Disparities in Childhood Abuse Between Transgender and Cisgender Adolescents

The Gender Minority Youth (GMY) Study was a cross-sectional online survey of TGAs and CGAs in the United States conducted from July to October 2018.²⁹  Participants were recruited via advertisements on Facebook and Instagram, social media platforms used by the vast majority of adolescents.³⁰  Two sets of advertisements targeted users ages 14 to 18. One had additional targets to reach TGAs using interest labels such as “transgender,” “gender-specific and gender-neutral pronouns,” “genderqueer,” and “passing (gender).” Almost all TGAs entered the survey through the TGA-specific advertisement, and CGAs who entered through the TGA-specific advertisement were more likely to identify with minority sexual orientations. Thus, the methods of the GMY Study oversampled TGAs and sexual minority CGAs, two difficult-to-reach populations that are underrepresented in adolescent health research.³¹ 

All participants provided assent (with a waiver of parental permission) before completing the GMY Study. Participants could enter a drawing for a $50 gift card and were provided with resources related to mental health, child abuse, and sexual assault. The University of Pittsburgh’s Human Research Protection Office approved this study.

Advertisements were served 377 469 times, and 8747 clicks were recorded (2.48% click-through rate). A total of 5642 participants assented, entered the survey, and began responding to questions. Adolescents were screened out of the survey if they were outside the targeted age range; in light of underrepresentation of TGAs assigned male at birth in the early period of data collection, additional screening was used toward the end of recruitment to allow only TGAs assigned male at birth to participate. In total, 1997 participants were screened out of the survey.

Multiple steps were taken to ensure the quality of the collected data. First, Internet Protocol (IP) addresses were used to identify potential duplicate cases, and cases with the same IP address were hand-checked. Duplicates with the same demographic characteristics and height and weight were removed (n = 320). Second, outlier analysis indicated that no cases had evidence of values outside the expected range on variables reported as counts. Third, free-response text was reviewed, and 7 cases that had inappropriate responses to survey questions were removed. Additional details of the GMY Study and data set are available elsewhere.²⁹ 

The current analysis included 1836 participants who completed the survey through the childhood abuse questions. Compared with the full sample of 3318, these 1836 participants were older, more likely to report female sex assigned at birth, and more likely to identify as bisexual or pansexual.

Participants reported their sex assigned at birth as either male or female. Participants selected all gender identities they currently identify with from the following options: male, female, transgender, female-to-male transgender/FTM, male-to-female transgender/MTF, trans male/transmasculine, trans female/transfeminine, genderqueer, gender expansive, intersex, androgynous, nonbinary, two-spirited, third gender, agender, not sure, and other. A 7-category gender identity variable was created, including cisgender male, cisgender female, transgender male (including participants who reported female sex assigned at birth and male, female-to-male transgender/FTM, and/or trans male/transmasculine identities), transgender female (including participants who reported male sex assigned at birth and female, male-to-female transgender/MTF, and/or trans female/transfeminine identities), nonbinary assigned female at birth, nonbinary assigned male at birth, and questioning gender identity (including participants who selected “not sure” and no other gender identities). Adolescents were categorized as nonbinary if they reported a genderqueer, gender expansive, intersex, androgynous, nonbinary, two-spirited, third gender, or agender current gender identity and no binary gender identities. In other words, adolescents were not categorized as nonbinary if they selected any of the binary identities. We have found empirical support for this approach to categorization in previous analyses of this data set.³  Questioning adolescents could not be divided by sex assigned at birth because of small cell sizes.

Abuse items were adapted from the Adverse Childhood Experiences questionnaire.³²  Psychological abuse (being sworn at, insulted, put down, or humiliated or being threatened and made afraid of physical harm), physical abuse (being pushed, grabbed, slapped, or having something thrown at you or being hit so hard you had marks or were injured), and sexual abuse (being touched, fondled, or forced to touch an adult in a sexual way or attempted or actual oral, anal, or vaginal sex with an adult at least 5 years older) during childhood by a parent or other adult were assessed separately with 2 questions each. Each abuse category was dichotomized (0: never experienced this form of abuse; 1: any experience of this form of abuse), a coding strategy implemented in previous research with the Adverse Childhood Experiences questionnaire.³³ 

Participants reported their age, race (coded as white, Black, Hispanic, Asian American or Pacific Islander, mixed, and Native American or other), and sexual orientation (coded as heterosexual, gay or lesbian, bisexual or pansexual, and queer, questioning, or other). Participants were coded as a sexual minority if they endorsed any sexual orientation other than heterosexual. Subjective social status (SSS) was measured with the MacArthur Scale of Subjective Social Status, a measure of adolescents' perceptions of their family's social status compared with all other families in American society, visualized by a 10-rung ladder.³⁴  Demographic covariates were selected because they have conceptual relevance to childhood abuse, evidenced significant associations with constructs of interest in analyses, and/or were associated with attrition in the survey before completing the abuse items.

Descriptive demographic information for the full sample, for heterosexual CGAs, for sexual minority CGAs, for TGAs, and for each gender identity subgroup is presented in Table 1. According to zip codes, participants lived in all 50 states, as well as Washington, DC, and Puerto Rico. CGAs in the sample were similar to nationally representative data regarding race. Compared with CGAs, TGAs were more likely to report white race, minority sexual orientations, older age, and lower SSS.

First, descriptive data for each child abuse outcome were examined for heterosexual CGAs, sexual minority CGAs, TGAs, and each gender identity subgroup. Second, χ² tests, including pairwise comparisons between heterosexual CGAs, sexual minority CGAs, and TGAs as well as between each of the 7 gender identity subgroups on each childhood abuse outcome were estimated. Bonferroni corrections were applied to significance levels to account for multiple comparisons (n = 3 in 3-category comparisons [P < .017]; n = 21 in 7-category comparisons [P < .002]). Third, multivariable logistic regression models were estimated to examine odds of each abuse outcome for TGAs (aggregated into 1 group) and sexual minority CGAs compared with heterosexual CGAs, while adjusting for sex assigned at birth, age, SSS, and race. Finally, multivariable logistic regression models were estimated for each dichotomized abuse outcome predicted by gender identity (coded as 7 subgroups) while controlling for sexual orientation, age, SSS, and race.

Percentages, along with 95% confidence intervals (CIs), of heterosexual CGAs, sexual minority CGAs, TGAs, and participants in gender identity subgroups endorsing dichotomized abuse outcomes are presented in Figs 1–3. Means and SEs were multiplied by 100 to transform to percentage metric before calculating CIs. Within subgroup comparisons, each χ² omnibus test was significant (all P values <.001), indicating each abuse outcome varied significantly across the 3 broad sexual orientation and gender identity categories and the 7 specific gender identity categories.

In unadjusted comparisons, TGAs (aggregated into one group) reported higher rates of psychological abuse than heterosexual CGAs, and sexual minority CGAs did not differ from heterosexual CGAs. In the adjusted model, TGAs had higher odds of psychological abuse than heterosexual CGAs (odds ratio [OR] = 1.84), but heterosexual and sexual minority CGAs did not differ (see Table 2). Pairwise 7-category gender identity comparisons indicated that transgender males and nonbinary adolescents assigned female at birth reported higher levels of psychological abuse than both cisgender males and females. Within adjusted analyses examining subgroup differences, cisgender males had lower odds of reporting psychological abuse compared with cisgender females. In the same model, transgender males and nonbinary adolescents assigned female at birth had higher odds of reporting psychological abuse than either cisgender group (see Table 3).

In unadjusted comparisons, TGAs reported higher rates of physical abuse than heterosexual CGAs, and sexual minority CGAs did not differ from heterosexual CGAs. Similar results were observed in the adjusted model after controlling for covariates because TGAs had higher odds of reporting physical abuse than heterosexual CGAs (OR = 1.61). Adolescents questioning their gender identity reported higher levels of physical abuse than CGAs in unadjusted pairwise comparisons by subgroup, and questioning adolescents had higher odds of physical abuse in an adjusted model compared with cisgender female adolescents.

In unadjusted comparisons, both TGAs and sexual minority CGAs had higher odds of reporting sexual abuse compared with heterosexual CGAs, and these results persisted within adjusted logistic regression analysis (OR = 2.04 for TGAs and OR = 1.87 for sexual minority CGAs). Nonbinary adolescents assigned female at birth reported higher levels of sexual abuse than CGAs in unadjusted pairwise comparisons. After adjusting for covariates, cisgender males had lower odds of reporting sexual abuse compared with cisgender females, and no TGA subgroups had significantly different odds of reporting sexual abuse compared with cisgender females.

TGAs are more likely to report psychological, physical, and sexual abuse during childhood compared with heterosexual CGAs. Our findings align with those of previous studies finding high rates of childhood abuse among transgender individuals.^22,24,25  Growing evidence indicates TGAs experience mental health problems at higher rates than CGAs, and childhood abuse likely contributes to the onset of mental health problems among TGAs. In the future, researchers should examine the role of childhood abuse in the etiology of mental health problems among TGAs.

In addition, our data set enabled us to examine subgroup differences in childhood abuse between TGAs and CGAs. In particular, TGAs assigned female at birth were more likely to report psychological abuse by parents or other adults in the household. Researchers have hypothesized that individuals who have experienced physical and/or sexual abuse will also endorse psychological abuse items because of overlap between these constructs, although they may not have experienced discrete psychological abuse.³⁵  Given this possibility, we conducted post hoc analyses examining multivariate associations between our 7-category gender identity variable and psychological abuse while removing all participants who endorsed physical or sexual abuse. In these analyses, transgender males retained higher odds of psychological abuse compared with female CGAs (OR = 1.80; 95% CI: 1.27–2.56), but nonbinary adolescents assigned female at birth no longer had higher odds of psychological abuse (OR = 1.35; 95% CI: 0.89–2.04). These findings indicate transgender males have higher risk for psychological abuse, which is separate from their risk for physical or sexual abuse, but nonbinary adolescents assigned female at birth no longer have higher risk for psychological abuse when accounting for the potential cumulative effect of other forms of abuse on this outcome. Among CGAs, nationally representative data indicate individuals assigned female at birth are more likely to report childhood psychological maltreatment,³⁶  and the same could be true of TGAs assigned female at birth. In the future, researchers should examine how parent-adolescent relationships are associated with mental health outcomes among TGA subgroups.

As noted above, the GMY Study included a large sample of adolescents who were racially diverse, and participants resided in all 50 states. The current study achieved a 2.5% click-through rate during recruitment. Although this rate is similar to other recent online studies of sexual minority and transgender youth,^37,38  it is possible that adolescent social media users who have experienced mental or physical health problems were more likely to click on an ad for a “health” study than those who have not, and these youth may disproportionately report childhood abuse. However, rates of physical and sexual abuse reported by participants in the data set are comparable to rates within other data sets of adolescents. For example, 27% of sexual minority CGAs reported physical abuse (19%–33% reported physical abuse in a large meta-analysis),²³  and ORs between TGAs and CGAs in adjusted models were similar to those found in a recent statewide survey.²⁵  These qualities of the GMY Study data set indicate our results could generalize to the broader population of adolescents in the United States. However, our sample is not nationally representative, and it is critical that future nationally representative surveys of adolescents in the United States, such as the Youth Risk Behavior Surveillance System, assess both the sex assigned at birth and current gender identity to enable the accurate identification and categorization into subgroups of TGAs in a nationally representative data set.

In addition, the current study is limited by its cross-sectional design. Although abuse experiences could be more common among TGAs before their identification as TGAs because of gender nonconformity during childhood,^19,20  abuse could also have onset and/or exacerbation after their identification as TGAs. For example, TGAs often report parental rejection after disclosure of their gender identity to parents,¹⁸  potentially resulting in onset of abuse by parents. Longitudinal studies of TGAs across development could shed light on how gender identity development, abuse, and mental health are related over time. Furthermore, the GMY Study battery included only assessments of psychological, physical, and sexual abuse, and did not include assessment of other experiences, such as parental neglect or witnessing domestic abuse between parents. This limitation makes it difficult to examine polyvictimization in this data set. Finally, our sample included fewer TGAs assigned male at birth than TGAs assigned female at birth, and this altered sex ratio is common within samples of TGAs.³⁹  However, this may have limited our power to detect differences between TGAs assigned male and CGAs in the present analyses.

It is recommended that pediatric medical and mental health professionals screen for child abuse to recognize and respond to ongoing maltreatment among children and adolescents,⁴⁰  and this recommendation should apply to TGAs presenting for care. Given the higher risk for psychological abuse by parents and other adults among TGAs assigned female at birth in this sample, providers should pay particular attention to parent-adolescent relationships when treating this population. Because some families of TGAs are rejecting of their gender identity,¹⁸  providers should assess gender identity privately without parents present if possible, including level of parental knowledge of gender identity among TGAs. This approach will optimize the possibility that TGAs are open about their gender identity with providers and limit the chances that interactions with providers could increase family discord related to gender identity, which has the potential to exacerbate childhood abuse among TGAs. Family-based cognitive-behavioral interventions could be implemented to reduce the risk of future abuse within families of TGAs,⁴¹  and TGAs who have experienced abuse should be provided access to evidence-based treatment, such as trauma-focused cognitive-behavioral therapy, to reduce mental health symptoms.^42,43  Providers should carefully tailor their interventions to ensure their care is validating of TGAs’ gender identity while limiting rejection and fostering acceptance by family members.^44,45 

TGAs are more likely to report childhood abuse than heterosexual CGAs, and risk for psychological abuse is highest among TGAs assigned female at birth. Clinicians should be aware that higher levels of abuse could contribute to disproportionate mental health problems among TGAs, and future research should examine how childhood abuse contributes to higher levels of mental health problems among TGAs over time.

CGA

cisgender adolescent

CI

confidence interval

GMY

gender minority youth

OR

odds ratio

SSS

subjective social status

TGA

transgender adolescent