (e.g., courageous) 81 (12%) 50 (14%) 31 (11%) 32 (17%) 18 (10%) 12 (11%) 19 (11%)
Independent (e.g., self-reliant) 81 (12%) 44 (12%) 37 (13%) 28 (15%) 16 (9%) 15 (14%) 22 (12%)
Powerful (e.g., dominant) 75 (11%) 47 (13%) 28 (10%) 22 (12%) 25 (14%) 9 (8%) 19 (11%)
Alcohol Consumption (e.g., drinks heavily) 70 (11%) 49 (13%) 21 (7%) 30 (16%) 19 (11%) 12 (11%) 9 (5%)
Tall (e.g., tall) 72 (11%) 35 (9%) 37 (13%) 15 (8%) 20 (11%) 11 (10%) 26 (15%)
Competitive (e.g., competitive) 60 (9%) 34 (9%) 26 (9%) 18 (10%) 16 (9%) 6 (5%) 20 (11%)
Sexual (e.g., has sex) 54 (8%) 40 (11%) 14 (5%) 23 (12%) 17 (9%) 9 (8%) 5 (3%)
Smart (e.g., intelligent) 53 (8%) 26 (7%) 27 (9%) 17 (9%) 9 (5%) 11 (10%) 16 (9%)
Involvement with Vehicles (e.g., interested in cars) 49 (7%) 32 (9%) 17 (6%) 21 (11%) 11 (6%) 3 (3%) 14 (8%)
Financially Able (e.g., financial success) 43 (7%) 35 (10%) 8 (3%) 19 (10%) 16 (9%) 2 (2%) 6 (3%)
Lifts Weights/Bodybuilder (e.g., lifts) 45 (7%) 16 (4%) 29 (10%) 11 (6%) 5 (3%) 17 (15%) 12 (7%)
Heterosexual (e.g., dates women) 41 (6%) 24 (7%) 17 (6%) 17 (9%) 7 (4%) 11 (10%) 6 (3%)
Clothing (e.g., wears dark colors) 42 (6%) 22 (6%) 20 (7%) 13 (7%) 9 (5%) 7 (6%) 13 (7%)
Assertive (e.g., assertive, pushy) 34 (5%) 27 (7%) 7 (2%) 9 (5%) 18 (10%)a 3 (3%) 4 (2%)a
Loud (e.g., loud, yelling) 33 (5%) 21 (6%) 12 (4%) 10 (5%) 11 (6%) 6 (5%) 6 (3%)
Open in a new tab

Superscripts denote proportions that are significantly different at p < .001 from proportions with the same superscript

Table 2.

Frequency counts for participant descriptions of femininity

Total n in category Total Sample Total Sample
 Women
 Men
Women Men College Mturk College Mturk
655 366 289 189 177 110 179
Emotional (e.g., gets worked up easily) 228 (35%) 121 (33%) 107 (36%) 73 (39%) 48 (27%) 40 (36%) 67 (37%)
Clothing (e.g., into fashion) 184 (28%) 103 (28%) 81 (28%) 46 (24%) 53 (30%) 19 (17%) 51 (28%)
Small (e.g., petite) 169 (26%) 99 (27%) 70 (24%) 34 (18%) 39 (22%) 24 (22%) 52 (29%)
Make-up (e.g., wears make-up) 149 (23%) 73 (20%) 76 (26%) 22 (12%)a 47 (27%) 13 (12%) 40 (22%)a
Caring or Kind (e.g., considerate) 141 (21%) 81 (22%) 60 (21%) 42 (22%) 39 (22%) 20 (18%) 40 (22%)
Sensitive or Empathic (e.g., sympathetic) 122(19%) 69 (19%) 53 (18%) 37 (20%) 31 (18%) 14 (13%) 40 (22%)
Good looking (e.g., pretty) 122 (19%) 68 (19%) 54 (19%) 35 (19%) 21 (12%) 16 (15%) 28 (16%)
Concerned about/works on appearance (e.g., cares about appearance, well groomed) 112(17%) 74 (20%) 38 (13%) 51 (27%) 52 (29%) 24 (22%) 57 (32%)
Nurturing (e.g., provides support) 100 (15%) 56 (15%) 44 (15%) 33 (17%) 41 (23%) 11 (10%) 27 (15%)
Long hair (e.g., long hair) 86 (13%) 42 (11%) 44 (15%) 26 (14%) 16 (9%) 16 (15%) 28 (16%)
Thin (e.g., slender) 74 (11%) 42 (11%) 32 (11%) 14 (7%) 21 (12%) 11 (10%) 23 (13%)
Nice or Sweet (e.g., sweet, friendly) 72 (11%) 24 (6%) 48 (17%) 22 (12%) 20 (11%) 13 (12%) 19 (11%)
Weak (e.g., weak, less strong) 69 (10%) 35 (10%)a 34 (12%)a 17 (9%) 7 (4%)b 19 (17%) 29 (16%)b
Quiet (e.g., quiet, soft-spoken) 69 (10%) 45 (12%) 24 (8%) 22 (12%) 23 (13%) 10 (9%) 14 (8%)
Soft (e.g., soft) 66 (10%) 50 (14%) 16 (5%) 14 (7%) 15 (8%) 5 (5%) 10 (6%)
Delicate (e.g., fragile) 65 (10%) 41 (11%) 24 (8%) 10 (6%) 12 (7%) 1 (1%) 17 (9%)
Polite (e.g., well-mannered) 59 (9%) 39 (11%) 20 (7%) 19 (10%) 22 (12%) 6 (5%) 18 (10%)
Smart (e.g., intelligent) 55 (8%) 27 (7%) 28 (10%) 15 (8%) 12 (7%) 12 (11%) 16 (9%)
Gentle (e.g., gentle, does not fight) 51 (8%) 26 (7%) 25 (9%) 17 (9%) 17 (10%) 7 (6%) 5 (3%)
High Voice (e.g., high voice) 50 (8%) 20 (5%) 30 (10%) 13 (7%) 12 (7%) 11 (10%) 12 (7%)
Flirtatious or Sexy (e.g., seductive) 49 (7%) 29 (8%) 20 (7%) 21 (11%) 5 (2%) 12 (12%) 13 (7%)
Dependent (e.g., dependent, helpless) 48 (7%) 25 (7%) 23 (8%) 21 (11%) 18 (10%) 6 (5%) 14 (8%)
Cooks (e.g., cooking) 46 (7%) 34 (9%) 12 (4%) 14 (7%) 16 (9%) 6 (5%) 9 (5%)
Clean or Cleans (e.g., cleans, not dirty) 45 (7%) 30 (8%) 15 (5%) 16 (8%) 13 (7%) 9 (8%) 11 (6%)
Submissive (e.g., yielding, not dominant) 44 (7%) 29 (8%) 15 (5%) 27 (14%) 23 (13%) 6 (5%) 10 (6%)
Loving (e.g., loves) 41 (6%) 28 (8%)a 13 (4%)a 10 (5%) 10 (6%) 11 (10%) 19 (11%)
Shops (e.g., likes shopping) 40 (6%) 22 (6%) 18 (6%) 11 (6%) 17 (10%) 4 (4%) 9 (5%)
Open in a new tab

Table 3.

Masculinity domains

(1) Physical differences related to biological sex (that people have little control over)
 Tall
 Hairy
 Deep Voice (cited in two categories)
(2) Emphasized physical differences (behavior impacts these physical differences)
 Deep voice (cited in two categories)
 Strong (cited in three categories)
 Muscles
 Clothing
(3) Activities and interests focused on the physical body
 Athletic/plays sport
 Sport-related
 Works out/physically fit
 Lifts weights/body builds
 Alcohol consumption
 Sexual
(4) Powerful or oriented toward power
 Powerful
 Financially-able
 Aggressive
 Assertive
 Competitive
 Independent
 Strong (cited in three categories)
(5) Emotion-control or emotionally-limited
 Emotionally-limited
 Confident
 Tough
 Brave
 Strong (cited in three categories)
Other
 Interested in vehicles
 Heterosexual
 Smart
Open in a new tab

Table 4.

Femininity domains

(1) Physical differences related to biological sex (that people have little control over)
 Small
 High Voice (cited in two categories)
(2) Emphasized physical differences (behavior impacts these physical differences)
 Concerned about/works on appearance
 Clothing
 Make-up
 Long hair
 Thin
 Soft
 Good looking
 High Voice (cited in two categories)
 Delicate (cited in two categories)
 Weak (cited in two categories)
 Clean or cleans (cited in two categories)
(3) Oriented toward other people
 Caring or kind
 Sensitive or empathetic
 Nurturing
 Nice or sweet
 Flirtatious
 Loving
 Polite
 Gentle (cited in two categories)
(3) Lacking power
 Submissive
 Dependent
 Quiet
 Weak (cited in two categories)
 Gentle (cited in two categories)
 Delicate (cited in two categories)
Other
 Emotional
 Cooks
 Shops
 Clean or cleans (cited in two categories)
 Smart
Open in a new tab

Fig. 1.

Fig. 1

Open in a new tab

Participants’ conceptions of domains of masculinity and most commonly mentioned codes in these domains

Fig. 2.

Fig. 2

Open in a new tab

Participants’ conceptions of domains of femininity and most commonly mentioned codes in these domains

Conceptions of Masculinity

For masculinity, the most common codes, listed by more than 20% of participants, were strong (listed by 48% of participants), having muscles (32% of participants), athletic/plays sports (25% of participants), and hairy (22%). Moderately common codes, listed by 10%-20% of participants, were emotionally-limited (19%), confident (19%), tough (18%), works out/physically fit (17%), aggressive (16%), sport-related (15%), deep voice (14%), brave (14%), independent (12%), powerful (11%), alcohol consumption (11%), and tall (11%).

Two common codes for masculinity, tall and hairy, are average physical differences related to biological sex that people have little control over. We assigned these codes to a domain we labeled “physical differences related to sex.” Though these codes appear to be purely biological (i.e., independent of social influence), women and men exist on a continuum in regard to these physical characteristics, and thus some women are taller, and hairier in some areas (e.g., legs), than some men. So, although there are average biological sex differences in height and hairiness between men and women, when participants frame these characteristics as masculine and feminine they are emphasizing more of a dichotomy and less of a continuum.

We put some codes focused on physical differences, including muscles and (physical) strength, into a domain we labeled “emphasized physical differences.” There are average physical differences between men and women in regard to the codes in this domain, but people have more control over these physical differences than they do over height and hairiness (before hair removal). For example, men in contemporary U.S. society often enhance their musculature and strength via lifting weights and other types of physical exertion, while women less often do so.

We placed the common code of “deep voice” in both the “physical differences related to sex” and “emphasized physical differences” domains. We did this because although changing the degree to which one’s voice is deep or high is relatively difficult, many people do make efforts to control, and do modify, the degree to which their voices sound deep or high (e.g., [10].

Also related to physical aspects of masculinity, we placed four of the most common codes into a domain we labeled “activities and interests focused on the physical body.” Three of these codes are related to fitness and sport: athletic/plays sport, sport-related (e.g., “likes sport”), and works out/physically fit. Another common code we placed in this category was alcohol consumption.

The fourth domain for masculinity we titled “powerful or oriented toward power.” Four of the most common codes in this category were: powerful, aggressive (which commonly involves using force in pursuit of power), independent (because being independent means being free from outside control and reliance on others), and strong (in terms of physical and social power).

Lastly, we placed five of the most common codes in a domain we titled “emotion-control or emotionally-limited.” These five codes are: emotionally-limited, confident (as this involves controlling emotions related to doubt), brave (as this involves not feeling, or overcoming, the emotion of fear), tough (as this involves controlling emotions related to feeling pain and fatigue), and strong (as strength is sometimes framed as being in control of one’s emotions).

See Table 1 for the percentage of participants who mentioned each code. See Table 3 for the placement of these codes into domains. And, see Fig. 1 for a visual representation of the domains along with the most commonly mentioned codes.

Regarding similarities across groups in our sample, among the 130 comparisons of proportions of ratings of masculinity across groups, 60 were equivalent and an additional 34 were close. Comparing men and women in general, proportions of codes for masculinity were equivalent for 12 of the 26 codes, and close for an additional 6 codes. Significant differences emerged by gender for a few descriptions of masculinity, including being muscular and having a deep voice (more common among women) and being confident and lifting weights (more common among men). Examining the four comparisons of groups by gender and data-source, of the 104 comparisons, 48 were equal and an additional 28 were close. Only three statistical differences emerged; college men and mturk men were more likely to include athleticism than college women and mturk women, respectively, and mturk women were more likely to include assertive than mturk men. Thus, the overall pattern indicated very substantial overlap in descriptors of masculinity from women and men participants in general and across data sources. Comparison data are presented in the supplemental tables.

Conceptions of Femininity

For femininity, the most common codes were emotional (listed by 35% of participants), clothing (28%), small (26%), make-up (23%), and caring/kind (21%). Moderately common codes, listed by 10–20% of participants, included sensitive/empathic (19%), good-looking (19%), concerned about/works on appearance (17%), nurturing (15%), long hair (13%), thin (11%), nice or sweet (11%), weak (10%), quiet (10%), soft (10%), and delicate (10%).

Two of the domains for femininity paralleled the domains for masculinity. First, we placed one of the most common femininity codes, “small,” in the domain “physical differences related to sex.” Although “small” may refer to weight and musculature, and thus be somewhat malleable, we conceived of the code of “small” as parallel to the masculine code of “tall,” and thus view this code as less malleable than other codes related to femininity.

We placed nine of the most common codes for femininity into the category “emphasized physical differences,” which were: concerned about/works on appearance, clothing, make-up, long hair, thin, soft, good looking, delicate, and weak (in terms of physical strength). All of these codes are related to behavior. For example, people can wear make-up or not, wear their hair long or short, and do things to develop their strength or not do so (and thus become physically weak and delicate).

We called a third domain for femininity “oriented toward other people.” We placed four of the most common codes in this domain: nurturing, caring or kind, sensitive or empathetic, and nice or sweet.

We placed three of the most common codes in the domain we called “lacking power.” These codes were: weak (in terms of both social and physical power), delicate (as delicate people are often perceived as physically and emotionally weak), and quiet (so not speaking up and thus not influencing the situation).

Lastly, there was one very common code that we did not place into a domain, which was “emotional.” Although this code may be related to some of the domains mentioned above, it could also reflect emotions such as happiness, sadness, disgust, surprise and anger, which do not necessarily fit with the domains mentioned above.

See Table 2 for the percentage of participants who mentioned each code. See Table 4 for the placement of these codes into domains. And, see Fig. 2 for a visual representation of the domains along with the most commonly mentioned codes.

Regarding similarities of codes among groups of our participants, of the 135 comparisons 72 were equivalent and an additional 41 were close. Comparing women and men in general, proportions of the codes for femininity were equivalent for 15 of the 27 codes, and close for an additional 10 codes. Thus, men and women were remarkably similar in the proportions of most codes listed. Significant differences emerged for the codes weak (more common among men than women) and submissive (more common among women than men). Regarding the four groups divided by gender and data-source, of 108 comparisons, 57 were equal and an additional 31 were close. Only two differences emerged between these groups, with mturk women being more likely than college women to list make-up, and mturk men being more likely than mturk women to list weak. Thus, again, the pattern of overall results suggested substantial overlap, with few significant differences by gender overall and few differences by gender and data-source. Comparison data are presented in the supplemental tables.

Discussion

For the present study, we gathered written qualitative data on U.S. men and women’s self-generated descriptions of femininity and masculinity. Then, we coded these descriptions. We then assessed interconnections among the codes, placing the most common codes into larger domain categories. Below, we discuss our findings, devoting special attention to how our results compare to established measures of gender role ideology and conformity that are commonly used in the contemporary period.

We found strong similarities across gender and sampling frame in terms of the predominance of the codes. In other words, one of our main findings is that characteristics/behaviors believed to be feminine and masculine varied little between our men and women participants and between our college and mturk participants. When considering the small number of gender differences in responses, we see only one possible pattern: It may be that participants are more aware of the socially constructed nature of gender that is associated with their own gender identity, as participants were more apt to name less malleable physical gender characteristics/behaviors not associated with their own gender identity and more behaviorally-malleable gender characteristics/behaviors that are associated with their own gender identity (e.g., for masculine characteristics/behaviors, women participants were more apt to name muscularity while men were more apt to name lifting weights). Future scholarship could explore this possibility.

When we compare our findings to existing measures, we focus on three of the most popular measures of masculinities: the MRNI (the original measure being cited 294 times since 1992 as of August 2019 and the revision being cited 187 times since 2007), the CMNI (the original being cited 1227 times since 2003 and the 46-item abbreviated cited 251 times since 2009), and the GRCS (cited 1150 times since 1986); and two measures of femininities: the CFNI (the original cited 416 times since development in 2005, and the 45-item revision cited 92 times since 2010) and FIS (cited 95 times since 2007).

Similarities and Differences in Conceptions of Femininity and Masculinity

Some common descriptions of masculinity articulated by our participants focused on physical sex differences over which individuals have little control, such as being tall, as well as physical differences which can be accentuated such as muscularity, physical strength, and presence of hair. Muscle size has both biological (i.e., greater testosterone production among males facilitates the development of larger muscles) and social (e.g., greater socialization of men into muscularity-enhancing behaviors such as weightlifting) causality (e.g., [50]). Similarly, while the presence of a deeper voice, and body and facial hair, are linked, in part, to testosterone levels, these bodily features are also influenced by social forces (e.g., [10, 56]). Although our participants named a variety of physical gender differences, none were highly dichotomous biological sex differences. That is, participants did not list more dichotomous physical features as masculine (e.g., having a penis, absence of breasts capable of lactation).

Whereas most theories of the social construction of gender de-emphasize physical differences, our participants emphasized physical differences in their characterizations of masculinity. It is tempting to dismiss the two physical domains of masculinity that emerged from our findings as simply related to biological sex and unrelated to social gender differences. But, all of the commonly mentioned physical differences exist on a continuum, as some women exhibit more of these characteristics than some men (e.g., some women are taller than some men, stronger than some men, etc.), and many can be—and commonly are—enhanced or minimized via behavior (e.g., lifting weights, singing falsetto). As such, though some domains named by our participants were focused on physical differences, these were nevertheless constructions of gender [13].

Some scholars have suggested that as increasing numbers of activities, dispositions and the like have become acceptable for women in the U.S., there are fewer activities, dispositions and the like that are available to convey masculinity, and in response physical differences between men and women have increasingly been emphasized to convey masculinity [14, 35, 37]. Our findings support this theoretical premise. Yet, the most commonly used measures of masculinities have focused on gender roles as cognitive phenomena and not on physical manifestations of gender expression. Thus, that physical aspects of gender formed principal aspects of participants’ descriptions of masculinity runs contrary to the thematic content of popular measures of masculinity, which do not include subscales that focus on physical manifestations of gender expression. Although a subscale related to physical strength was present in the original development of the CMNI, it dropped out of the factor solution [32]. Two scales that are less commonly used in the contemporary period, [11, 15], do have physical characteristics subscales.

Related to perceptions that masculinity was associated with physical bodies, our participants’ perceived activities and interests focused on the physical body as central to their conceptions of masculinity. Sports and physical fitness were the most common aspects of this conception, with another especially common physical activity being alcohol consumption. The most common (recently) used measures of masculinity do not focus on activities and interests, including activities and interests related to the physical body. Contrary to scholars who emphasize a connection between athleticism and masculinity (e.g., [36]), as well as our participants’ responses, the commonly utilized measures of masculinity that we studied do not include sports and fitness, or alcohol use, as aspects of masculine gender roles. Participation in sports is not an item in any of the major masculinity measures that we studied. The MRNI has an item reflecting passive engagement in sport—although only in juxtaposition to a more feminized alternative (“Men should watch football instead of soap operas”). Alcohol use items do not appear in any of these measures of masculinity, though relations between masculinity and alcohol use are often studied (e.g., [49]).

Our participants’ descriptions of masculinity included the domain of power and orientation toward power, which involved such characteristics as powerful, aggressive, strong, and independent. Although being strong is not an item on the commonly used measures of masculinity that we studied, aspects of social dominance and power are present in the MRNI (dominance), CMNI (power over women, winning), and GRCS (success, power, and competition). Aggression is also present on the CMNI (violence), and independence is present in the MRNI (self-reliance through mechanical skills) and CMNI (self-reliance). Thus, our domain focused on power and power orientation was present in established measures of masculinities.

Lastly, our participants conveyed that masculinity involves emotional control and limited emotions. Restricted emotionality is present in all three measures of masculinity listed above. It is notable that only the MRNI contains a subscale on toughness. Concepts such as bravery and confidence, while not included in the three measures of masculinity, have appeared in scholarly work on masculinity, both in terms of being potentially problematic manifestations of masculinity that serve to deny struggles (e.g., [34]) and aspects of positive masculinity (e.g., [20]). Further work is needed to delineate traits such as bravery or confidence as acknowledging and overcoming emotions such as fear and doubt, versus denying those emotions, and how such aspects of functioning may manifest in and affect men and women as well as affect society.

As with masculinity, some of our participants’ descriptions of femininity focused on physical characteristics. Participants noted fewer less-malleable physical characteristics and many malleable physical characteristics, such as wearing make-up, having long hair, and being delicate. In fact, concerning oneself with one’s appearance and working to mold one’s appearance (i.e., constructing a feminized body) was perceived by our participants as central to femininity. Two of the common codes in our findings, being small and wearing make-up, are captured in the existing measures of feminine gender role ideology and conformity that we studied. The FIS contains a Stereotypic Images and Activities subscale, with items referencing having a petite body, whereas the CFNI contains a subscale related to investment in appearance, with items reflecting spending time on make-up and hair. The specific content of the long hair code found in our study is not present in either femininity measure, though other research has linked long hair to perception of femininity (e.g., [54]). Thinness emerged in the present study, and is a facet of femininity in the CFNI. Codes related to delicate, soft, and physically weak do not have direct parallels in the two existing measures of femininity.

That the existing measures of femininity that we studied do reflect a focus on physical expression (although not all of the forms of physical expression that we found) stands in contrast to the measures of masculinity that we studied, which do not include aspects of physical gender expression. Perhaps the difference in the emphasis on physical gender role indicators between measures of femininity and masculinity reflects the large body of literature on women’s body ideals, efforts to attain these ideals, and problems associated with these ideals/efforts (e.g., [38]), in contrast to the more nascent growing body of literature on men’s body ideals, efforts to attain these ideals, and problems associated with these ideals/efforts (e.g., [47]). Alternatively, these differential emphases may reflect a narrative of women’s deficiencies; in this case, that women and not men experience body-rated concerns and potential pathology [24]. Lastly, the greater focus on physical characteristics of femininity may reflect a historical shift toward more emphasis on physical ideals of femininity in U.S. society [7].

Another domain deemed central to femininity by our participants is being oriented toward other people, captured in our codes of nurturing, caring/kind, sensitive/empathetic, and sweet/nice. The CFNI contains numerous parallels to these codes, including items in the Nice in Relationships and Care for Children subscales, and other items. These facets of femininity are assessed in the FIS only tangentially in some items pertaining to emotionality and traditional roles (e.g., “It is expected that a single woman is less fulfilled than a married woman”). Our findings and the CFNI item content are consistent with feminine socialization, which emphasizes relationship development (e.g., [31, 33]).

Our participants indicated that femininity is associated with lack of power. Consistent with this finding, the FIS contains a subscale focused on Dependency/Deference (items related to not being financially independent and not being competitive), and the CFNI contains a Modesty subscale (items reflecting reluctance to be seen as prideful or boastful). While the domain of lack of power was reflected in the two existing femininity measures in the manners just discussed, our participants mentioned other associations between femininity and lack of power, including being weak, delicate, and quiet, which were not included in the CFNI or FIS.

The last finding we will discuss here is that femininity was perceived by our participants to be associated with being emotional. Although some aspects of being emotional are associated with femininity domains already discussed, such as being oriented toward other people, many emotions do not fit within these domains so we left the characteristic of being emotional as a stand-alone code. A subscale parallel to this facet of femininity is not present in the CFNI, although this facet of femininity is tangentially present in some FIS items in the emotionality and traditional roles subscale (e.g., “It is expected that women will have a hard time handling stress without getting emotional”).

In the discussion above, we identified some domains of femininity and masculinity that were commonly mentioned by our participants that were absent or minimally present in commonly used existing measures of gender conformity and ideology that we studied. In addition, the existing measures of gender conformity and ideology that we studied included some characteristics that were not considered central to femininity and masculinity by our participants. In particular, the CFNI contains a subscale focused on sexual fidelity, a characteristic that was not seen by our participants as central to femininity. In contrast, seven percent of our participants mentioned that being flirty/sexy was a characteristic of femininity. This finding may be evidence of a shift away from cultural mores emphasizing chastity for women (e.g., [9]), at least for this predominantly White, young adult U.S. sample. Similarly, risk-taking is present in the CMNI but was not mentioned by a significant number of participants in the present sample. It is possible that risk-taking is a relatively situation-dependent manifestation of masculinity, and as participants were sitting at their computers taking the present survey they may not have had risk-taking thoughts particularly primed (e.g., [42]). Power over women is also a subscale of the CMNI, and although the more general concept of power was mentioned by the participants in our study, the power was not named to be specifically over women. Although means for risk-taking in studies using the CMNI and its short forms with men have tended toward the mid-range of the response scale, means for power over women have tended to be low [45, 46]. However, studies demonstrate that many men in the U.S. engage in power-related behaviors toward women, such as interrupting women in conversations (e.g., [25]). Further investigation is warranted into why such behavior is observed by many researchers, and yet did not arise in the present study and receives generally low endorsement in studies using the CMNI and its short forms. Perhaps men’s power over women is often taken for granted and thus goes unnoticed in the U.S. Or, perhaps the feminist movement in the U.S. has rendered public recognition of this power less acceptable.

Scholars have long observed that many men and women in the U.S. view gender as dichotomous (e.g., [4]). It is important to note the apparently dichotomous nature of many of the feminine and masculine characteristics discovered in this study. Some physical characteristics are framed dichotomously, including strong versus weak/delicate and small/thin versus tall/muscular. The domain of femininity as lacking power contrasts in a dichotomous manner with the domain of masculinity as powerful or oriented toward securing power. Conceiving of femininity as emotional stands in stark opposition to conceiving of masculinity as emotionally-limited and controlling emotions. Some specific characteristics are also dichotomous, such as dependent versus independent and nurturing versus competitive. Yet, not all of our findings reflect dichotomous thinking about gender roles. For example, femininity was not portrayed as abstaining from alcohol consumption or fitness activities, and masculinity was not portrayed as being unkind and unconcerned about appearance. Further, named characteristics were not absolutely dichotomous characteristics (e.g., chromosome pairs or the presence or absence of specific anatomy), but rather were continuous or malleable characteristics. This leads us to the conclusion that our participants most often framed gender in a dichotomous way, but not in all cases.

A critical question for scholars who study gender is how the characteristics of femininity and masculinity articulated by our participants relate to gender inequality. Toward that end, it is important to note that many of these characteristics are aligned with [21] theory of benevolent sexism (which is part of their theory of ambivalent stereotypes), according to which gender inequality is legitimized by perceptions that men are competent and women are warm. More specifically, in our study, masculinity is associated with power and limited emotions, which evidences high competence and low warmth, whereas femininity is associated with powerlessness, emotionality and orientation toward others, which evidences low competence and high warmth. And, because warmth has a positive connotation, the inequality is more difficult to detect and address [21]. More narrowly, although many purport to value empathy among women, the societal maldistribution of empathy between men and women reinforces gender inequality [31]. Further, some of the physical aspects of gender identified by our participants are often used to legitimate gender inequality, such as when physical strength is deemed important to jobs (e.g., [39]). Overall, many of the dimensions, as well as specific characteristics, of femininity and masculinity revealed in our study are utilized to generate, maintain, and/or legitimate gender inequality.

Limitations and Future Research Directions

The results of the present study must be interpreted in light of its limitations. The following aspects of our question prompts may have moved our participants to think of some aspects of femininity and masculinity rather than other aspects these constructs: the wording of the prompts (i.e., “characteristics” and “behaviors”), asking participants about both masculinity and femininity, and inquiring about these two constructs in the same order. Other limitations of our study relate to our sample. Our sample includes only people from the U.S., and was not representative of people in U.S. society. Our sample included only younger adults, and conceptualizations of gender likely differ by age cohort and by experiences related to age (e.g., raising children, retiring). Related to this point, while we inquired about characteristics/behaviors our participants thought other young adults believed were masculine/feminine, we do not know if our participants believed these characteristics/behaviors were applicable to people of all ages or only to particular age groups. Our sample was exclusively heterosexual and predominately White; and important group and intersectional similarities and differences exist and are critical to understanding the construction of gender. Further, it is likely that occupation, education level, household income, and many other factors (e.g., specific socialization experiences) impact perceptions of gender, and our data did not allow us to examine these differences. Finally, the nature of the online data collection did not allow us to probe participants for clarification on responses, which is a practice often used in qualitative research to improve data quality.

Despite these limitations, information derived from the present study can be used to inform further development and refinement of the study of gender in the U.S., both in qualitative and quantitative efforts. Our results both resemble and diverge from popular measurements of gender utilized in the U.S. Consistent with the popular measurements of gender we utilized for comparison purposes, participants in our study described some cognitive aspects of femininity and masculinity. But our participants also described physical, activity and interest domains. The latter domains are absent or minimal in most popular measures of gender role ideology and conformity. We recommend that scholars constructing new measures of gender roles, or those modifying existing measures of gender roles, consider adding measures of physical, activity, and interest domains.

Because gender roles and ideologies change over time, we recommend that scholars use currently existing scales to determine which characteristics and subscales in these measures rise and fall in societal importance over time. Further, we hope that scholars in the United States who create new scales to measure femininity and masculinity, as well as scholars who modify existing scales, are able to use our inductive qualitative findings in their work. It seems necessary to periodically use information drawn from large-scale qualitative studies such as ours to examine existing scales to see if they are capturing contemporary conceptions of femininity and masculinity.

Conclusions

Scholars of gender must take into account that gender roles and ideologies vary in different societies, in different social contexts within societies, and among different demographic groups. Equally important, it is critical to recognize that gender roles and ideologies within societies change over time. In fact, propelled by the feminist movement, there has been a major transformation in the lives of women in the United States since the 1970s that continues up to this day. It would be surprising if this transformation had no effects on gender role ideologies and conformities. Thus, it is important to use existing quantitative measures of gender to examine different demographic groups, social contexts, and changes in gender over time. Further, inductive research focused on gender roles and ideologies is useful to determine whether existing quantitative measures need to be modified. We hope that our study contributes toward this end.

Supplementary Material

Supplementary Material

Footnotes

Electronic supplementary material The online version of this article (https://doi.org/10.1007/s12147-019-09246-y) contains supplementary material, which is available to authorized users.

References

  • 1.Auster CJ, & Ohm SC (2000). Masculinity and femininity in contemporary American society: A reevaluation using the bem sex-role inventory. Sex Roles, 43, 499–528. 10.1023/A:1007119516728. [DOI] [Google Scholar]
  • 2.Bem SL (1974). The measurement of psychological androgyny. Journal of Consulting and Clinical Psychology, 42, 155–162. 10.1037/h0036215. [DOI] [PubMed] [Google Scholar]
  • 3.Berger A, & Krahe B (2013). Negative attributes are gendered too: Conceptualizing and measuring positive and negative facets of sex-role identity. European Journal of Social Psychology, 43, 516–531. 10.1002/ejsp.1970. [DOI] [Google Scholar]
  • 4.Bosson JK, & Michniewicz KS (2013). Gender dichotomization at the level of ingroup identity: What it is, and why men use it more than women. Journal of Personality and Social Psychology, 105, 425. 10.1037/a0033126. [DOI] [PubMed] [Google Scholar]
  • 5.Brooks C, & Bolzendahl C (2004). The transformation of US gender role attitudes: cohort replacement, social-structural change, and ideological learning. Social Science Research, 33(1), 106–133. 10.1016/S0049-089X(03)00041-3. [DOI] [Google Scholar]
  • 6.Brownlie EB (2006). Young adults’ constructions of gender conformity and nonconformity: A Q methodological study. Feminism & Psychology, 16(3), 289–306. 10.1177/0959353506067848. [DOI] [Google Scholar]
  • 7.Brumberg JJ (1997). The big project: An intimate history of American girls. New York: Random House. [Google Scholar]
  • 8.Buhrmester M, Kwang T, & Gosling SD (2011). Amazon’s mechanical turk: A new source of inexpensive, yet high-quality, data? Perspectives on Psychological Science, 6(1), 3–5. 10.1177/1745691610393980. [DOI] [PubMed] [Google Scholar]
  • 9.Buss DM, Shackelford TK, Kirkpatrick LA, & Larsen RJ (2001). A half century of mate preferences: The cultural evolution of values. Journal of Marriage and Family, 63, 491–503. 10.1111/j.1741-3737.2001.00491.x. [DOI] [Google Scholar]
  • 10.Cartei V, Cowles HW, & Reby D (2014). Spontaneous voice gender imitation in adult speakers. PLoS ONE, 7, 1–12. 10.1371/journal.pone.0031353. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Cejka MA, & Eagly AH (1999). Gender-stereotypic images of occupations correspond to the sex segregation of employment. Personality and Social Psychology Bulletin, 25, 413–423. 10.1177/0146167299025004002. [DOI] [Google Scholar]
  • 12.Connell RW (2005). Masculinities. Los Angeles: University of California Press. [Google Scholar]
  • 13.Crawley SL, Foley LJ, & Shehan CL (2008). Gendering bodies. Lantham: Rowman & Littlefield. [Google Scholar]
  • 14.Davis LR (1997). The swimsuit issue and sport: Hegemonic masculinity in Sports Illustrated. Albany: State University of New York. [Google Scholar]
  • 15.Deaux K, & Lewis LL (1983). Assessment of gender stereotypes: Methodology and components. Psychological Documents, 13, 1–23. [Google Scholar]
  • 16.Diekman AB, & Eagly AH (2000). Stereotypes as dynamic constructs: Women and men of the past, present, and future. Personality and Social Psychology Bulletin, 26, 1171–1188. 10.1177/0146167200262001. [DOI] [Google Scholar]
  • 17.Diekman AB, Eagly AH, Mladinic A, & Ferreira MC (2005). Dynamic stereotypes about women and men in Latin America and the United States. Journal of Cross-Cultural Psychology, 36, 209–226. 10.1177/0022022104272902. [DOI] [Google Scholar]
  • 18.DeVellis RF (2016). Scale development: Theory and applications. Thousand Oaks: SAGE Publications. [Google Scholar]
  • 19.Donnelly K, & Twenge JM (2017). Masculine and feminine traits on the Bem Sex-Role Inventory, 1993–2012: A cross-temporal meta-analysis. Sex Roles, 76, 556–565. 10.1007/s11199-016-0625-y. [DOI] [Google Scholar]
  • 20.Englar-Carlson M, & Kiselica MS (2013). Affirming the strengths in men: A positive masculinity approach to assisting male clients. Journal of Counseling & Development, 91, 399–409. 10.1002/j.1556-6676.2013.00111.x. [DOI] [Google Scholar]
  • 21.Glick P, & Fiske ST (2001). Ambivalent stereotypes as legitimizing ideologies: Differentiating paternalistic and envious prejudice. The psychology of legitimacy: Emerging perspectives on ideology, justice, and intergroup relations (pp. 278–306). New York: Cambridge University Press. [Google Scholar]
  • 22.Haines EL, Deaux K, & Lofaro N (2016). The times they are a-changing…or are they not? A comparison of gender stereotypes, 1983–2014. Psychology of Women Quarterly, 40, 353–363. 10.1177/0361684316634081. [DOI] [Google Scholar]
  • 23.Helmreich RL, Spence JT, & Wilhelm JA (1981). A psychometric analysis of the personal attributes questionnaire. Sex Roles, 7, 1097–1108. 10.1007/BF00287587. [DOI] [Google Scholar]
  • 24.Hyde JS (1994). Should psychologists study gender differences? Yes, with some guidelines. Feminism & Psychology, 4(4), 507–512. 10.1177/0959353594044004. [DOI] [Google Scholar]
  • 25.Kiesling S (2007). Men, masculinities, and language. Language and Linguistics Compass, 1, 653–673. 10.1111/j.1749-818X.2007.00035.x. [DOI] [Google Scholar]
  • 26.Landis JR, & Koch GG (1977). The measurement of observer agreement for categorical data. Biometrics, 33, 159–174. 10.2307/2529310. [DOI] [PubMed] [Google Scholar]
  • 27.Levant RF (2011). Research in the psychology of men and masculinity using the gender role strain paradigm as a framework. American Psychologist, 66(8), 765–776. 10.1037/a0025034. [DOI] [PubMed] [Google Scholar]
  • 28.Levant RF, Hirsch LS, Celentano E, Cozza TM, Hill S, MacEachern M, et al. (1992). The male role: An investigation of contemporary norms. Journal of Mental Health Counseling, 14, 325–337. [Google Scholar]
  • 29.Levant RF, Richmond K, Cook S, House AT, & Aupont M (2007). The Femininity Ideology Scale: Factor structure, reliability, convergent and discriminant validity, and social contextual variation. Sex Roles, 57(5–6), 373–383. 10.1007/s11199-007-9258-5. [DOI] [Google Scholar]
  • 30.Levant RF, Smalley KB, Aupont M, House AT, Richmond K, & Noronha D (2007). Initial validation of the male role norms inventory-revised (MRNI-R). The Journal of Men’s Studies, 15, 83–100. 10.3149/jms.1501.83. [DOI] [Google Scholar]
  • 31.Lobb A (2013). The agony and the empathy: The ambivalence of empathy in feminist psychology. Feminism & Psychology, 23, 426–441. 10.1177/0959353513497556. [DOI] [Google Scholar]
  • 32.Mahalik JR, Locke BD, Ludlow LH, Diemer MA, Scott RPJ, Gottfried M, et al. (2003). Development of the conformity to masculine norms inventory. Psychology of Men & Masculinity, 4, 3–25. 10.1037/1524-9220.4.1.3. [DOI] [Google Scholar]
  • 33.Mahalik JR, Morray EB, Coonerty-Femiano A, Ludlow LH, Slattery SM, & Smiler A (2005). Development of the conformity to feminine norms inventory. Sex Roles, 52(7–8), 417–435. 10.1007/s11199-005-3709-7. [DOI] [Google Scholar]
  • 34.Melzer S (2018). Manhood impossible: Men’s struggles to control and transform their bodies and work. New Brunswick: Rutgers University. [Google Scholar]
  • 35.Messner MA (1992). Power at play: Sports and the problem of masculinity. Boston: Beacon. [Google Scholar]
  • 36.Messner MA, & Sabo DF (1990). Sport, men, and the gender order: Critical feminist perspectives. Champaign: Human Kinetics. [Google Scholar]
  • 37.Mishkind ME, Rodin J, Silberstein LR, & Streigel-Moored RH (1986). The embodiment of masculinity: Cultural, psychological, and behavioral dimensions. American Behavioral Scientist, 29, 545–562. 10.1177/000276486029005004. [DOI] [Google Scholar]
  • 38.Moradi B, & Huang Y (2008). Objectification theory and psychology of women: A decade of advances and future directions. Psychology of Women Quarterly, 32, 377–398. 10.1111/j.1471-6402.2008.00452.x. [DOI] [Google Scholar]
  • 39.Ness K (2012). Constructing masculinity in the building trades: ‘Most jobs in the construction industry can be done by women’. Gender, Work & Organization, 19(6), 654–676. 10.1111/j.1468-0432.2010.00551.x. [DOI] [Google Scholar]
  • 40.Norona JC, Preddy TM, & Welsh DP (2016). How gender shapes emerging adulthood. In Arnett JJ & Arnett JJ (Eds.), The Oxford handbook of emerging adulthood (pp. 62–86). New York: Oxford University Press. [Google Scholar]
  • 41.Paolacci G, Chandler J, & Ipeirotis PG (2010). Running experiments on Amazon mechanical turk. Judgment and Decision Making, 5, 411–419. [Google Scholar]
  • 42.Parent MC, Kalenkoski C, & Cardella E (2018). Risky business: Precarious manhood and financial investment decisions. Psychology of Men & Masculinity, 19, 195–202. 10.1037/men0000089. [DOI] [Google Scholar]
  • 43.Parent MC, & Moradi B (2009). Confirmatory factor analysis of the conformity to masculine norms inventory and development of the conformity to masculine norms inventory-46. Psychology of Men & Masculinity, 10, 175–189. 10.1037/a0015481. [DOI] [Google Scholar]
  • 44.Parent MC, & Moradi B (2010). Confirmatory factor analysis of the Conformity to Feminine Norms Inventory and development of an abbreviated version: The CFNI-45. Psychology of Women Quarterly, 34, 97–109. 10.1111/j.1471-6402.2009.01545.x. [DOI] [Google Scholar]
  • 45.Parent MC, & Moradi B (2011). An abbreviated tool for assessing conformity to masculine norms: Psychometric properties of the conformity to masculine norms inventory-46. Psychology of Men & Masculinity, 12, 339–353. 10.1037/a0021904. [DOI] [Google Scholar]
  • 46.Parent MC, Moradi B, Rummell CM, & Tokar DM (2011). Evidence of construct distinctiveness for conformity to masculine norms. Psychology of Men & Masculinity, 12, 354–367. 10.1037/a0023837. [DOI] [Google Scholar]
  • 47.Parent MC, Schwartz EN, & Bradstreet TC (2016). Men’s body image. In Wong YJ & Wester SR (Eds.), APA handbook of men and masculinities (pp. 615–626). Washington DC: American Psychological Association. [Google Scholar]
  • 48.Patton MQ (2002). Qualitative research and evaluation methods. Thousand Oaks: Sage. [Google Scholar]
  • 49.Peralta RL (2007). College alcohol use and the embodiment of hegemonic masculinity among European American men. Sex Roles, 56, 741–756. [Google Scholar]
  • 50.Ricciardelli LA, McCabe MP, & Ridge D (2006). The construction of the adolescent male body through sport. Journal of Health Psychology, 11(4), 577–587. 10.1177/1359105306065018. [DOI] [PubMed] [Google Scholar]
  • 51.Steiger JH (2004). Beyond the F test: Effect size confidence intervals and tests of close fit in the analysis of variance and contrast analysis. Psychological Methods, 9, 164–182. 10.1037/1082-989X.9.2.164. [DOI] [PubMed] [Google Scholar]
  • 52.Strauss A, & Corbin J (1998). Basics of qualitative research: Techniques and procedures for developing grounded theory. Thousand Oaks: Sage. [Google Scholar]
  • 53.Strough J, Leszczynski JP, Neely TA, Flinn JA, & Margrett J (2007). From adolescence to later adulthood: Femininity, masculinity, and androgyny in six age groups. Sex Roles, 57, 385–396. 10.1007/s11199-007-9282-5. [DOI] [Google Scholar]
  • 54.Synnott A (1987). Shame and glory: A sociology of hair. The British Journal of Sociology, 38, 381–413. 10.2307/590695. [DOI] [Google Scholar]
  • 55.Thompson EH Jr., & Bennett KM (2015). Measurement of masculinity ideologies: A (critical) review. Psychology of Men & Masculinity, 16, 115–133. 10.1037/a0038609. [DOI] [Google Scholar]
  • 56.Toerien M, & Wilkinson S (2003). Gender and body hair: Constructing the feminine woman. Women’s Studies International Forum, 26(4), 333–344. 10.1016/S0277-5395(03)00078-5. [DOI] [Google Scholar]
  • 57.Tolman DL, & Porche MV (2000). The Adolescent femininity ideology scale: Development and validation of a new measure for girls. Psychology of Women Quarterly, 24, 365–376. 10.1111/j.1471-6402.2000.tb00219.x. [DOI] [Google Scholar]
  • 58.Twenge JM (1997). Changes in masculine and feminine traits over time: A meta-analysis. Sex Roles, 36, 305–325. 10.1007/BF02766650. [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material
NIHMS1913572-supplement-Supplementary_Material.xlsx (22.3KB, xlsx)

RESOURCES