1. Introduction
This paper investigates the relationship between parental attitudes toward gender expression and adolescent gender nonconformity. At birth, infants are assigned a sex based on their genitalia. Sex assigned at birth can be congruent or incongruent with gender identity, “a deeply felt, inherent sense of being a boy, a man, or a male; a girl, a woman, or a female; or an alternative gender (e.g., genderqueer, gender nonconforming, gender-neutral)” (American Psychological Association, 2015). A child’s gender identity typically emerges between the ages of three and five (Kohlberg, 1966; Ruble, Taylor, Cyphers, Greulich, Lurye, & Shrout, 2007) and is often exhibited through the child’s gender expression (i.e., how a person publicly presents his or her gender including through the dress and behavior). Although an individual’s sex assigned at birth, gender identity, and gender expression are correlated, two or all three of these can be misaligned. When a person’s gender expression is different from what is stereotypically associated with their sex assigned at birth, these behaviors are described as gender nonconforming behaviors. Even among cisgender children (i.e., children whose sex assigned at birth and gender expression aligned), gender-nonconforming behaviors and interests are common, with one study showing that 23 percent of boys and 39 percent of girls exhibit gender-atypical behaviors to some degree (Sandberg, Meyer-Bahlburg, Ehrhardt, & Yager, 1993).
It is well known that socialization plays an influential role in adolescent development, including the development of gender identity (Blakemore, 2008). Hearing comments from adults and peers such as “don’t be so girly” or “why are you always so masculine” put pressure on adolescents to conform to gender stereotypes. Even implicit messaging such as having the father in a household always drive the car, and the mother in the household always cook can influence adolescent gender development, including gender expression (Mesman & Groeneveld, 2018). In addition, this pressure can affect psychosocial outcomes for children and adolescents, with consequences persisting through adulthood. Previous research indicates that adolescents who engage in more consistent gender-nonconforming behaviors are more at risk for psychopathological development compared to adolescents who engage in more gender-conforming behaviors (Roberts, Rosario, Slopen, Calzo, & Austin, 2013; D’Augelli et al., 2006). This risk disproportionately affects sexual and gender minority youths, as they tend to express more gender-nonconforming traits (Bailey & Zucker, 1995; Rieger et al., 2008). Research on parental acceptance-rejection theory has shown that children’s perceptions of parental warmth correlate with psychological adjustment and personality dispositions (Khaleque, 2013). In particular, this research has indicated that parental acceptance of atypical gender expression safeguards gender-nonconforming adolescents, while parental rejection leads to increased risk for psychopathological development (Roberts et al., 2013; D’Augelli, Grossman, & Starks, 2006; van Beusekom, Bos, Overbeek, & Sandfort, 2015).
Parental influence on the gender expression of young children has been shown to differ based on the sex of the child. In particular, the degree of gender-nonconformity among female children does not impact the degree to which parents try to change the child’s behavior, while among male children, the more gender-nonconforming the child is, the more the parents try to change the child’s behavior (Kane, 2006; Spivey, Huebner, & Diamond, 2018). Fathers of gender-nonconforming children especially are more likely to try to change their male children’s behavior than their female children’s behavior (Kane, 2006).However, these studies have been limited in two ways. First, they have asked parents about their attitudes but not asked adolescents about their perspectives of their parents’ attitudes. Second, they have asked parents to reflect on how they responded to their children before they reached adolescence rather than asking them to reflect on their current attitudes and behaviors. The current study addresses these limitations by examining current attitudes and behaviors of parents of adolescents and of the adolescents themselves. In particular, this study explores how parents’ current attitudes, as evaluated by both parents and adolescents, relate to the adolescents’ gender expression and whether this relationship varies depending on the adolescent’s sex assigned at birth.
The current era is seeing much more widespread acceptance of diversity of gender identity and expression, with more people understanding that gender is socially constructed. It remains to be seen how this cultural shift will affect the ways gender is expressed. Therefore, this study aims to contribute to the literature on determinants of gender expression, both as a basic science question as well as in its possible applications within clinical psychology.
2. Methods
2.1. Participants
Adolescents and at least one of their parents were recruited for this study through convenience sampling. Participants included 33 parents and thirty adolescents. Given that for all but three adolescents, only one parent chose to participate, the second parent was eliminated from the study in these three cases, resulting in 30 dyads. Additionally, two dyads were discarded due to incomplete questionnaires. Therefore, analyses were conducted on a total of 28 dyads.
The parents included thirteen cisgender fathers and fifteen cisgender mothers. Parents ranged in age from 35 to 55 (M = 48.38, SD = 4.37). Fifteen of the adolescents were assigned male at birth and thirteen were assigned female at birth, and all were cisgender. Adolescents ranged in age from twelve to seventeen years (M = 14.83, SD = 1.72). The sample of participants was racially relatively heterogeneous. Of the parents, 39.3% identified as White (non-Hispanic), 3.6% Black, 7.1% Hispanic or Latino, 32.1% Asian or Pacific Islander, 3.6% Native American, and 14.3% mixed race. Information about parents’ religious affiliation, annual household income, and education, as well as information about adolescents’ religious affiliation is presented in Table 1.
2.2. Materials, Procedure, and Scoring
Participants completed a survey through Google Forms. The survey began with a consent form and instructions to complete the survey privately. The survey consisted of five sets of questions drawn from four scales. Parents completed the first four sets of questions drawn from three different scales: the Gender Role Stereotypes Scale (Mills, Culbertson, Huffman, & Connell, 2012), the Personal
![]()
Table 1. Demographic characteristics of parent participants.
Attributes Questionnaire (Spence, Helmreich, & Stapp, 1975), and two sets of questions created as a modified version of the Gender Identity Questionnaire (Johnson, Bradley, Birkenfeld-Adams, Kuksis, Maing, Mitchell, & Zucker, 2004). These questions evaluated parents’ views on gender roles, parents’ gendered traits, and parent views about and responses to gender nonconformity in their children. Adolescents completed one set of questions modified from the Recalled Childhood Gender Identity/Gender Role Questionnaire (Zucker et al., 2006). These questions measured how often the adolescent engaged in gender-non-conforming behaviors.
The first set of questions that parent participants completed was from the Gender Role Stereotypes Scale (GRSS). The GRSS is an eight-item, two-dimensional (female and male stereotypes), three-point scale that assesses whether the survey participant has more traditional or egalitarian views on gender roles. Unlike other scales that also assess views on gender roles, the GRSS assesses a participant’s views on gender roles based upon current gender stereotypes rather than gender stereotypes from decades prior. As a result, the GRSS provides a more modern assessment of gender role attitudes compared to previous scales (Mills-Koonce et al., 2018). Participants were asked to determine whether a man or woman should complete a certain household chore, or whether both a man and woman should share responsibility for the chore. For example, participants were asked whether a man or woman should “handle financial matters, such as paying the bills’’ or if a man and woman should share responsibility. For each participant, responses were totaled to create a gender role total score, with higher scores indicating more traditional views about gender roles. Each participant’s gender role total score was averaged, and Z-scores were calculated for each participant to create gender role Z-scores.
The second set of questions that parent participants completed was from the Personal Attributes Questionnaire (PAQ). The PAQ assesses gendered traits among parent participants by evaluating gendered (i.e., traditionally masculine or feminine) attributes on a five-point scale (Spence, Helmreich, & Stapp, 1975). Male participants with a higher masculinity subscale score and female participants with a higher femininity subscale score tend to hold traits that are more traditionally aligned with their gender, whereas male participants with a lower masculinity subscale score and female participants with a lower femininity subscale score tend to hold traits that are less traditionally aligned with their gender. Each participant’s gendered trait total score was averaged, and Z-scores were calculated for each participant to create gendered trait Z-scores.
The third set of questions that parent participants completed was a modified version of the Gender Identity Questionnaire (GIQ). The GIQ was used to assess parent views on gender-nonconformity within their own children. The questionnaire was originally designed to measure gender-atypical behaviors, traits, and interests exhibited by children. For example, the questionnaire asks the frequency with which the child plays with toys that are associated with the opposite gender, such as Barbie dolls for boys and G.I. Joes for girls (Johnson et al., 2004). However, in order to assess parent views on gender-nonconformity within their children, the questions were altered so that parents were asked how negatively or positively they viewed their children engaging in gender-non-conforming behaviors, having gender-nonconforming traits, or having gender-nonconforming interests. For instance, participants were asked how negatively or positively they viewed their son playing with a Barbie doll, or their daughter playing with a G.I. Joe as a child. Higher scores indicate that the participant views gender-nonconformity within their child more negatively, while lower scores indicate that the participant views gender-nonconformity within their child more positively. For each participant, the scores of all the questions were totaled to create a gender-nonconformity viewpoint (GNV) total score. Each participant’s GNV total score was averaged, and Z-scores were calculated for each participant to create GNV Z-scores.
The fourth and final set of questions that parent participants completed was also modified from the GIQ. These questions assessed parent responses to gender-nonconformity within their own children by asking how frequently the parent would engage in behavior to discourage gender-nonconforming behaviors, traits, or interests within their child. For example, participants were asked how frequently they would say or do anything to change “their son playing with a Barbie doll, or their daughter playing with a G.I. Joe as a child”. Higher scores indicate that the participant would change gender-nonconforming behavior or traits within their child more frequently. For each participant, the scores of all the questions were totaled to create a gender-nonconformity response (GNR) total score. Each participant’s GNR total score was averaged, and Z-scores were calculated for each participant to create GNR Z-scores.
Adolescents completed a set of modified questions from the Recalled Childhood Gender Identity/Gender Role Questionnaire (Zucker et al., 2006). This questionnaire was initially designed to measure recalled gender-type behavior during childhood (e.g., how often an adolescent remembered playing with Barbies as a child). However, the questions were modified to evaluate the frequency with which participants currently engage in gender-nonconforming behaviors. For example, participants were asked how often they currently use cosmetic products. Higher scores indicate greater engagement in behaviors that are traditionally aligned with a particular gender (such as a girl using makeup) and less engagement in behaviors that are gender-nonconforming (such as a boy using make-up). Scores were totaled to create an adolescent gender-nonconformity (AGN) score. Each participant’s AGN total score was averaged, and Z-scores were calculated for each participant to create AGN Z-scores.
3. Results
3.1. Parent Attitudes and Adolescent Gender Nonconformity
Once Z-scores were calculated for each of the five sets of questions, a parent composite Z-score was calculated for each parent participant. The average Z-scores were calculated by averaging together each parent participant’s Gender Role Z-score, Gendered Trait Z-score, GNV Z-score, and GNR Z-score. Bivariate correlations between parent composite Z-scores and adolescent Z-scores were calculated separately for male and female adolescents. A Fisher’s r to z transformation was then performed to compare the strengths of these relationships. In addition, bivariate correlations were calculated between each of the component parent Z-scores (gender role Z-score, gendered trait Z-score, GNV Z-score, and GNR Z-score) and the male and female adolescent composite Z-scores. Each male adolescent bivariate correlation was compared to the respective female adolescent bivariate correlation.
3.1.1. Parent Attitudes and Adolescents Gender Expression
A Pearson correlation coefficient was computed to assess the linear relationship between parent attitude and adolescent gender non-conforming behaviors. There was a positive high correlation between the two variables, (r = 0.81,p < 0.01, Figure 1).
![]()
Figure 1. Correlation between parent attitude and adolescent gender non-conforming behaviors.
Next, each component of the parent assessment was examined separately in its relationship to adolescent gender expression. Pearson correlation coefficients were calculated to assess the linear relationship between each of the non-composite parent Z-scores (Gender Role Z-score, Gendered Trait Z-score, GNV Z-score, and GNR Z-score) and the adolescent Z-scores (Table 2). Every component in the parent composite exhibited a significant positive correlation to the adolescent Z-Scores, with the GNR Z-Scores exhibiting the strongest correlation and the GNV Z-scores, gendered role Z-scores, and gender trait Z-Scores exhibiting weaker but still significant correlations.
3.1.2. Parent Attitudes and Adolescents Sex Assigned at Birth
Finally, parent attitudes were compared between adolescents’ sex assigned at birth. First, a Pearson correlation coefficient was calculated to assess the linear relationship between the male adolescent composite Z-scores and the parent composite Z-scores. There was a positive high correlation between the two variables, (r = 0.78,p < 0.01, Figure 2). Then, a Pearson correlation coefficient was calculated to assess the linear relationship between the female adolescent composite Z-scores and the parent composite Z-scores. There was a positive high correlation between the two variables, (r = 0.89,p < 0.01, Figure 3). The correlation coefficient of the parents and their male children was smaller than the correlation coefficient of the parents and their female children. This indicates that the relationship between adolescent gender expression and parental views towards gender nonconformity differs for male and female adolescents. A one-way ANOVA between the composite Z-scores of parents with male adolescent children and the composite Z-scores of parents with female adolescent children revealed a significant effect of adolescents’ sex assigned at birth on parent attitudes toward gender nonconforming behaviors (F(1, 26) = 5.33, p = 0.03).
![]()
Table 2. Linear relationship between each of the non-composite parent Z-scores and the adolescent Z-scores.
![]()
Figure 2. Linear relationship between the male adolescent composite Z-scores and the parent composite Z-scores.
![]()
Figure 3. Linear relationship between the female adolescent composite Z-scores and the parent composite Z-scores.
3.2. Adolescent Gender Nonconformity
Of the 28 adolescents included in the sample, 24 (85.71%) self-reported at least one gender-nonconforming behavior. Similar to a previous report (Spivey et al., 2018), nine adolescents (32.14%) reported two or more gender-nonconforming behaviors, and female adolescents were more likely than male adolescents to do so (53.85% vs. 13.33%, respectively). A one-way ANOVA between the female composite Z-scores and male composite Z-scores revealed a significant effect of gender on number of gender non-conforming behaviors (F(1, 26) = 8.60,p < 0.01).
4. Discussion
The purpose of this study was to examine the relationship between parental attitudes toward gender non-conformity and the degree to which adolescents exhibit gender-nonconforming behaviors. The present study replicates previous findings that most adolescents exhibit at least one gender-nonconforming behavior and that female adolescents typically do so more than male adolescents. The study also finds that parents exhibit greater discomfort toward gender nonconforming behaviors with their male children than their female children, which is consistent with previous research (Sandberg et al., 1993; Spivey et al., 2018).
Additionally, this study found a positive correlation between parent attitudes toward gender nonconformity and adolescent gender expression for both male and female adolescents. The relationship between adolescent gender expression and parental views differed for male and female adolescents, with parental views better-predicting gender expression in females. One explanation for this difference is that modern Western society is generally more accepting of gender conformity in females. Put another way, parents are more likely to believe that “boys will be boys” than that “girls will be girls” (Ehrensaft, 2011). Adults that reported having displayed more gender-nonconforming behaviors in childhood also reported having had more controlling parents (Alanko, Santtila, Harlaar, Witting, Varjonen, Jern, Johansson, von der Pahlen, & Sandnabba, 2008). Additionally, research suggests that parents use more physical control strategies, such as grabbing, pushing, or spanking with boys than with girls (Endendijk, Groeneveld, van der Pol, van Berkel, Hallers-Haalboom, Bakermans-Kranenburg, & Mesman, 2017). Therefore, one possibility is that parents’ differential beliefs about gender nonconformity in boys and girls lead to differential control strategies, which leads to differential behaviors in children and adolescents.
The present study assessed parents on four measures: their views on gender roles generally, their own gender conformity, their views on their adolescents’ gender expression, and their behaviors relating to their adolescents’ gender expression. Although all four measures predicted adolescent gender expression, parents’ own gender conformity was the weakest predictor, whereas parents’ propensity to try to change their adolescents’ behaviors was the strongest predictor. This is likely because of the more direct relationship between parents’ behaviors and adolescents’ expression.
Limitations and Future Research
This study included a small sample size, with only one parent in each family surveyed, and excluding 2 families due to incomplete questionnaires, resulting in a total of 28 dyads. In addition, previous research has shown that fathers of girls are less likely to influence their gender-nonconforming behaviors than fathers of boys or mothers in general. However, the current study does not assess the differences between the parental attitudes of gender nonconformity of fathers and mothers. Future studies should expand surveys to include all parents/guardians within an adolescent’s household, as well as increase the number of dyads surveyed.
All participants surveyed reside within the same east coast suburban community in the United States. Future studies should expand the current study to include different regions and communities in order to determine whether geographic location plays a role in the relationship between parental attitudes towards gender nonconformity and adolescent gender nonconformity. Other demographic categories, such as parental income, religion, age, and education level should also be assessed in future research, especially since previous research has suggested that these factors affect adolescent development (Zucker et al., 2002).
Finally, the current study did not consider the makeup of a family when assessing the relationship between parental attitudes and their children’s gender-nonconforming behaviors. Previous research has indicated that the structure of a family (e.g., the number of children or guardians within the family) plays a role in the long-term effects of adolescent gender-role attitudes, which can affect adolescent gender expression (Kiecolt & Acock, 1988). Future studies should examine if there are differences in gender-nonconformity among adolescents in households with a single child, multiple children of the same sex, or multiple children of different sexes, as well as households with two parents or one parent. These extensions of the current study would further elucidate the familial predictors of adolescent gender expression.