Greater Gender Diversity among Autistic Children by Self Report and Parent Report

Abstract

Emerging research suggests over-representation of gender diversity among autistic youth. Previous gender diversity research with autistic children has relied on parent-report based on a single question. The Gender Diversity Screening Questionnaire self-report (GDSQ-S) and parent-report (GDSQ-P) assessed gender diversity experiences from 244 children (140 autism spectrum disorder (ASD) and 104 typically developing (TD)) between 10-to-13 years, and their parents. Parent-report Child Behavior Checklist Item-110, “Wishes to be the opposite sex” was also collected. Autistic children endorsed higher GDSQ-S Binary Gender Diversity, t(223.21)=−2.83, adjusted p=0.02, d=−0.35, and Nonbinary Gender Diversity, t(191.15)= −3.79, adjusted p= 0.001, d=−0.46, than TD children. Similarly, for GDSQ-P, there was a significant Gender Body Incongruence difference between the groups, t(189.59)= −2.28, adjusted p=0.05, d=−0.30. Within-group analyses revealed that parents of autistic females-assigned-at-birth reported significantly more gender Body Incongruence than males-assigned-at-birth, t(32.91)= −3.78, p<0.001, d=−1.11. Moreover, distinct within-group profiles of gender experiences and internalizing symptoms were revealed, showing associations between gender profiles and symptoms of anxiety, depression, and suicidality. Results extend reports showing increased rates of gender diversity in autistic children based on both self- and parent-report, underscoring the need to better understand and support the unique and complex needs of autistic children who experience gender diversity.

Lay Abstract

Gender diversity broadly refers to the way in which an individual experiences (expressions and/or identities) their gender distinctly to that which would be expected based upon social norms for their gender assigned at birth. Recent research has shown a higher representation of gender diversity among autistic youth. Previous research in this area has relied on parent-report based on a single question from the Child Behavior Checklist Item-110, asking if their child “Wishes to be the opposite sex”. The Gender Diversity Screening Questionnaire self-report and parent-report were used to assess the experience of gender diversity in 244 children (140 autism spectrum disorder [ASD] and 104 typically developing [TD]) between 10-to-13 years. The Item-110 was also collected. Results showed that autistic children endorsed much higher rates of binary gender diversity (less identification with their designated sex and more with the other binary sex) and nonbinary gender diversity (identification as neither male nor female) than TD children. Similarly, parents of autistic children reported significantly more gender-body incongruence experienced by their child than parents of TD children. Specifically, parents of autistic females-assigned-at-birth reported significantly more gender-body incongruence than autistic males-assigned-at-birth. Parent- and self-report measures were largely related. Moreover, statistical comparisons between and within the groups revealed associations between gender profiles and symptoms of anxiety, depression and suicidality. Results extend previous reports showing increased rates of gender diversity in autistic children, now based on both self- and parent-report, and highlight the need to better understand and support the unique and complex needs of autistic children who experience gender diversity.

Introduction

In many cultural traditions, gender has been considered a binary concept: the state of being male or female based on sex designated at birth and with reference to social and cultural norms. However, there is increasing acknowledgment that many individuals do not align or experience gender with such finite distinctions (Whyte, Brooks, & Torgler, 2018). In fact, survey-based research estimate as many as 1–2% of adolescents identify as gender diverse (i.e., gender non-conforming or transgender; Clark et al., 2014; Rider, McMorris, Gower, Coleman, & Eisenberg, 2018; Shields et al., 2013). While terminology in the field is evolving, gender diversity broadly refers to an individual’s experience of aspects of their gender as different from their assigned sex at birth. Gender diversity experiences in youth can include gender incongruence, in which a person’s gender identity differs from their gender assigned at birth (Claahsen-van der Grinten et al., 2021). Gender dysphoria presents when gender incongruence contributes to significant clinically relevant distress, requiring clinical gender supports (Becker, Ravens-Sieberer, Ottova-Jordan, & Schulte-Markwort, 2017). Those with clinical gender needs may require an evaluation with a gender development specialist and, in some cases, may pursue gender-affirming interventions. For others, gender diversity may be an expression of underlying gender experiences that to varying degrees color and shape an individual’s identity or expression without the need of gender affirming interventions. There is evidence that youth who experience gender diversity are at greater risk for a range of mental health and broader life challenges related to stress and victimization (Kuper, Adams, & Mustanski, 2018; Lowry et al., 2018; van der Miesen, Nabbijohn, Santarossa, & VanderLaan, 2018). Therefore, it is critical to understand the experiences, needs, and potential protective factors of gender diverse youth.

There is growing evidence of an over-occurrence of co-occurring gender diversity and autism in youth (e.g., Nabbijohn et al., 2018; van der Miesen, de Vries, Steensma, & Hartman, 2018). Autistic children¹ have persistent deficits in the area of social communication and a repertoire of repetitive, restricted or stereotyped interests and behaviors (RRBI; APA, 2013), though there is notable heterogeneity. A recent report by the Centers for Disease Control estimated that 1-in-54 children are diagnosed with autism spectrum disorder (ASD) in the United States (Maenner et al., 2020), yet rates of ASD among gender-referred youth appear to be much higher (Strang, Janssen, et al., 2018; van der Miesen, de Vries, et al., 2018). Greater gender diversity experiences have also been reported in well-characterized autistic child samples, though reports have been limited to parent report (Janssen, Huang, & Duncan, 2016; Strang et al., 2014).

An Affirming Approach to the Autism and Gender Diversity Co-Occurrence

Some of the early case reports of co-occurring autism and gender diversity (e.g., Landen & Rasmussen, 1997; Williams, Allard, & Sears, 1996) offered interpretations through the lens of pathology: that autism characteristics, such as RRBI produced gender diversity-related “behaviors” (i.e., gender diversity as an RRBI). Recent advances in the understanding of the breadth of gender diversity experiences in the general population (e.g., Rider et al., 2018) as well as qualitative accounts of the lived experiences of gender diversity amongst autistic youth (Strang, Powers, et al., 2018), adults (Cooper, Mandy, Butler, & Russell, 2021) and self-advocates (Autistic Self Advocacy Network, 2016; Falk, 2018) have shifted the focus in conceptualization of the co-occurrence from a disorder-based framework, to conceptions that emphasize neurodiversity and gender diversity experiences. A longitudinal qualitative study of autistic gender-diverse adolescents identified themes of deeply felt and integral gender-related experiences among the autistic youth as well as a range of gender trajectories over time (Strang, Powers, et al., 2018). Gender identity formation from an autistic perspective might produce different gender experiences or trajectories of gender development (van Schalkwyk, Klingensmith, & Volkmar, 2015) or allow for greater forthrightness about underlying gender diversity-related experiences shifts the narrative from “comorbidity” to human diversity. It is from this gender diversity and neurodiversity affirming perspective that the current study authors approach investigation into the co-occurrence of autism and gender diversity experiences.

Autism as Assessed in Gender Diverse Children

Clinical gender programs in several countries have reported over-representation of autism and autistic characteristics among youth referred for services (e.g., Holt, Skagerberg, & Dunsford, 2016; Kaltiala-Heino, Sumia, Tyolajarvi, & Lindberg, 2015; Shumer, Reisner, Edwards-Leeper, & Tishelman, 2016; van der Miesen, de Vries, et al., 2018). The first systematic study was led by de Vries, Noens, Cohen-Kettenis, van Berckelaer-Onnes, and Doreleijers (2010), who investigated ASD incidence in children and adolescents referred for gender diversity services. Specifically, it was shown that 7.8% of the sample met strict diagnostic criteria for ASD, a figure 10 times higher than the previously reported prevalence of ASD in the general population at that time (Baird et al., 2006; Fombonne, 2005). Recently, van der Miesen and colleagues (2018) compared ASD symptoms using the Children’s Social Behavior Questionnaire (CSBQ) in large samples of autistic and typically developing (TD) youth with gender dysphoria (GD). Overall, youth with GD had higher scores on the CSBQ (i.e., more ASD symptoms) compared to TD youth, but significantly lower symptoms than the ASD group. When applying a cutoff score on the CSBQ potentially suggestive of an ASD diagnosis, 14.5% of the GD group met that threshold.

A systematic review of the co-occurrence (Glidden, Bouman, Jones, & Arcelus, 2016) included nineteen studies, twelve with youth, and showed the prevalence of ASD or autistic characteristics is higher in gender diverse youth. In a separate literature review, van der Miesen and colleagues (2016) found that autism and gender diversity coexist more often than would be expected by chance. However, autism screeners used in some studies of the co-occurrence could lead to false-positives for autism due to the endorsement of social or emotional challenges that may be attributed to other factors (Turban & van Schalkwyk, 2018). Strang, Janssen, and colleagues (2018) in their reanalysis of available studies omitted studies employing ASD screeners, instead identifying studies of gender-diverse and transgender youth that included rates of existing clinical autism diagnoses or ASD diagnosed through more comprehensive approaches: Of all 7 available studies, rates of clinical ASD were significantly greater as compared to population base-rates for ASD (Strang, Janssen, et al., 2018).

Gender Diversity Experiences as Assessed in Autistic Children

To date, four studies have examined rates of gender diversity in autistic children (Hisle-Gorman et al., 2019; Janssen et al., 2016; May, Pang, & Williams, 2017; Strang et al., 2014), all by parent report. Strang and colleagues (2014) examined the rates of binary “gender variance” (specifically, CBCL Item-110:”Wishes to be of opposite sex [sic]”, which will be referred to as “gender variance”) using item-level analysis of the Child Behavior Checklist (CBCL) in a chart review study. The sample included children with ASD (N=147), ADHD (N=126) or a neurodevelopmental medical disorder (N=116) and two non-referred samples: standardization (N=1,605) and control sample (N=165). The results, based on parent-report, showed a significantly greater proportion of “gender variance” in ASD (5.4%) and ADHD (4.8%) than the neurodevelopmental medical group (1.7%) or non-referred control (0 – 0.7%). The autistic children were 7.59 times more likely to have parent reported “gender variance” supporting the hypothesis of over-representation of gender diversity in ASD. Janssen and colleagues (2016) replicated part of this approach in a larger sample of autistic youth. Comparing 492 autistic children to a non-referred standardization sample, they found 5.1% of autistic youth had parent reported gender variance, 7.76 times the rate of the comparison sample. May and colleagues (2017) also replicated this approach with 176 autistic children and found that 4% of autistic youth had parent-reported “gender variance” but noted that this over-representation did not differ from rates of gender variance in clinically referred youth in general.

More recently, Hisle-Gorman et al. (2019) utilized the Military Health System to investigate GD in a large sample of children diagnosed with ASD compared to a matched-control non-ASD sample of children and found autistic children were over four times more likely to be diagnosed with GD.

Gender Diversity and Internalizing Symptoms

Many gender diverse youth experience significant rates of internalizing disorders (e.g., anxiety, depression). For example, a study in the Netherlands found 21% of gender dysphoric adolescents met criteria for an anxiety disorder, while 12.4% met criteria for a mood disorder (de Vries, Doreleijers, Steensma, & Cohen-Kettenis, 2011). The most common diagnosis in this sample was social anxiety disorder, at 9.5%. Additionally, poor peer relations, peer victimization, and social ostracism are all strong predictors of emotional health outcomes in gender diverse youth (Cohen-Kettenis, Owen, Kaijser, Bradley, & Zucker, 2003; de Vries, Steensma, Cohen-Kettenis, VanderLaan, & Zucker, 2016; Steensma et al., 2014). Addressing the sources of this emotional distress in these youth is critical given high prevalence of suicidal ideation and attempts (Grossman & D’Augelli, 2006; Khatchadourian, Amed, & Metzger, 2014), especially in those of female gender identity (Strang et al., 2021). Overall, it is estimated that rates of internalizing disorders in youth with GD surpasses the averages for the general population (Chodzen, Hidalgo, Chen, & Garofalo, 2018).

ASD and Internalizing symptoms

Many autistic youth experience symptoms of anxiety and depression (e.g., Gotham, Brunwasser, & Lord, 2015), with rates exceeding those of the general population (Mayes, Calhoun, Murray, Ahuja, & Smith, 2011; Simonoff et al., 2008). Due to the high rates of internalizing symptoms and stigma among both autistic and gender diverse youth, it would be expected that individuals with the co-occurrence might experience even greater mental health concerns and stigmatization. Indeed, the combination of ASD and gender-dysphoric traits has been associated with poorer mental health outcomes in adults than autistic people without gender-dysphoric traits (George & Stokes, 2018). Strang et al. (2014) found that gender diverse children with neurodevelopmental disorders had higher levels of parent-reported anxiety and depression than gender typical children with neurodevelopmental disorders. However, autistic gender diverse children in this study were less anxious than gender diverse children with non-autism-based neurodevelopmental conditions, purportedly as a result of diminished social awareness as to how they are perceived by others. More recently, however, a study of autistic transgender adolescents reported that more than half demonstrated self-reported internalizing symptoms above clinical threshold (Strang et al., 2021).

Current Study: Assessing Self and Parent Report of Gender Diversity in Children

While parent-report gender-identity-item-level analysis of the CBCL (Achenbach, 2001) has been shown to be related to GD and gender diversity experiences (Cohen-Kettenis et al., 2003), the approach has been criticized since the measure is not specific to GD (Turban & van Schalkwyk, 2018), and the item is specifically binary, omitting nonbinary experience. One study has investigated self-reported binary gender diversity in adolescents and adults using the self-report version of the CBCL binary gender item (van der Miesen, Hurley, Bal, & De Vries, 2018). However, to date, no study has investigated self-report of gender in autistic children. Further, no study has studied self-report of both binary and nonbinary gender experience in autistic youth. Finally, the extent to which parent perceptions are consistent with child-report of gender diversity experiences is unknown and warrants investigation, as parents often determine a child’s access to gender-related supports (e.g., identifying gender specialized providers, scheduling/facilitating appointments). Therefore, the current prospective study addresses these next steps using a gender screening measure developed and tested in autistic gender diverse and autistic cisgender youth (Strang et al., 2017) with parent- and self-report versions. The study focuses on four aspects of gender diversity experiences via parent report and two aspects based on self-report in a well-defined sample of autistic and TD children. Based on previous findings, the following hypotheses are made: 1) there will be significant differences between the ASD and TD parental groups such that parents of autistic children will report more gender diversity experiences in their child (e.g., Janssen et al., 2016; Strang et al., 2014); 2) autistic children will endorse more gender diversity experiences than TD peers; and 3) self- and parent-reported gender diversity will relate to higher rates of internalizing symptoms.

Methods

Participants

Participants included 244 children between 10–13 years, 140 with ASD (mean age=11.42, SD=1.03) and 104 TDs (mean age=11.71, SD=1.21). In the ASD group, there were 104 males and 36 females, and in the TD group, there were 58 males and 46 females. Families were recruited from the Central Southern part of the United States. Recruitment was from the community, research registries, ASD diagnostic clinics, and local autism/disability organizations. Participants were required to have an intelligence quotient (IQ) score ≥70. The racial and ethnic representation of the sample was 83.2% White, 7.8% Black, 0.4% Asian, and 8.6% mixed race. Regarding medication status, 67% of the ASD group and 17.3% of the TD group were on any medications. Demographic and diagnostic characteristics are presented in Table 1. One of the autistic participants was exploring gender identity diversity outwardly.

Table 1.

Demographic and Diagnostic Information

Measure	TD	ASD	dfa	t	p-value	95% CI	d b
	M (SD)	M (SD)
Age	11.71 (1.21)	11.42 (1.03)	200.57	1.91	0.06	[−0.01, 0.57]	0.25
Estimated FSIQ	116.93 (13.89)	100.96 (20.77)	239.01	7.18	<0.001	[11.59, 20.36]	0.88
ADOS Total	--	12.58 (4.57)	--	--	--	--	--
SCQ	2.69 (2.49)	17.53 (8.35)	171.50	−19.86	<0.001	[−16.32, −13.37]	−2.26
CBCL Affective	55.48 (7.24)	66.18 (8.74)	236.97	−10.41	<0.001	[−12.73, −8.68]	−1.32
CBCL Anxiety	54.94 (7.51)	66.45 (7.90)	240	−11.44	<0.001	[−13.49, −9.53]	−1.49
C-SSRS Total Suicidality	0.13 (0.56)	0.54 (1.19)	206.43	−3.59	<0.001	[−0.64, −0.19]	−0.43
	Proportion	Proportion	N	X2	p-value	φ
Sex: Female	0.442   46/104	0.257   36/140	244	9.17	0.002	0.19
Race			244	12.06	0.007	0.22
White	0.856 (89/104)	0.814 (114/140)
Black	0.019 (2/104)	0.121 (17/140)
Multiracial	0.125 (13/104)	0.057 (8/140)

Note: Autism Spectrum Disorder = ASD; CBCL = Child Behavior Checklist; CI = Confidence Interval; C-SSRS = Columbia Suicide Severity Rating Scale; FSIQ = Estimated Full Scale Intelligence Quotient. SCQ = Social Communication Questionnaire; TD = Typical Development.

^aWelch degrees of freedom when assumption of equal variances was violated.

^bCohen’s d

Diagnostic Criteria

The diagnosis of ASD was based on the Diagnostic and Statistical Manual-5 (APA, 2013) and established by: (1) a previous diagnosis by a psychologist, psychiatrist, or behavioral pediatrician with autism expertise; (2) current clinical judgment, and (3) corroborated by the Autism Diagnostic Observation Schedule (ADOS-2; Lord et al., 2012), administered by research-reliable personnel, an approach consistent with standards set by the NICHD/NIDCD Collaborative Programs of Excellence In Autism (Lainhart et al., 2006). Typically developing children were operationalized as individuals with an absence of ASD using DSM criteria (APA, 2013). Participants were screened based on parent-report and clinical observation to ensure an absence of ASD, neurodevelopmental disorders (e.g., ADHD), neuropsychiatric conditions (e.g., psychosis) or previous head trauma. The research was carried out in accordance with the Code of Ethics of the World Medical Association (Declaration of Helsinki). The Vanderbilt Institutional Review Board approved the study. Informed written consent and assent was obtained from all parents and study participants, respectively, prior to inclusion in the study.

Diagnostic Procedures

Autism Diagnostic Observation Schedule-Second Edition (ADOS-2; Lord et al., 2012) is a semi-structured, play and interview-based instrument used to support the diagnosis of ASD. A score of 7 or above on Module 3 (fluent speech) was required for inclusion.

Social Communication Questionnaire (SCQ, Rutter, Bailey, & Lord, 2003) is an autism screener (score of 15 is suggestive of ASD and a score of 22 is suggestive of autistic disorder). TD children with a score of ≥ 10 were excluded.

Wechsler Abbreviated Scale of Intelligence, Second Edition (WASI-II, Wechsler, 2011) is a measure of cognitive ability that was used to obtain an efficient, reasonable estimate of the child’s intellectual functioning (IQ≥70 required).

Dependent Measures

Gender Diversity

The Gender Diversity Screening Questionnaire (GDSQ) is a brief self-report (GDSQ-S) and parent-report (GDSQ-P) screener developed through an iterative participatory process with autistic and TD gender-diverse and cisgender youth as part of an ongoing multi-site initiative to calibrate gender characterization tools for diverse populations, including autism (Pelphrey, 2017; Strang et al., 2017). The GDSQ-S and GDSQ-P have strong psychometrics, with dimensions developed through data-driven factor analytics in autistic and non-autistic youth and their parents/caregivers. Internal reliability ranges from acceptable to excellent. Details regarding the development and calibration of the GDSQ are available through Open Science Framework: https://osf.io/qh25d/?view_only=c0ce41d07bca4af1b792e074d51b7ded and https://osf.io/qfza7/?view_only=c5c7503b622042389ac1ca26c90ff313. The parent-report provides four domain scores: Gender Expression Diversity (i.e., gender diversity in a child’s gender expression as compared to their sex assigned-at-birth), Gender Identity Diversity (i.e., the experience of not being the sex assigned-at-birth and experience of being a different gender), Gender-related Social and Play Interests (i.e., how much a child’s social and play interests conform to common interests associated with their assigned sex at birth versus the other binary sex), and Body-related Gender Incongruence (i.e., body-related gender dysphoria). The self-report, presented as a set of printed questions for the child to look at and respond to as the examiner reads the questions, provides two domain scores: Binary Gender Diversity and Nonbinary Gender Diversity. Nonbinary Gender Diversity captures the degree to which a child experience themself as neither male nor female. Binary Gender Diversity captures how much the child feels like their assigned sex versus the other binary gender (male vs. female). Higher scores on each of the domains represent greater experience of gender diversity. In the case of Binary Gender Diversity, a higher score indicates increased alignment with features of the other binary gender on the male-female continuum of binary gender experience. In the case of Nonbinary Gender Diversity, higher scores indicate increased experience of gender that is neither male nor female. The simultaneity of binary and nonbinary gender experiences is not uncommon, as in clinical practice, many gender diverse youth report both a primary binary gender (e.g., “male”) and concurrent experience of nonbinariness (e.g., “I am a nonbinary male”). The GDSQ differentiated autistic gender service-seeking youth from autistic and typically developing cisgender youth (Goldstein et al., 2017).

Child Behavior Checklist (CBCL; Achenbach, 2001) is a well-established, frequently used behavioral inventory for parents. For the gender diversity part of the study, Item-110 “Wishes to be the opposite sex [sic]” was examined. The CBCL item-level approach has been used to identify children, adolescents, and young adults who experience “gender variance” in ASD (Janssen et al., 2016; van der Miesen et al., 2018), ADHD, and control groups (Strang et al., 2014) and a twin study showing strong genetic heritability (van Beijsterveldt, Hudziak, & Boomsma, 2006). Additionally, the item has been used in gender care clinics demonstrating strong positive correlations between endorsement on the CBCL index and GID diagnosis (Cohen-Kettenis et al., 2003).

Internalizing Symptoms

To examine internalizing symptoms, the CBCL for ages 6 – 18 years was used (Achenbach, 2001). The CBCL is a common measure of anxiety, depression, and other emotional and behavioral problems. Previous studies have shown elevations in children with ASD, at or above the clinical range for withdrawn/depressed, social, thought and attention problems scales (Mazefsky, Anderson, Conner, & Minshew, 2011). Affective and the Anxiety T scores were used.

Columbia Suicide Severity Rating Scale (C-SSRS; Posner et al., 2011) assesses suicidal ideation and suicidal behavior within the last 3 months and lifetime, conducted by trained clinicians. If participants endorsed suicidal thoughts or behaviors, further assessment was provided as needed by the study PI and lead clinician. The total lifetime suicidality (number of “yes” answers for all 10 suicidal ideation and behavior categories) was used. The C-SSRS has not been validated in autistic samples specifically; however, several studies with autistic adults and adolescents suggest that the C-SSRS provides one of the most comprehensive assessments of suicidality of the measures available (Cassidy, Bradley, Bowen, Wigham, & Rodgers, 2018a, 2018b; Howe, Hewitt, Baraskewich, Cassidy, & McMorris, 2020). In addition to the measure itself, administration of the C-SSRS by trained psychiatrists and psychologists with expertise in autism in the present study affords a more robust approach for screening suicidality in this cohort.

Statistical Analyses

Independent sample t-tests with 95% confidence intervals were conducted to test for differences between groups on the demographic and diagnostic variables. If the assumption of normality was violated, the equivalent nonparametric test was used. The assumption of homogeneity of variance was tested with Levene’s test, and the Welch degree of freedom approximation was used when the assumption was violated.

To correct for multiple comparisons, p-values were adjusted using the Benjamini-Hochberg procedure (Benjamini & Hochberg, 1995). To examine associations between CBCL and GDSQ elements, a correlation matrix using Spearman’s rho correlations was created (Table 2). Separate correlation matrices are also presented for ASD and TD (see Table 3). To test the hypothesis that self- and parent-reported gender diversity will relate to higher rates of internalizing symptoms, multivariable linear regression models predicting internalizing symptoms were conducted with GDSQ as predictors and adjusted for age, gender, and diagnosis. Assumptions of each model have been checked and met. In order to gauge the sample size adequacy and statistical precision achieved in the current study, we refer the reader to examine the 95% confidence intervals and effect sizes provided in the results section.

Table 2.

Spearman Rho Correlation Matrix for GDS-Q and Internalizing Symptoms

	1.	2.	3.	4.	5.	6.	7.	8.	9.	10.
1. Gender Expression-P	1.0
2. Gender Identity- P	0.12	1.0
3. Social and Play Interests – P	0.25**	−0.11	1.0
4. Gender Incongruence – P	0.17*	0.04	0.10	1.0
5. Binary Gender Diversity – S	0.04	0.12	0.01	0.11	1.0
6. Nonbinary Gender Diversity – S	0.05	0.10	−0.05	−0.09	0.27**	1.0
7. CBCL Item 110	0.22**	0.15*	0.17*	0.26**	0.07	0.02	1.0
8. CSSRS Total	0.11	0.04	0.06	0.15*	0.14	0.04	0.04	1.0
9. CBCL Affective	−0.02	0.06	0.10	0.25**	0.11	0.03	0.20**	0.24**	1.0
10. CBCL Anxiety	0.01	0.04	0.10	0.23**	0.07	0.08	0.13	0.24**	0.68	1.0

^*p≤0.05,

^**p≤0.01

Table 3.

Spearman Rho Correlation Matrix for GDS-Q and Internalizing Symptoms by Diagnosis

Autism Spectrum Disorder	1.	2.	3.	4.	5.	6.	7.	8.	9.	10.
1. Gender Expression-P	1.0
2. Gender Identity- P	0.04	1.0
3. Social and Play Interests – P	0.23**	−0.14	1.0
4. Gender Incongruence – P	0.13	0.02	0.19*	1.0
5. Binary Gender Diversity – S	−0.03	0.13	−0.01	0.11	1.0
6. Nonbinary Gender Diversity – S	0.06	0.06	−0.16	−0.16	0.27**	1.0
7. CBCL Item 110	0.24**	−0.04	0.23**	0.29**	−0.02	−0.11	1.0
8. CSSRS Total	0.03	−0.05	0.05	0.10	0.01	−0.03	−0.09	1.0
9. CBCL Affective	−0.08	0.02	0.04	0.22*	−0.07	−0.21*	0.17	0.19*	1.0
10. CBCL Anxiety	−0.12	0.00	0.01	0.23**	−0.02	−0.04	0.15	0.17	0.45**	1.0
Typically Developing	1.	2.	3.	4.	5.	6.	7.	8.	9.	10.
1. Gender Expression-P	1.0
2. Gender Identity- P	0.21*	1.0
3. Social and Play Interests – P	0.24*	−0.07	1.0
4. Gender Incongruence – P	0.22*	0.05	−0.03	1.0
5. Binary Gender Diversity – S	0.09	0.06	0.01	0.08	1.0
6. Nonbinary Gender Diversity – S	−0.04	0.13	0.08	−0.05	0.19	1.0
7. CBCL Item 110	0.19	0.58**	0.06	0.22*	0.23*	0.30**	1.0
8. CSSRS Total	0.20	0.22	0.00	0.15	0.36**	0.04	0.42**	1.0
9. CBCL Affective	−0.08	0.04	0.11	0.15	0.26*	0.16	0.19	0.10	1.0
10. CBCL Anxiety	−0.01	0.02	0.10	0.14	0.09	−0.01	0.13	0.10	0.63**	1.0

^*p≤0.05,

^**p≤0.01

Community Involvement

The study was carried out in consultation with key stakeholders for the aims of the project. Specifically, the development of the GDSQ parent and child versions, which is foundational to the study, was informed by an active group of autistic youth and adult gender-diverse and cisgender community members (Pelphrey, 2017; Strang et al., 2017). Moreover, several co-authors on the study team are members of the LGBTQ community.

Results

As shown in Table 1, there were no significant differences in age between groups. However, there were significant differences in IQ (p<0.001). The ASD sample presented with cognitive functioning solidly within the average range (mean 100.96); however, the TD group exhibited average-to-above-average intelligence (mean 116.93), a common challenge in academic research (e.g., Corbett, Muscatello, & Baldinger, 2019). There were no associations between IQ and the dependent variables (all p>0.05); thus, IQ was not included in the statistical models.

GDSQ - Parent

Parental differences were predicted between the ASD and TD groups across all GDSQ-P variables. There was a marginally significant difference with a small effect with the ASD parents endorsing more gender Body Incongruence, t(189.59)= −2.28, adjusted p=0.054, d=−0.30, compared to TD parent reports (Figure 1). Within-group analyses revealed significant sex differences with a large effect size in the ASD group such that parents of youth assigned female-at-birth reported more Incongruence than males, t(32.91)= −3.78, p<0.001, d=−1.10 (Figure 2). As shown in Table 4, there were no significant differences between the groups for GDSQ-P Expression, Identity, and Social Play. Furthermore, there were no significant differences on the CBCL Item-110, t(201.69)= −1.24, p=0.26, d=−0.15. These results appear in part consistent with May et al. (2017), Janssen et al. (2016), and Strang et al. (2014) in which their parental samples endorsed more gender diversity experiences in autistic children; however, there were no differences on the CBCL Item-110 as shown in these studies. Inconsistencies across studies lie in the effect sizes and the measures in which differences were found.

Figure 1.

GDSQ Parent-Report of Gender Incongruence Experiences in autism spectrum disorder (ASD) and typical development (TD) Groups.

Figure 2.

GDSQ Parent Report of Gender Incongruence for autism spectrum disorder (ASD) group.

Table 4.

Gender Diversity Means and Standard Deviations with Independent Samples T-Tests

Measure	TD	ASD	dfa	t	p-valueb	95% CI	d c
	M (SD)	M (SD)
GDSQ-P Expression	1.56 (1.76)	1.91 (1.87)	202	−1.37	0.24	[−0.85, 0.15]	−0.19
GDSQ-P Identity	0.13 (0.69)	0.24 (1.18)	202	−0.82	0.41	[−0.38, 0.14]	−0.12
GDSQ-P Social and Play Interests	3.77 (2.33)	4.25 (2.40)	202	−1.44	0.26	[−1.14, 0.17]	−0.20
GDSQ-P Incongruence	0.47 (0.78)	0.81 (1.36)	189.59	−2.28	0.05	[−0.64, −0.05]	−0.30
GDSQ-S Binary Diversity	0.53 (1.17)	1.11 (1.91)	223.21	−2.83	0.02	[−0.97, −0.17]	−0.35
GDSQ-S Nonbinary Diversity	0.20 (0.63)	0.73 (1.43)	191.15	−3.79	0.001	[−0.80, −0.25]	−0.46
CBCL 110	0.01 (0.10)	0.04 (0.22)	201.69	−1.24	0.26	[−0.07, 0.01]	−0.15

Note: ASD = Autism Spectrum Disorder; CBCL = Child Behavior Checklist; GDSQ-P = Gender Diversity Screening Questionnaire – Parent; GDSQ-S = Gender Diversity Screening Questionnaire – Self; TD = Typical Development.

^aWelch degrees of freedom when assumption of equal variances was violated.

^bBenjamini-Hochberg adjusted

^cCohen’s d

In regard to response rate on the CBCL Item-110, there were no significant differences (p>0.05); specifically, 2.9% of parents of children with ASD and 1% of parents of TD children reported gender diversity experiences. Also, 242 parents completed the CBCL (TD=103, ASD=139); whereas 206 parents completed the GDSQ (TD=90 and ASD=116) resulting in an 85.12% response rate or a 14.88% difference in completion rates on the GDSQ.

GDSQ - Self

Regarding self-report of gender diversity, there were significant differences and small-to-medium effects between the groups: autistic children endorsed higher rates of experienced Binary Gender Diversity, t(223.21)=−2.83, adjusted p=0.02, d=−0.35, and Nonbinary Gender Diversity, t(191.15)= −3.79, adjusted p=0.001, d=−0.46, which supports the hypothesis (Figure 3).

Figure 3.

GDSQ Self-Report of Gender Nonbinary and Gender Incongruence Experiences in autism spectrum disorder (ASD) and typical development (TD) Groups.

GDSQ Parent and Self-Report Correlations

Spearman’s correlations were conducted between GDSQ-P and GDSQ-S variables on the total sample. There were no significant correlations between the GDSQ-P and GDSQ-S (see Table 2).

Spearman’s rho correlations GDSQ-P and internalizing symptoms

To examine hypothesized association between gender diversity and internalizing symptoms, Spearman’s correlations were conducted using the GDSQ-P raw score indices that differentiated the groups and the CBCL Affective and Anxiety subscales and Columbia Suicidality Total. For the total sample, there were small positive correlations between Incongruence-P and CBCL Affective, 𝜌 = .25, p<0.001, CBCL Anxiety, 𝜌 = .23, p=0.001, and Suicidality Total, 𝜌 = .15, p=0.04.

Spearman’s rho correlations GDSQ-S and internalizing for Total Group

Spearman’s correlations were conducted between the CBCL Affective and Anxiety subscales and the GDSQ-S. For the total sample there were small positive correlations between GDSQ-S Binary Gender Diversity and Suicidality Total, 𝜌= .14, p=0.05.

GDSQ predicting Internalizing symptoms

To further analyze the extent of the relationship between GDSQ and internalizing symptoms, multivariable linear regression models predicting internalizing symptoms were conducted and adjusted for age, sex, and diagnosis. ASD diagnosis and higher Gender Body Incongruence were predictive of higher CBCL Affective T scores, 𝛽=10.42, adjusted p<0.0001, 95% CI [8.01, 12.83], and, 𝛽=1.56, p=0.02, 95% CI [0.49, 2.64], respectively.

ASD diagnosis was also predictive of CBCL Anxiety T scores, 𝛽=10.86, adjusted p<0.0001, 95% CI [8.48, 13.25]. While Gender Body Incongruence was correlated with Anxiety, when age, sex and diagnosis were included in the regression model, they were no longer statistically significant after correcting for multiple comparisons, 𝛽=1.33, adjusted p=0.06, 95% CI [0.27, 2.40].

For total Suicidality scores, age and ASD diagnosis were predictive without adjusting for false discovery rates but were no longer significant after adjustment.

Within group correlations

To explore associations within groups, correlations between GDSQ and CBCL were examined in the ASD and TD sample separately. For the ASD sample, GDSQ-P Incongruence was positively correlated with CBCL Affective, 𝜌= .22, p= 0.02, and Anxiety, 𝜌= .24, p= 0.01, domains. Yet, the GDSQ-S Nonbinary was negatively correlated with Affective, 𝜌= −.21, p= 0.03.

For the TD sample GDSQ-S Diversity was positively correlated with Affective, 𝜌 = .26, p=0.02, and Suicidality, 𝜌 = .36, p < 0.001.

Discussion

The aim of the study was to prospectively examine gender diversity experiences through self- and parent-report in well-characterized autistic and typically developing children. Comparisons between groups (ASD and TD) and respondents (parent and child) regarding endorsement of gender diversity and nonbinary experiences were analyzed. Additionally, the extent to which internalizing symptoms were related to gender diversity was explored.

There were significant differences in parent-reported gender Body Incongruence on the GDSQ-P in autistic children compared to TD children. The finding appears broadly consistent with existing studies (Janssen et al., 2016; May et al., 2017; Strang et al., 2014) in which large parental samples of autistic children reported more gender diversity experiences in their child compared to non-ASD comparison groups. Despite distinctions in Body Incongruence, no other parental differences emerged on the remaining GDSQ-P scales. Moreover, there were no differences on the CBCL Item-110 and a comparable percentage of parents of children with and without ASD reported gender diversity experiences based on the index. The response rate of 2.9% is smaller than existing studies, in which a range between 4.0–5.4% of parents endorsed gender diversity experiences in their autistic child based on this index (Janssen et al., 2016; May et al., 2017; Strang et al., 2014). Higher levels of body incongruence (GDSQ-P) reported by parents of autistic children along with the absence of differences on Item 110 of the CBCL could potentially be explained by the parents’ perceptions of negative reactions of the children to bodily changes related to puberty independent of gender identity incongruence.

Importantly, not all parents elected to complete the GDSQ-P, whereas most completed the CBCL. Specifically, 242 parents completed the CBCL yet only 206 parents completed the GDSQ resulting in a 14.88% difference in completion rates on the GDSQ. Some parents overtly stated they did not want to complete the GDSQ due to religious or undisclosed reasons. Conversely, none of the parents in the current sample who completed the CBCL selectively failed to complete Item-110. Taken together, under-reporting of gender diversity is plausible suggesting that cultural context may need to be considered in interpreting the parental report. It is noteworthy that the current dataset was from families in the Central Southern region of the United States, an area of relatively reduced tolerance for gender diversity and LGBTQ-related issues from both legal/policy and cultural perspectives (Campaign, 2017; Hughto, Murchison, Clark, Pachankis, & Reisner, 2016; Wang, Geffen, & Cahill, 2016). Indeed, both the Janssen and Strang studies were conducted in two of the most progressive areas in the United States (New York City, NY and Washington, DC), amongst the highest levels of acceptance of LGBTQ people (Campaign, 2017, 2018). It is also plausible that the GDSQ parent report instrument is not sensitive or may be perceived as complex or cumbersome to complete. Future studies examining a broader demographic in combination with parental feedback regarding completion of the tasks would help to disentangle these potential contribution factors.

Another intriguing finding emerged in the autism group based on the sex of the child at birth. Specifically, within group analyses revealed that parents of autistic youth assigned-female-at-birth reported more body incongruence than autistic males assigned-at-birth. Although the subsample was admittedly small, the finding seems consistent with a report of autistic adults showing a greater number of females endorsing atypical gender roles (Cooper, Smith, & Russell, 2018). Similarly, Walsh, Krabbendam, Dewinter, and Begeer (2018) found a higher rate of transgender and nonbinary identities in autistic adults especially females assigned-at-birth. Notwithstanding having fewer autistic females in the study and a smaller sex-based ratio in autism (APA, 2013), a higher percentage of females exhibiting characteristics of body incongruence is notable.

Regarding the child self-reports, there were significant group differences with autistic children endorsing greater gender diverse experiences, both binary and nonbinary, compared to TD peers, which supported the hypothesis. These results are broadly consistent with the literature to date showing higher rates of gender diversity incongruence in autistic youth than comparison age groups with no ASD symptoms (Janssen et al., 2016; Nabbijohn et al., 2018; Strang et al., 2014). However, this is the first study to ask children of this age-range themselves about their experienced gender, including both binary and nonbinary experiences.

While there is strong emerging literature revealing that autistic individuals experience greater gender diversity, there is less research focused on gender nonbinary experiences. Participants with ASD in the current study endorsed significantly higher rates of nonbinary experiences. These findings are consistent with Cooper and colleagues (2018) demonstrating a large percentage of autistic adults identify as nonbinary. For example, 26.5% of the autistic people assigned female at birth who experienced gender diversity endorsed some degree of nonbinary identification.

One of the important aims of the current study was to determine the extent to which parents and children endorse gender diversity experiences. GDSQ-P Gender Identity and GDSQ-S Binary Gender Diversity were not correlated, suggesting inconsistency across respondents.

Another important aim of the study was to examine associations between endorsement of gender diversity experiences and internalizing symptoms. For the total sample parent-report, there were small positive correlations between gender incongruence experiences compared to measures of anxiety, depression and suicidal symptoms suggesting the more body dysphoria experiences the higher presentation of internalizing symptoms. The youth with ASD demonstrated elevated internalizing symptoms and suicidality relative to the TD participants.

Within-group correlations revealed that this was primarily reported by the parents of autistic youth. For the self-report indices, within-group correlations revealed unique profiles. For the TD sample, GDSQ-S Binary Gender Diversity was positively associated with symptoms of depression and suicidality, but not in the ASD sample. Moreover, gender nonbinary experience in the autistic youth was negatively correlated with symptoms of depression. These findings suggesting that children who experience body-related gender dysphoria are at an increased risk of depressive symptoms. While these findings seem to be in opposition with previous studies in transgender allistic and autistic adolescents (Strang et al., 2021), nonbinary youth were not included, emphasizing the need for further studies to understand how gender identity and gender roles may contribute to affective symptoms in autistic youth.

The aforementioned findings showing higher rate of gender variance and internalizing symptoms in ASD may be similar to Strang and colleagues (2014) who found that children with ASD and co-occurring “gender variance” had reduced anxiety symptoms as compared to youth with allistic (i.e., non-autistic) neurodevelopmental conditions and co-occurring “gender variance”. It is plausible that, due to their social differences, children with ASD may be less concerned with how they are perceived by others. In fact, children with ASD often show an atypical and diminished stress pattern to social evaluation by peers and adults (Corbett, Muscatello, Kim, Patel, & Vandekar, 2020; Corbett, Schupp, & Lanni, 2012; Edmiston, Blain, & Corbett, 2017; Lanni et al., 2012). Recently, Corbett and colleagues (2019) showed perception of facial affect (i.e., neutral affect) mediates the stress response to social evaluative threat in children with and without ASD. Thus, the extent to which a child is able to discern social threat impacts their emotional and physiological response to the context. While hypothetical, some autistic, gender-diverse children may experience less anxiety or depression than expected due to reduced awareness of social judgment. Of concern, however, is the longer-term trajectory of individuals who are autistic and gender diverse or dysphoric, as findings among adults suggest that the combination of autism and gender dysphoria results in the highest levels of mental health symptoms, greater than in autistic gender typical or allistic gender dysphoric adults individually (George & Stokes, 2018).

Importantly, the current study did not assess clinical gender dysphoria, though some autistic youth reported levels of Binary Gender Diversity on the GDSQ-S, which may be consistent with autistic youth who were gender-referred in another sample (Goldstein et al., 2017). However, the remaining parent and child variables were not associated. It may be the case that parent-reports are inherently observational whereas the child variables are self-reports of inner experiences that parents might not see or be aware of in their child, especially autistic children who may struggle to self-advocate and self-express (Strang, Powers, et al., 2018; Sturrock, Chilton, Foy, Freed, & Adams, 2021). Finally, as noted some parents refused to complete the GDSQ-P due to religious or undisclosed reasons highlighting the challenge in how to provide adequate support, education and recommendations regarding gender diversity within a cultural context that may not be receptive.

Identifying the impact of daily experience on behavioral health in youth with both characteristics of autism and gender diversity may improve understanding of sources for and risk factors of emotional distress. Additionally, examining gender experience over time is important (van Schalkwyk et al., 2015) as poignantly revealed by Strang, Powers, and colleagues (2018) who conducted in-depth interviews of 22 well-characterized autistic gender-diverse adolescents and demonstrated that gender experiences can be fluid. Beyond the unique challenges, interviewees in the study expressed shared perspectives, including: prepubertal gender nonconformity, intense gender dysphoria experiences, acknowledgement of the unique challenges of neurodiversity (ASD) and gender diversity co-occurrence, and challenges expressing their gender due to their neurodiversity characteristics. Recently, Cooper and colleagues (2021) conducted a qualitative study of 21 adults with ASD who identified as transgender or non-binary whose lived experience revealed several themes encompassing a lack of acceptance of both gender- and neurodiversity. Critically, puberty was noted as a particularly distressing developmental time with the occurrence of unwanted bodily changes.

Previously, no clinical care model has been in place (Van Der Miesen et al., 2016). Initial clinical guidelines have been proposed for youth with co-occurring ASD and GD or incongruence developed via expert guidance using a two-stage Delphi procedure (Strang, Meagher, et al., 2018). Primary themes from the guidelines include the importance of collaboration between gender and autism specialists, psychoeducation for children and parents, exploration of gender identity over time, an extended diagnostic period and consideration of potential medical treatments. A more recent community-based participatory procedure utilized ongoing needs assessments during a group program to identify key priorities of youth and parents for developing a clinical care model (Strang et al., 2020). Eleven clinical approaches were developed from youth and parent-input (Strang et al., 2020) along the themes of 1) targeting and supporting youth gender-related needs in group, 2) targeting broader support needs, 3) importance of youth connection, and 4) providing parents with a support group, too.

It is also imperative that autistic individuals who experience gender diversity continue to receive intervention services for social communication and self-advocacy skills. Autism-related neurodiversity may complicate or magnify challenges with both gender discernment and advocacy (Strang, Powers, et al., 2018); therefore, interventions that provide positive, peer-mediated social support may be valuable (e.g., Corbett et al., 2016). Also, training in psychosexual education that includes information about gender expression and gender identity diversity and is explicitly developed for adolescents on the spectrum, such as Tackling Teenage Training, may be valuable (Dekker et al., 2015; Visser et al., 2015).

Limitations and Future Directions

Despite the compelling findings, there are limitations to acknowledge. The study may be viewed as preliminary as it included a relatively small sample size compared to studies that have employed chart review (Janssen et al., 2016; Strang et al., 2014) or online reporting methods (van der Miesen, de Vries, et al., 2018). A strength of the study lies in the careful characterization of the children. Another limitation is that the sample consisted of primarily prepubertal children between 10-to-13 years limiting generalizability beyond this age range. Importantly, the sample is part of a longitudinal study, which provides the opportunity to follow the cohort over this critical developmental period. The study did not provide comprehensive interviews with participants or attempt to carefully characterize the presence of gender dysphoria (APA, 2013; Strang, Powers, et al., 2018) because the aim was to sample gender experiences via self- and parental-report. Thus, the extent to which the findings extend to individuals who meet criteria for co-occurring ASD and GD is undetermined. While the current study examined correlation between parent- and youth-scores at the domain-level, future studies are expected to assess consistency between groups in item-level responses in a larger sample of youth and adults. Within-group results suggest associations between gender diversity and depression/anxiety; however, examination of diagnosis and gender variance interactions in predicting internalizing profiles will be necessary in a larger sample to elucidate the nature of gender experiences and mental health outcomes in autistic youth. In addition, the gender distribution of the sample was significantly biased toward autistic male children and the racial distribution was disproportionately white. Therefore, future studies are warranted which more representative based on sex (e.g., females assigned at birth) and race (e.g., Black, Asian).

In summary, results extend recent reports showing increased rates of experienced gender diversity in autistic children based on self-report and parent-report. Findings corroborate clinical and research observations and underscore the need to better understand and support the unique and complex needs of autistic children who experience gender diversity. It is apparent that the co-occurrence may result in greater clinical, theoretical and ethical challenges. Nevertheless, initial clinical guidelines are emerging for children and adolescents experiencing ASD and GD (Strang et al., 2020; Strang, Meagher, et al., 2018) to create a framework for supportive assessment and treatment.