912147

research-article2020
AUT0010.1177/1362361320912147AutismBeck et al.

Mental Health Across the Lifespan

Autism

Looking good but feeling bad: 2020, Vol. 24(4) 809­–821

© The Author(s) 2020
Article reuse guidelines:
“Camouflaging” behaviors and mental sagepub.com/journals-permissions
DOI: 10.1177/1362361320912147
https://doi.org/10.1177/1362361320912147

health in women with autistic traits journals.sagepub.com/home/aut

Jonathan S Beck , Rebecca A Lundwall, Terisa Gabrielsen,

Jonathan C Cox and Mikle South

Abstract
Autistic traits are associated with frequent psychological distress, suicidal ideation, and everyday functional challenges.
These associations may be especially prominent in women. Some women with autistic traits “camouflage” feelings of
awkwardness in typical social situations by effortfully engaging in expected social behaviors. We explored camouflaging
because emerging evidence posits an association between camouflaging behaviors and poorer outcomes related to
mental health, daily functioning, and access to mental health care. We recruited a sample of 58 women (age M = 25 years;
IQ M = 115) who reported that they find social situations confusing and who scored high on a measure of broad
autistic traits (Broad Autism Phenotype Questionnaire score > 3). The majority of participants reported significant levels of
psychological distress, suicidal ideation, and daily functioning difficulties. Regression models showed that camouflaging
efforts (Camouflaging Autistic Traits Questionnaire) and autistic traits (Social Responsiveness Scale, Second Edition) modestly
but significantly predicted psychological distress and functional challenges, respectively. In a subgroup of participants with
high Camouflaging Autistic Traits Questionnaire scores, camouflaging scores were significantly associated with psychological
distress and functional challenges. Camouflaging may help predict when clinicians should be concerned about higher
mental health distress in autistic women and may be beneficial to measure as part of a comprehensive, multimethod
assessment of mental health in women who report difficulties fitting into social situations.

Lay Abstract
Women who try to hide or “camouflage” their autistic traits are likely to report that they feel distressed, think of
suicide, and/or struggle to function in everyday life. We asked 58 women with autistic traits to complete questionnaires
about camouflaging and mental health. Most of these women did not have a formal diagnosis of autism, yet a majority
reported that they camouflaged autistic traits, and a large majority reported significant mental health challenges. Some
researchers have suggested that women with autistic traits are more likely than autistic men to experience mental health
challenges because women may try more to “fit in” socially by camouflaging their autistic traits. Analyses showed that
camouflaging was associated with feeling distressed (depressed, anxious, and/or stressed). For women who reported
above-average levels of camouflaging, camouflaging was also associated with having thoughts about suicide and struggling
to function in everyday life. Trying to camouflage autistic traits was associated with mental health challenges, regardless
of whether those traits were very mild or more severe. The findings of this study may influence how mental health
professionals evaluate and treat women with autistic traits.

Keywords
autism spectrum, camouflage, diagnosis, disorders, female, mental health, women

There is growing concern that the mental health of many Brigham Young University, USA
autistic women is poor, and perhaps worse than that of
autistic men (e.g., Tint et al., 2017). Consistent with Corresponding author:
Rebecca A Lundwall, Department of Psychology and Neuroscience
reported sex differences in mental health in the general Center, Brigham Young University, 1064 KMBL, Provo, UT 84602,
population (Kessler et al., 2005), autistic females are more USA.
likely than autistic males to experience certain psychiatric Email: [email protected]
810 Autism 24(4)

concerns such as anxiety and depression (Lai et al., 2019; community (Hickey et al., 2018; Hurlbutt & Chalmers,
Mandy et al., 2012; Oswald et al., 2016; Solomon et al., 2002; Kanfiszer et al., 2017; Leedham et al., 2020; Portway
2012). Suicide rates among autistic women are reported to & Johnson, 2005; Webster & Garvis, 2017).
be much higher than in other clinical and non-clinical pop- Whether or not they have been previously diagnosed,
ulations (Cassidy et al., 2014; Hirvikoski et al., 2019). autistic women and men often report attempting to “cam-
And, in contrast to general population trends that show ouflage” themselves in social situations (Hull et al., 2019).
women are less likely than men to die by suicide Their camouflaging efforts may involve suppressing autis-
(Hedegaard et al., 2018), autistic women are just as likely, tic tendencies and/or effortfully engaging in “typical”
if not more likely, to die by suicide as autistic men social behaviors (Hull et al., 2017). For example, to appear
(Hirvikoski et al., 2016; Kirby et al., 2019). It is not well more typical, an autistic woman may avoid talking too
understood if and how internal distress and suicidality much about a favorite topic or may practice looking at a
impact the daily functioning (e.g., performing work tasks, conversation partner’s facial expressions before respond-
maintaining relationships, and engaging with the commu- ing. Broadly speaking, camouflaging efforts themselves
nity) of autistic women (Howlin & Magiati, 2017; Kirby are typical social behaviors in that neurotypical adults
et al., 2016; Taylor et al., 2015); evidence from non-clini- report employing camouflaging strategies that are qualita-
cal and majority-male autistic samples suggests that the tively similar to the strategies used by autistic women and
impact may be significant and negative (Katon, 2009; men (Livingston et al., 2019). Adaptive camouflaging
Wallace et al., 2016). To support the mental health of autis- efforts may help autistic women fit in and avoid a range of
tic women, there is an urgent need to better understand the social stressors associated with mental health concerns.
factors that significantly impact their psychological dis- For instance, a woman who consciously avoids talking
tress, suicidality, and functional capacity. excessively about topics of special interest may be more
One factor that may impact the mental health of autistic likely to make and keep friends, and a woman who effort-
women is the severity of autistic traits. While few studies fully makes eye contact and carefully evaluates a conver-
have recruited substantial samples of autistic women, stud- sation partners’ facial expressions may be more likely to
ies of autistic individuals across the lifespan suggest that interview well and obtain employment. Camouflaging
autism spectrum disorder is associated with increased risk may also produce less-tangible benefits. One qualitative
for psychiatric concerns, such as anxiety, attention-deficit/ study quoted an autistic woman who shared that she feels
hyperactivity, and affective disorders (Lai et al., 2019; camouflaging is rewarding for her because she “[shows
Leitner, 2014; Lever & Geurts, 2016; Nylander et al., people] that autistic people can have people skills and be
2018; Stewart et al., 2006; White et al., 2009). Even mild, good role models” (Hull et al., 2017). The numerous short-
subclinical autistic traits (i.e., traits from the broader and long-term benefits of camouflaging efforts reinforce
autism phenotype; Piven et al., 1997) are associated with the behavior and make future efforts more likely.
increased prevalence of psychiatric concerns (e.g., Cassidy, Despite the advantages of successful camouflaging,
Bradley, Shaw, & Baron-Cohen, 2018; Ingersoll & camouflaging efforts may also contribute to poor mental
Hambrick, 2011; Nylander et al., 2018; Pine et al., 2008). health, and the relationship between camouflaging efforts
Women with autistic traits may experience psychiatric and mental health in autistic women is of interest for sev-
concerns that are unrelated to those traits, of course, and eral reasons. First, while camouflaging is not unique to
the relationships between autistic traits and mental health autistic women, it seems to be especially common among
concerns remain unclear (Bolton et al., 1998; Piven & them (Cassidy, Bradley, Shaw, & Baron-Cohen, 2018;
Palmer, 1999). Hull et al., 2020; Schuck et al., 2019). Second, women
Difficulty fitting in with neurotypical social situations who camouflage often report that they experience mental
may also contribute to mental health concerns. Such diffi- health concerns such as low self-esteem, stress, exhaus-
culty can lead to autistic women experiencing reduced tion, anxiety, and depression (Cage & Troxell-Whitman,
support in their social networks, which in turn increases 2019; Milner et al., 2019). Therefore, camouflaging may
the likelihood of mental health concerns (see Howlin & partly explain the prevalence of internalizing disorders and
Magiati, 2017). While some autistic women state that they suicidality in autistic women. Third, camouflaging may
are unconcerned by such social challenges, many others also partly explain why autistic women are diagnosed with
are keenly aware of these challenges and feel different, autism later, on average, than their male counterparts
misunderstood, or lonely (Bargiela et al., 2016; Gotham (Begeer et al., 2013). For some women, social challenges
et al., 2014). This may be especially true for undiagnosed may outstrip camouflaging abilities only later in life
autistic women who are more likely to attribute social (Giarelli et al., 2010). Also, just as a depressed patient who
challenges to personal traits rather than to traits shared presents a smile may not be referred for treatment (despite
with other autistic people, with the former often leading to significant internal distress), an autistic woman who suc-
poor self-concept and feelings of isolation, and the latter cessfully camouflages and presents as socially competent
potentially leading to self-awareness and a sense of may not be referred for clinical assessment. Fourth, while
Beck et al. 811

interventions for autistic adults potentially ameliorate dis- Table 1. Participant demographics.
tress and support successful functioning (Jones et al.,
n (%)
2014), many interventions encourage camouflaging
behaviors that may perpetuate distress. Previous autism diagnosis
Although camouflaging is theoretically associated with Diagnosed in childhood/adolescence 14 (24.1)
mental health benefits and risks, investigations of the rela- Diagnosed in adulthood 4 (6.9)
tionship between camouflaging and mental health in autistic Educational background
women have yielded mixed but generally negative associa- Some high school 2 (3.4)
tions. One study of combined autistic men and women High school diploma or GED 4 (6.9)
showed that autistic men (n = 30) who manifested camou- Associate's degree 5 (8.6)
flaging were more likely to report depression (r = 0.3), but College student 27 (46.6)
Bachelor’s degree 15 (25.9)
camouflaging among autistic women (n = 30) was not asso-
Graduate degree 5 (8.6)
ciated with depression (Lai et al., 2017). However, another
Relationship status
study (N = 306; approximately 1/3 men) showed significant
Single 36 (62.1)
associations between self-reported camouflaging and symp-
In a dating relationship 7 (12.1)
toms of depression and anxiety (Hull et al., 2019). Another Married or engaged 13 (22.4)
recent study of autistic adults (N = 262; 52% women) Divorced or separated 2 (3.4)
reported that individuals who camouflage inconsistently Employment status
across contexts (e.g., camouflage at work but not in less for- Unemployed 13 (22.4)
mal contexts) experience greater psychological distress Part-time work 40 (69.0)
(Cage & Troxell-Whitman, 2019). Regarding suicidality, Full-time work 5 (8.6)
some research suggests that camouflaging may play a role Socioeconomic background
(Camm-Crosbie et al., 2019; Cassidy, Bradley, Shaw, &  At least one parent earned a bachelor’s 50 (86.2)
Baron-Cohen, 2018; Kirby et al., 2019). Cassidy, Bradley, degree
Shaw, and Baron-Cohen (2018) reported that, for autistic  At least one parent earned an advanced 33 (56.9)
men and women (N = 164) who were first diagnosed in degree
adulthood, self-reported camouflaging efforts were predic- Both parents earned bachelor's degrees 25 (43.1)
tive of suicidality even after controlling for sex, depression, Both parents earned advanced degrees 5 (8.6)
anxiety, age, and employment. Finally, camouflaging efforts Ethnicity
may contribute to challenges in functioning well at work, Hispanic or Latino 4 (6.9)
Not Hispanic or Latino 54 (93.1)
home, and other contexts. A majority of autistic women with
Race
average intellectual functioning may struggle to secure or
Asian 1 (1.7)
maintain full-time employment and so rely on extensive
Black or African American 1 (1.7)
support from family (Engström et al., 2003; Howlin, 2000).
White 55 (94.8)
To our knowledge, no study has directly investigated the More than one race 1 (1.7)
relationship between camouflaging in women and everyday
functional challenges. GED: General Education Development.

Study aims mental health in autistic women is to help clinicians know

whom to reach out to and how to ease distress and improve
In response to calls to explore mental health concerns in functioning, perhaps by assessing and addressing prob-
autistic female samples that include undiagnosed autistic lematic camouflaging.
women (see Halladay et al., 2015; Livingston & Happé,
2017), we recruited a sample that included many women
in a diagnostic “gray zone” who potentially are undiag- Methods
nosed because of efforts to camouflage their autistic traits.
All participants reported feeling awkward and confused in
Measures
social situations. Approximately one third of participants Screening measure. The Broad Autism Phenotype Ques-
had previous autism diagnoses (see Table 1). In this unique tionnaire (BAPQ; Hurley et al., 2007) is a 36-item self-
sample, we aimed to model psychological distress, suici- report measure of autism-related traits (e.g., “I like being
dality, and daily functioning challenges as predicted by around other people”) with high inter-item reliability
camouflaging behavior, while controlling for autistic traits. (Cronbach’s α = 0.95). We chose the BAPQ as a screening
We specifically hypothesized that camouflaging efforts measure because, in contrast to other autism measures
would significantly predict poor mental health even after administered for this project, it was designed to be sensi-
accounting for clinician-rated autism severity and self- tive even to subclinical autistic traits. In the Hurley et al.
reported autistic traits. Our aim in seeking predictors of (2007) validation sample (n = 150 women), using a total
812 Autism 24(4)

score cutoff of 3.25 yielded 75% sensitivity and 87% spec- original 36-item assessment, the WHODAS 2.0 explains
ificity for autistic personality traits in female relatives of 81% of the variance of the 36-item version, has excellent
autistic individuals. A more recent study (n = 18 women) test–retest reliability (intra-class coefficient of 0.98), and
used a lower total score cutoff of 3.17 and reported 33% has been validated in autistic adults without intellectual
sensitivity and 100% specificity for women (Sasson et al., disability (Park et al., 2019).
2013). In order to include all potential participants with
significant autistic traits, we valued sensitivity over speci- Clinician-rated autistic traits. The Autism Diagnostic Obser-
ficity and used a liberal cutoff of 3.0. vation Schedule, Second Edition, Module 4 (ADOS-2; Lord
et al., 2012) is a valid and reliable instrument that is interna-
Psychological distress. The Depression Anxiety Stress Scales tionally recognized as the gold-standard diagnostic measure
21 (DASS-21; Lovibond & Lovibond, 1995) is the 21-item for adolescents and adults with ASD (Gotham et al., 2006).
form of the 42-item self-report DASS. The DASS-21 invites Due to its superior sensitivity (90.5%) and specificity
respondents to rate their experience of seven symptoms (82.2%), we used the revised algorithm to classify partici-
each of depression, anxiety, and stress over the past week. pants as autistic or not (Hus & Lord, 2014). The examiners
The DASS-21 total score is a valid measure of general psy- all had received formal clinical ADOS-2 training and were
chological distress with high inter-item reliability (Cron- supervised by clinicians who have attained ADOS-2 research
bach’s α = 0.93; Henry & Crawford, 2005). reliability. ADOS-2 administrations were video recorded,
and 86% were scored by a second clinician. Raters achieved
Suicidality. The Suicidal Behavior Questionnaire–Revised mean inter-rater reliability of 0.88 across all items.
(SBQ-R; Osman et al., 2001) is a brief, four-item self-
report measure of suicidal behavior that has been validated Self-reported autistic traits. The Social Responsiveness
in clinical and non-clinical samples. Two items assess life- Scale, Second Edition, Adult Self-Report (SRS-2; Constan-
time incidence of suicidal ideation, threats, and attempts; tino, 2012) is a 65-item self-report questionnaire that asks
one item assesses suicidal ideation over the past year; and respondents to consider their behavior over the past 6
one item assesses the probability of future attempts. For months and rate the presence of a variety of autistic traits,
the future-oriented item (“How likely is it that you will including social communication deficits, restricted inter-
attempt suicide someday?”), two of the six response ests, and repetitive behaviors. Mandell et al. (2012)
options are similar: “Never” and “No chance at all.” With reported that a cut score of > 84 (raw total score), which is
the approval of Augustine Osman, PhD, one of the meas- interpreted within a moderate-to-severe severity range,
ure’s authors, we changed the wording of these response maximized specificity (0.60) and sensitivity (0.86). Test–
options to “Never thought about suicide” and “Thought retest correlations range from 0.88 to 0.95, and across par-
about suicide, but no chance at all.” For a non-clinical, ents, spouses, and relatives, inter-rater reliability
undergraduate sample, a total cutoff score of seven differ- coefficients range from 0.61 to 0.92 (Bruni, 2014).
entiates well between those at risk for suicide and non- The Autism-Spectrum Quotient (AQ; Baron-Cohen
suicidal adults, with sensitivity of 93% and specificity of et al., 2001) is a 50-item self-report measure of personal
95%. The SBQ-R has not yet been validated in an adult traits related to autism, such as social skills, attention
autistic sample, but neither has any other suicide screening switching, attention to detail, communication, and imagi-
measure (Cassidy, Bradley, Bowen, et al., 2018). nation. The authors reported good test–retest (r = .70) and
inter-rater (no significant differences between adult- and
Functional challenges. The World Health Organization Dis- parent-report) reliability (Baron-Cohen et al., 2001). A
ability Assessment Schedule, Second Edition (WHODAS total score cutoff of > 32 distinguishes most male and
2.0; Üstün, 2010) uses 12 face-valid questions to measure female adults with autism from controls.
disability or dysfunction in six areas: cognition, mobility,
self-care, getting along, life activities (e.g., “taking care of Camouflaging. The Camouflaging Autistic Traits Question-
household responsibilities”), and community participa- naire (CAT-Q; Hull et al., 2019) is a 25-item self-report
tion. The respondent rates how much difficulty (ranging measure of social camouflaging behaviors (i.e., strategies
from “None” to “Extreme or Cannot do”) has been caused used to compensate for or mask autistic characteristics to
by health conditions over the past 30 days for each of the facilitate social assimilation). Unlike measures of camou-
12 tasks. The measure explicitly includes “mental or emo- flaging derived from behavioral observations, the CAT-Q
tional problems” that are “short or long lasting” in its defi- measures both successful and failed camouflaging attempts
nition of health conditions. Before being entered into based on intention. The CAT-Q was developed for online
regression models, total scores reflecting percent decreased administration, and the normative sample included 354
functioning were multiplied by 100 for ease of interpreta- autistic and 478 non-autistic adults (recruited through
bility (i.e., a score of 100 means the respondent cannot do social media, a clinic, and word-of-mouth, similar to our
any of the 12 activities). The 12-item version of the sample). Internal consistency (α = 0.94) and test–retest
Beck et al. 813

reliability (r = 0.77) are acceptable. In the measure’s vali-

dation sample of autistic adults, the mean total score was 156 expressed interest in participating
4.79 (SD = 0.99). The measure’s authors reported that Phone intake -11 declined to participate
-15 ineligible (e.g., not fluent in English)
autistic adults scored higher on the CAT-Q than non-autis-
tic adults, and autistic females scored higher than autistic
males (Hull et al., 2019, 2020). 130 were invited to complete the online
survey
Initial online survey
-22 did not complete the survey
IQ. The Wechsler Abbreviated Scale of Intelligence, Sec- -10 scored < 3.0 on the BABQ

ond Edition (WASI-II; Wechsler & Hsiao-Pin, 2011) is a

valid, brief measure of IQ. We administered the two-scale 98 were invited to complete an in-person
full-scale IQ (FSIQ-2) which includes the Vocabulary and In-person visit visit
68 completed the in-person battery
Matrix Reasoning subtests. The WASI-II has excellent psy-
chometric properties, including test–retest reliability of
0.88 for FSIQ-2 in adults. Post-visit online 68 were invited to complete a post-visit
survey survey
(CAT-Q, AQ) 58 completed it
Procedure
All study procedures were approved by the university
Figure 1. Summary of the process of sample formation.
Institutional Review Board (IRB), and written consent BAPQ: Broader Autism Phenotype Questionnaire; AQ: Autism-Spectrum
was obtained for all participants in accord with the Quotient; CAT-Q: Camouflaging Autistic Traits Questionnaire. Of the 30
Declaration of Helsinki. This project was part of a larger women who completed the initial online survey but not the in-person
visit, about half did not respond to outreach or reported they had
series of studies that included additional self-report, moved out of state.
behavioral, and parent-report measures detailed in the
supplementary materials.
Women participants were recruited through university- Table 2. Participant characteristics.
based and private mental health clinics, social media, and M (SD) Range
an intense residential support program for autistic young
adults. The recruiting flyer read, “In social situations, do Age in years 25.17 (6.17) 18–42
you often find yourself confused, anxious, or exhausted?” FSIQ-2a 114.60 (11.27) 89–140
An existing autism diagnosis was not required for partici- CAT-Q total score 4.73 (0.86) 2.32–6.28
pation. Interested participants initially completed self- DASS-21 total score 55.28 (20.78) 12–102
report questionnaires using the Qualtrics online survey SBQ-R total score 7.98 (3.48) 3–17
WHODAS 2.0 total score 0.27 (0.14) 0.20–0.63
platform (Qualtrics, Inc., Provo, Utah), including BAPQ,
ADOS-2 total score 6.07 (5.41) 0–18
SRS-2, DASS-21, and a custom questionnaire regarding
Self-reported autistic traits
demographic information, treatment history, and additional
SRS-2 raw total score 90.24 (22.36) 33–139
topics related to health and social history. Women with a
AQ total score 28.57 (6.56) 13–43
BAPQ total score > 3.0 indicating the presence of autistic BAPQ total score 3.90 (0.45) 3.08–5.28
traits (3.0 is a liberal cutoff; see Sasson et al., 2013) were
invited to our lab to complete an in-person assessment FSIQ-2: Two-scale full-scale IQ; CAT-Q: Camouflaging Autistic Traits
involving the WASI-II, ADOS-2, WHODAS 2.0, and SBQ. Questionnaire; DASS-21: Depression Anxiety Stress Scales 21; SBQ-R:
Suicidal Behavior Questionnaire–Revised; WHODAS 2.0: World Health
Following the assessment, researchers invited participants Organization Disability Assessment Schedule, Second Edition; ADOS-2:
to complete the AQ and CAT-Q, also through Qualtrics. See Autism Diagnostic Observation Schedule, Second Edition; SRS-2: Social
Figure 1 for a summary of the sample formation process. Responsiveness Scale, Second Edition; AQ: Autism-Spectrum Quotient; BAPQ:
Participants were compensated $10/hour for their participa- Broader Autism Phenotype Questionnaire.
a
One participant did not complete IQ testing. She had earned a
tion in both online and in-person portions. bachelor’s degree, suggesting at least average intellectual functioning.

Participants were diagnosed in childhood, 4 in adolescence, and 4 in

The final sample reported here includes 58 adult women adulthood. A majority of participants (63.8%) reported
who completed all measures for this study. Participants receiving a formal psychiatric diagnosis other than autism
had a mean age of 25.2 years (range: 18–42; SD = 6.17; see spectrum disorder, such as generalized anxiety disorder
Table 2) and mean IQ of 114.60 (range: 89–140; SD = 11.27). (39.7%), major depressive disorder (31.0%), or attention-
All participants scored high on a measure of broad autistic deficit/hyperactivity disorder (17.2%; see Table 3). Most
traits (BAPQ score > 3). Eighteen participants (31%) had participants were White (94.8%), working part time
received a previous, formal diagnosis of autism spectrum (69.0%), and single or divorced (65.5%), and many were
disorder (autism or Asperger syndrome). Of these 18, 10 currently attending college (46.6%; see Table 1).
814 Autism 24(4)

Table 3. Participant-reported previous psychiatric diagnoses.

Diagnosis Full sample Previously diagnosed Not previously

(n = 58) with ASD diagnosed with ASD
(n = 18) (n = 40)

n (% sample) n (% group) n (% group)

Any psychiatric diagnosis (excluding ASD) 37 (63.8) 15 (83.3) 22 (55.0)
Generalized anxiety disorder 23 (39.7) 9 (50.0) 14 (35.0)
Major depressive disorder 18 (31.0) 6 (33.3) 12 (30.0)
ADHD 10 (17.2) 6 (33.3) 4 (10.0)
Social anxiety disorder 9 (15.5) 5 (27.8) 4 (10.0)
Obsessive-compulsive disorder 7 (12.1) 5 (27.8) 2 (5.0)
Eating disorder 4 (6.9) 1 (5.6) 3 (7.5)
Learning disorder 3 (5.2) 3 (16.7) 0 (0)
Specific phobia 3 (5.2) 2 (11.1) 1 (2.5)
Personality disorder 2 (3.4) 0 (0) 2 (5.0)
Bipolar disorder 1 (1.7) 0 (0) 1 (2.5)
Trichotillomania 1 (1.7) 0 (0) 1 (2.5)

ASD: autism spectrum disorder; ADHD: attention-deficit/hyperactivity disorder.

Of the two participants who reported a previous diagnosis of a personality disorder (PD), one had been diagnosed with borderline PD and the
other with schizotypal PD. Of the four participants who reported a previous diagnosis of an eating disorder, two indicated restrictive eating habits,
one bulimia, and one binging.

Data analyses
We tested ordinary least-squares linear regression models
of psychological distress (DASS-21), suicidality (SBQ-R),
and functional challenges (WHODAS 2.0) with clinician-
rated autistic traits (ADOS-2), insight into autistic traits
(SRS-2, AQ), and camouflaging (CAT-Q) as potential pre-
dictor variables. To examine associations between entered
dependent and predictor variables, we generated scatter-
plots to confirm linearity. Given the early stage of research
about these constructs, we completed simultaneous (ver-
sus hierarchical) regressions. Since camouflaging might Figure 2. DASS-21 subscale scores — full sample (N = 58).
A substantial majority of participants reported experiencing clinically
differentially predict distress depending on the level of significant psychological distress across subscales, with more than
severity of autistic traits (e.g., efforts to camouflage more one-third reporting severe or extremely severe levels of concern. The
severe traits may be more psychologically taxing), we numbers on the stacked bars represent percentages of the sample that
entered interaction terms in regressions involving the rel- scored in the respective interpretive range.
evant predictors. Before creating the product terms that
represent interactions, we mean-centered the involved using Stata version 15.1 (StataCorp, 2017). Deidentified
quantitative predictors (Warner, 2013). Centered scores data are available upon request.
were not entered as predictors. Shapiro–Wilk tests showed
normality in the distributions of regression residuals
(Thode, 2002), scatterplots of residuals against predicted/
Results
fitted variables did not show notable heteroscedasticity, This sample of women with significant self-reported autis-
and collinearity was not problematic as the variance infla- tic traits reported considerable psychological distress. On
tion factors (VIF) for all variables was < 3 (O’Brien, the DASS-21, 62% reported moderate or severe depres-
2007). sion, 67% reported moderate or severe anxiety, and 66%
There were no missing data among analyzed variables. reported moderate or severe stress (Figure 2). According to
Before regression analyses, extreme scores were Winsorized the SBQ-R, 62% of participants were at risk for suicide.
to 5th and 95th percentile values (Dixon & Yuen, 1974). This distress comes alongside significant everyday func-
Re-analysis of un-Winsorized data did not significantly tional challenges. Scores on the WHODAS 2.0 indicate that
change our findings. Power analyses are reported in the sup- 57% rated themselves as functioning in a range of every-
plementary materials. Statistical analyses were implemented day activities at three-fourths or less of their full capacity,
Beck et al. 815

Table 4. Linear regression models of mental health (N = 58).

Criterion Predictor B SE B t p β F df p R2 R2adj

Psychological distress 2.30 (6, 51) .048 .21 .12
ADOS-2 –.48 .50 –.95 .35 –.12
SRS-2 .30 .15 1.97 .06 .30
AQ –.50 .51 –.99 .33 –.16
CAT-Q 7.12 3.16 2.26 .03 .29
SRS-2 × CAT-Q –.19 .16 –1.14 .26 –.23
AQ × CAT-Q .18 .60 .30 .77 .06
Suicidality 1.43 (6, 51) .22 .14 .04
ADOS-2 .01 .08 .17 .86 .02
SRS-2 .04 .03 1.52 .14 .24
AQ .11 .09 1.25 .22 .21
CAT-Q –.33 .53 –.62 .54 –.08
SRS–2 × CAT-Q –.01 .03 –.54 .59 –.11
AQ × CAT-Q .07 .10 .72 .47 .15
Functional challenges 2.56 (6, 51) .03 .23 .14
ADOS-2 –.003 .003 –.98 .33 –.13
SRS-2 .002 < .001 2.17 .04 .33
AQ .001 .003 .33 .75 .05
CAT-Q .038 .020 1.90 .06 .24
SRS-2 × CAT-Q –.001 .001 –1.01 .32 –.20
AQ × CAT-Q .005 .004 1.25 .22 .25

Psychological distress = Depression Anxiety Stress Scales 21 total score; Suicidality = Suicidal Behavior Questionnaire–Revised total score; Functional
challenges = World Health Organization Disability Assessment Schedule, Second Edition total score; ADOS-2: Autism Diagnostic Observation Schedule, Second
Edition; SRS-2: Social Responsiveness Scale, Second Edition; AQ: Autism-Spectrum Quotient; CAT-Q: Camouflaging Autistic Traits Questionnaire.

and 15% of that group rated themselves as functioning at Of the three CAT-Q subscales, participants were most ele-
half or less. Participants were most likely to report diffi- vated on the masking subscale that assesses “strategies
culty on WHODAS 2.0 items related to social tasks (e.g., used to hide autistic characteristics or portray a non-autis-
“How much of a problem did you have in joining in com- tic persona,” with 74% of the sample reporting masking
munity activities . . .?”) or emotional health (“How much above the autistic sample’s mean (see Table S1).
have you been emotionally affected by your health
problems?”).
During the in-person assessment, many participants
Regression analyses
expressed their distress to examiners with comments such We ran three regression models. A regression model of
as “I’m more worried about making a [social] mistake psychological distress (DASS-21 total scores) was statisti-
than dying” or “I was tired of trying [to succeed socially] cally significant, F(6, 51) = 2.30, p = .048, and accounted
and making social mistakes, so I started avoiding people.” for 12% of the variance in distress (see Table 4). The
One participant who had a formal diagnosis of autism model showed that camouflaging efforts (CAT-Q total
shared that she felt “ashamed” about her diagnosis. scores) significantly predict such distress in the sample
Another described how she was glad to receive a diagno- (β = 0.29, p = .03), and self-reported autistic traits (SRS-2
sis, although it did not make social interactions easier: “I raw total scores) approached significance as a predictor
always wondered what was wrong with me. After my (β = 0.30, p = .06). A regression model of suicidality (SBQ-
diagnosis, I realized nothing was wrong, things are just R scores) was not statistically significant, F(6, 51) = 1.43,
harder for me. It is nice to have a reason for things. Now I p = 0.22. A regression model of daily functional chal-
know it is not my fault.” Further evidencing their distress, lenges (WHODAS 2.0 total scores) was statistically sig-
72% of participants had received psychotherapy, 74% had nificant, F(6, 51) = 2.56, p = 0.03, and accounted for 14%
taken medication for mental health concerns, and 64% of the variance in such challenges. In this model, self-
were receiving pharmacological or psychotherapy treat- reported autistic traits (SRS-2 raw total scores) signifi-
ment at the time of their participation in the study. cantly predicted functional challenges in the sample
On the CAT-Q, the present sample reported camouflag- (β = 0.33, p = .04), while camouflaging efforts (CAT-Q
ing to a degree very similar to that reported for the meas- total scores) approached significance as a predictor
ure’s autistic validation sample, with 55% of the sample (β = 0.24, p = .06). Neither clinician-rated autistic traits
reporting camouflaging above the autistic sample’s mean. (ADOS-2 scores) nor self-reported autistic traits per the
816 Autism 24(4)

AQ significantly predicted any of the mental health meas- Table 5. Correlations between CAT-Q total scores and
ures. Interactions between self-reported autistic traits and mental health measures for high- and low-camouflaging
camouflaging efforts were also not statistically significant subgroups.
in any model. High- Low-
camouflaging camouflaging
(n = 32) (n = 26)
Post hoc analysis of a high-camouflaging
subgroup r p r p

Investigation of the scatterplots created before the regres- DASS-21 total score 0.36 0.04 0.25 0.22
sion analyses (see Figure S1) suggested that camouflaging SBQ-R total score 0.37 0.04 0.05 0.81
might relate to measures of distress and functional chal- WHODAS 2.0 total score 0.46 < 0.01 0.08 0.69
lenges more strongly for participants who reported high lev- CAT-Q: Camouflaging Autistic Traits Questionnaire; DASS-21: Depression
els of camouflaging versus those who reported low levels. Anxiety Stress Scales 21; SBQ-R: Suicidal Behavior Questionnaire–Revised;
Therefore, we used the reported mean CAT-Q total score of WHODAS 2.0: World Health Organization Disability Assessment Schedule,
autistic adults in the validation sample (4.79) to split our Second Edition.
Members of high-camouflaging subgroup have CAT-Q total scores > 4.79
sample into two groups: high-camouflaging (n = 32) and (the mean score of autistic adults in the validation sample; Hull et al.,
low-camouflaging (n = 26). Nine participants of the high- 2019); members of the low-camouflaging subgroup have scores ⩽ 4.79.
camouflaging subgroup (28%) had received autism diagno-
ses prior to their participation; 12 members of this group for participants who were autistic per other diagnostic
(38%) scored above autism cutoffs on the ADOS-2. We measures, we classified participants for autism “caseness”
completed Doornik-Hansen omnibus tests to confirm bivar- by applying each tool’s published cut score or interpretive
iate normality in the relationships between CAT-Q total guideline: SRS-2 total scores in the Moderate range (total
scores and each mental health measure (DASS-21, SBQ-R, raw score > 84) and AQ scores ⩾ 32 were considered indic-
WHODAS 2.0; Doornik & Hansen, 2008). We then calcu- ative of autism. A review of Table S3 shows there is no
lated Pearson correlations between CAT-Q total scores and clear pattern between diagnostic stringency and average
scores on the mental health measures, first in the high-cam- scores on the mental health measures. Interestingly, approx-
ouflaging subgroup, and then in the low-camouflaging sub- imately 71% of the sample—including all participants pre-
group for comparison. In the high-camouflaging subgroup, viously diagnosed with autism—had scores in the range of
CAT-Q total scores were significantly associated with each severity consistent with autism on at least one measure, yet
measure of mental health. In the low-camouflaging sub- no single measure or combination of measures classified all
group, CAT-Q total scores were not associated with any of previously diagnosed participants as autistic.
the three measures of mental health (ps > 0.2; Table 5). Measures of mental health and camouflaging were more
strongly associated in a subgroup of women who did not
Post hoc analysis of subgroups based on have a diagnosis of autism (n = 40). In this group, CAT-Q
diagnostic classification total scores were significantly associated with psychologi-
cal distress (DASS-21 total scores) and functional chal-
We did not include diagnostic status as a binary independ- lenges (WHODAS 2.0 total scores); see Table S4.
ent variable in the regression models because we included
ADOS-2 scores, and diagnostic status is conflated with
Discussion
these scores. However, we wondered whether participant
characteristics varied significantly by ADOS-2 diagnostic A majority of the sample of women with autistic traits
classification. We completed t tests to compare the partici- reported significant psychological distress (depression, anx-
pants who scored above ADOS-2 cut score (revised iety, and stress), suicidality, and reduced daily functioning.
Module 4 algorithm overall score ⩾ 8) with those who did Self-reported camouflaging efforts (CAT-Q total scores) sig-
not. On average, those who met ADOS-2 criteria for autism nificantly predicted psychological distress while measures
spectrum disorder were older by about 5 years. Otherwise, of autistic traits (ADOS-2, AQ, and SRS-2) did not. CAT-Q
there were no significant differences on participant charac- scores also showed a trend toward predicting functional
teristics, including on the dependent variables in the challenges, and self-reported autistic traits (SRS-2 total
regression analyses (i.e., DASS-21, SBQ-R, WHODAS scores) significantly predicted such challenges. For a sub-
2.0). Group mean differences for non-normal variables sample of participants with high CAT-Q camouflaging
(age, FSIQ-2, ADOS-2 total scores, SRS-2 total raw scores) scores, CAT-Q scores were significantly associated with sui-
were confirmed with Whitney–Mann U tests. Full results cidality, in addition to psychological distress and reduced
are reported in the supplementary materials (Table S2). functioning. These strong associations between camouflag-
To explore whether participants’ scores on mental health ing and mental health in the high-camouflaging subgroup,
measures (the dependent variables) varied systematically in combination with the nonsignificant interactions between
Beck et al. 817

camouflaging efforts and self-reported autistic traits in the Camouflaging and mental health
full-sample regression models, suggest that camouflaging’s
relationship with mental health is linked to the degree of Our findings suggest that camouflaging efforts are perva-
camouflaging effort and not to the severity of autistic traits sive among women who perceive themselves as socially
being camouflaged. atypical, regardless of whether they have been or could be
Our findings echo results from adults diagnosed with diagnosed as autistic. Our findings also suggest that assess-
autism as reported in Cassidy, Bradley, Shaw, and Baron- ing camouflaging efforts is clinically relevant because,
Cohen (2018) showing that self-reported suicidality (SBQ- although camouflaging behaviors may often be adaptive
R) was not associated with AQ scores. Furthermore, we and facilitate social inclusion, they are significantly asso-
partly replicated their finding that self-reported camou- ciated with psychological distress, even after accounting
flaging (which they assessed using four brief questions) for severity of autistic traits. The direction of this associa-
was significantly associated with suicidality (assessed tion is unclear. Perhaps camouflaging efforts come about
using the SBQ-R) in women and men who reported a diag- in response to feeling social isolation, low self-concept,
nosis of autism spectrum disorder. In our intentionally and/or low self-acceptance, or perhaps vice versa. Notably,
broader sample that included women with mild autistic however, many women with autistic traits describe camou-
traits, we found a significant association between camou- flaging efforts as having many negative consequences. For
flaging and suicidality only in participants who reported example, a qualitative investigation of camouflaging in a
engaging in camouflaging behaviors to a high degree. sample of women and men (N = 92; 60% women) reported
themes regarding the long-term consequences of camou-
flaging that were all negative: “I fall to pieces,” “People
Evaluating mental health concerns in women have a stereotyped view [of me],” and “I’m not my true
with autistic traits self” (Hull et al., 2017).
Our finding that camouflaging efforts are more associ-
The high prevalence of mental health concerns in the pre- ated with mental health concerns in undiagnosed women
sent sample raises the question of whether these concerns supports the idea that undiagnosed women may be more
are common among women with autistic traits, or autistic likely to feel different, misunderstood, or lonely, possibly
traits are common among women with mental health con- because they are more likely to attribute social challenges
cerns, or both. What is clear from our results is that many to personal traits rather than lack of acceptance of one’s
women with symptoms of the broader autism phenotype community or other factors (Bargiela et al., 2016; Hickey
are in significant distress. This finding emphasizes the et al., 2018; Kanfiszer et al., 2017; Leedham et al., 2020).
importance of thorough, multimethod assessment of men- This finding is tentative due to the small sample size of
tal health concerns in women with autistic traits, especially formerly diagnosed autistic women, who did not show a
as such concerns may mask or exaggerate autistic traits. similar correlation between camouflaging and psychologi-
For example, Gotham et al. (2018) found that autistic cal distress or functional challenges (Table S4), but who
adults showed neural activity in response to emotional might also have mental health concerns associated with
stimuli that seemed different from neurotypical peers, yet camouflaging.
similar to depressed neurotypical peers. Their finding sug- Given growing evidence that camouflaging is generally
gests that traits often attributed to autism (e.g., ruminative associated with significant mental health concerns, clini-
thoughts; Gotham et al., 2014) may be better explained by cians who administer interventions that promote camou-
depression or other mental health concerns rather than by flaging efforts should consider whether these efforts are
autism traits per se. Clinicians must carefully consider manageable and in line with the patient’s values (Camm-
how mental health concerns affect presentation of autism Crosbie et al., 2019; Crane et al., 2019). Clinicians who
symptoms, and vice versa. One important strategy for encourage camouflaging efforts might also consider
doing this is to listen carefully and thoroughly to the indi- arranging wraparound services (e.g., support groups to dis-
vidual client (Crane et al., 2019). cuss possibly less desirable effects, social clubs that sup-
To reiterate, the fact that an autistic woman is verbally port autistic women in “being themselves”) to prevent
fluent, has IQ in the average range or above, and seems to possible mental health issues in the event camouflaging
function well does not indicate that she is functioning well itself is a causal contributor to psychological distress.
emotionally. Indeed, she may be in a significant amount of
distress. It is important to properly assess all concerns
Limitations and strengths
because, for many, a proper diagnosis leads to greater self-
acceptance, a sense of belonging in the autism community, Our study has notable limitations. First, the all-female
and access to appropriate care, including support in man- sample was relatively homogeneous in terms of age, cog-
aging suicidality, anxiety, and the many challenges inher- nitive ability, ethnicity, and socioeconomic background.
ent in navigating the neurotypical social space (Jones Therefore, the findings may not generalize to children,
et al., 2014). individuals with intellectual challenges, people of color, or
818 Autism 24(4)

individuals who experience poverty or other significant share regardless of diagnostic status. Along with increas-
psychosocial stressors unrepresented in this study. Second, ing public awareness of autism, there is increasing aware-
the validity of findings may be impacted by mono-opera- ness that loneliness is a widespread public health concern
tion bias as several of the constructs (e.g., camouflaging) (Beutel et al., 2017; Holt-Lunstad et al., 2017; Leigh-Hunt
were quantified using a single measure. Third, since all et al., 2017). As society adapts to address loneliness in the
measures were self-reported except for the ADOS-2 and neurotypical majority, there is opportunity to create inclu-
WASI-II, common-method variance possibly inflated the sive communities that can meet the social needs of all peo-
associations between most predictors and the dependent ple, including those with autistic traits. Future research can
variables. Fourth, although many analyses were suffi- explore how best to socially include women with autistic
ciently powered, the study has low (< 0.8) power to detect traits and whether increased social connection reduces
some effects of interest, especially in the small high-cam- psychological distress and increases functional capacity.
ouflaging subgroup. And fifth, our cross-sectional study
design does not allow us to investigate questions regarding Acknowledgements
causality or changing trajectories over time. It is important We are deeply indebted to the study participants. We thank the
that future studies investigate these constructs longitudi- research assistants who contributed to the work reported in this
nally across the lifespan. article, including Ava Dixon, Brinnley Lemmon, Eli Haun,
Our study also has notable strengths. The unique and Claire A. Chelladurai, Greyson Schuenman, Madyline Spencer,
inclusive sample of women with autistic traits allowed us to Marissa Farmer, Trevor DeMordaunt, Laurel Peacock, Katherine
investigate constructs dimensionally and trans-diagnosti- E. Christensen, Rachel Nuttall, Lacy Ekins, and Sarah Kamhout.
cally. Arguably, unlike many “clean” research samples, the
present sample represents a broad range of women with Declaration of conflicting interests
autistic traits who seek clinical services: some participants The author(s) declared no potential conflicts of interest with
were diagnosed autistic women, some are probably autistic respect to the research, authorship, and/or publication of this
women who are undiagnosed due to successful camouflag- article.
ing efforts, and some are neurotypical women with autistic
traits. We encourage other researchers to heed the calls for Funding
research that explores distress trans-diagnostically as this The author(s) disclosed receipt of the following financial support
approach will allow samples to include undiagnosed for the research, authorship, and/or publication of this article:
(potentially camouflaged) autistic women (Halladay et al., This work was supported by a Mentoring Environment Grant
2015; Livingston & Happé, 2017). Another strength of our from Brigham Young University to Drs. Lundwall and Cox, a
study is that it incorporated a measure of daily functional Graduate Research Fellowship from Brigham Young University
to Dr. Beck, and a graduate research grant from the Organization
challenges in addition to measures of psychological dis-
for Autism Research to Dr. Beck.
tress. To our knowledge, ours is the first study to report on
the relationship between camouflaging of autistic traits and
ORCID iDs
functional challenges.
Jonathan S Beck https://orcid.org/0000-0002-9815-0597
Mikle South https://orcid.org/0000-0003-0152-1257
Future directions
Supplemental material
The predictor variables included in our study accounted
for about a quarter or less of the variance in the clinically Supplemental material for this article is available online.
important dependent variables. Potential predictive factors
not included in the current study design are numerous, References
including the perceived ratio of successful to unsuccessful Bargiela, S., Steward, R., & Mandy, W. (2016). The experiences
camouflaging efforts. Our findings regarding camouflag- of late-diagnosed women with autism spectrum conditions:
ing and social difficulties (SRS-2) harmonize with Cassidy, An investigation of the female autism phenotype. Journal of
Bradley, Shaw, and Baron-Cohen’s (2018) finding that Autism and Developmental Disorders, 46(10), 3281–3294.
camouflaging and unmet support needs were risk factors https://doi.org/10.1007/s10803-016-2872-8
for severe psychological distress, as well as with a recent Baron-Cohen, S., Wheelwright, S., Skinner, R., Martin, J., &
Clubley, E. (2001). The Autism-Spectrum Quotient (AQ):
finding that loneliness is among the strongest predictors of
Evidence from Asperger syndrome/high-functioning
depression in autistic adults (Han et al., 2019). Together, autism, males and females, scientists and mathematicians.
these results suggest that self-perceived social atypicality Journal of Autism and Developmental Disorders, 31(1),
and social isolation likely are key predictors of mental 5–17. https://doi.org/10.1023/A:1005653411471
health challenges in individuals with autistic traits. Begeer, S., Mandell, D., Wijnker-Holmes, B., Venderbosch, S.,
Somewhat ironically, feeling chronically different from Rem, D., Stekelenburg, F., & Koot, H. M. (2013). Sex dif-
their peers may be something people with autistic traits all ferences in the timing of identification among children and
Beck et al. 819

adults with autism spectrum disorders. Journal of Autism depressive symptoms in a verbal adolescent-adult ASD sam-
and Developmental Disorders, 43, 1151–1156. https://doi. ple. Autism Research: Official Journal of the International
org/10.1007/s10803-012-1656-z Society for Autism Research, 7(3), 381–391. https://doi.
Beutel, M. E., Klein, E. M., Brähler, E., Reiner, I., Jünger, C., org/10.1002/aur.1377
Michal, M., . . . Tibubos, A. N. (2017). Loneliness in the Gotham, K., Risi, S., Pickles, A., & Lord, C. (2006). The Autism
general population: Prevalence, determinants and relations Diagnostic Observation Schedule: Revised algorithms
to mental health. BMC Psychiatry, 17(1), Article 97. for improved diagnostic validity. Journal of Autism and
Bolton, P. F., Pickles, A., Murphy, M., & Rutter, M. (1998). Developmental Disorders, 37(4), 613–627. https://doi.
Autism, affective and other psychiatric disorders: Patterns org/10.1007/s10803-006-0280-1
of familial aggregation. Psychological Medicine, 28(2), Gotham, K., Siegle, G. J., Han, G. T., Tomarken, A. J., Crist, R.
385–395. N., Simon, D. M., & Bodfish, J. W. (2018). Pupil response
Bruni, T. P. (2014). Test review: Social Responsiveness to social-emotional material is associated with rumination
Scale–Second Edition (SRS-2). Journal of Psychoeduca- and depressive symptoms in adults with autism spectrum
tional Assessment, 32(4), 365–369. https://doi.org/10.1177/ disorder. PLOS ONE, 13(8), Article e0200340. https://doi.
0734282913517525 org/10.1371/journal.pone.0200340
Cage, E., & Troxell-Whitman, Z. (2019). Understanding the rea- Halladay, A. K., Bishop, S., Constantino, J. N., Daniels, A. M.,
sons, contexts and costs of camouflaging for autistic adults. Koenig, K., Palmer, K., . . . Szatmari, P. (2015). Sex and gen-
Journal of Autism and Developmental Disorders, 49(5), der differences in autism spectrum disorder: Summarizing
1899–1911. evidence gaps and identifying emerging areas of priority.
Camm-Crosbie, L., Bradley, L., Shaw, R., Baron-Cohen, S., & Molecular Autism, 6(1), Article 36. https://doi.org/10.1186/
Cassidy, S. (2019). “People like me don’t get support”: s13229-015-0019-y
Autistic adults’ experiences of support and treatment for Han, G. T., Tomarken, A. J., & Gotham, K. O. (2019). Social
mental health difficulties, self-injury and suicidality. Autism, and nonsocial reward moderate the relation between autism
23, 1431–1441. https://doi.org/10.1177/1362361318816053 symptoms and loneliness in adults with ASD, depression,
Cassidy, S., Bradley, L., Bowen, E., Wigham, S., & Rodgers, J. and controls. Autism Research, 12(6), 884–896. https://doi.
(2018). Measurement properties of tools used to assess sui- org/10.1002/aur.2088
cidality in autistic and general population adults: A system- Hedegaard, H., Curtin, S. C., & Warner, M. (2018). Suicide mor-
atic review. Clinical Psychology Review, 62, 56–70. https:// tality in the United States, 1999–2017. NCHS Data Brief,
doi.org/10.1016/j.cpr.2018.05.002 330, 1–8.
Cassidy, S., Bradley, L., Shaw, R., & Baron-Cohen, S. (2018). Henry, J. D., & Crawford, J. R. (2005). The short-form version of
Risk markers for suicidality in autistic adults. Molecular the Depression Anxiety Stress Scales (DASS-21): Construct
Autism, 9(1), Article 42. https://doi.org/10.1186/s13229- validity and normative data in a large non-clinical sample.
018-0226-4 British Journal of Clinical Psychology, 44(2), 227–239.
Cassidy, S., Bradley, P., Robinson, J., Allison, C., McHugh, Hickey, A., Crabtree, J., & Stott, J. (2018). “Suddenly the first
M., & Baron-Cohen, S. (2014). Suicidal ideation and sui- fifty years of my life made sense”: Experiences of older
cide plans or attempts in adults with Asperger’s syndrome people with autism. Autism, 22(3), 357–367. https://doi.
attending a specialist diagnostic clinic: A clinical cohort org/10.1177/1362361316680914
study. Lancet Psychiatry, 1, 142–147. Hirvikoski, T., Boman, M., Chen, Q., D’Onofrio, B. M.,
Constantino, J. (2012). Social Responsiveness Scale, Second Mittendorfer-Rutz, E., Lichtenstein, P., Bölte, S., & Larsson,
Edition (SRS-2). Western Psychological Services. H. (2019). Individual risk and familial liability for suicide
Crane, L., Adams, F., Harper, G., Welch, J., & Pellicano, E. attempt and suicide in autism: A population-based study.
(2019). “Something needs to change”: Mental health expe- Psychological Medicine, 1-12. https://doi.org/10.1017/S003
riences of young autistic adults in England. Autism, 23(2), 3291719001405
477–493. https://doi.org/10.1177/1362361318757048 Hirvikoski, T., Mittendorfer-Rutz, E., Boman, M., Larsson, H.,
Dixon, W. J., & Yuen, K. K. (1974). Trimming and winsoriza- Lichtenstein, P., & Bölte, S. (2016). Premature mortality in
tion: A review. Statistische Hefte, 15(2–3), 157–170. https:// autism spectrum disorder. The British Journal of Psychiatry:
doi.org/10.1007/BF02922904 The Journal of Mental Science, 208(3), 232–238. https://
Doornik, J. A., & Hansen, H. (2008). An omnibus test for uni- doi.org/10.1192/bjp.bp.114.160192
variate and multivariate normality. Oxford Bulletin of Holt-Lunstad, J., Robles, T. F., & Sbarra, D. A. (2017).
Economics and Statistics, 70(s1), 927–939. Advancing social connection as a public health priority in
Engström, I., Ekström, L., & Emilsson, B. (2003). Psychosocial the United States. American Psychologist, 72(6), 517–530.
functioning in a group of Swedish adults with Asperger syn- Howlin, P. (2000). Outcome in adult life for more able individu-
drome or high-functioning autism. Autism, 7(1), 99–110. als with autism or Asperger syndrome. Autism, 4(1), 63–83.
https://doi.org/10.1177/1362361303007001008 Howlin, P., & Magiati, I. (2017). Autism spectrum dis-
Giarelli, E., Wiggins, L. D., Rice, C. E., Levy, S. E., Kirby, R. order: Outcomes in adulthood. Current Opinion in
S., Pinto-Martin, J., & Mandell, D. (2010). Sex differences Psychiatry, 30(2), 69–76. https://doi.org/10.1097/YCO.000
in the evaluation and diagnosis of autism spectrum disor- 0000000000308
ders among children. Disability and Health Journal, 3(2), Hull, L., Lai, M. C., Baron-Cohen, S., Allison, C., Smith, P.,
107–116. Petrides, K. V., & Mandy, W. (2020). Gender differences
Gotham, K., Bishop, S. L., Brunwasser, S., & Lord, C. (2014). in self-reported camouflaging in autistic and non-autistic
Rumination and perceived impairment associated with adults. Autism, 24, 352–363.
820 Autism 24(4)

Hull, L., Mandy, W., Lai, M.-C., Baron-Cohen, S., Allison, C., Psychiatry, 6(10), 819–829. https://doi.org/10.1016/S2215-
Smith, P., & Petrides, K. V. (2019). Development and vali- 0366(19)30289-5
dation of the Camouflaging Autistic Traits Questionnaire Lai, M.-C., Lombardo, M. V., Ruigrok, A. N., Chakrabarti, B.,
(CAT-Q). Journal of Autism and Developmental Disorders, Auyeung, B., Szatmari, P., & Baron-Cohen, S. (2017).
49(3), 819–833. https://doi.org/10.1007/s10803-018-3792-6 Quantifying and exploring camouflaging in men and
Hull, L., Petrides, K. V., Allison, C., Smith, P., Baron-Cohen, women with autism. Autism, 21, 690–702. https://doi.
S., Lai, M.-C., & Mandy, W. (2017). “Putting on my best org/10.1177/1362361316671012
normal”: Social camouflaging in adults with autism spec- Leedham, A., Thompson, A., Smith, R., & Freeth, M. (2020).
trum conditions. Journal of Autism and Developmental “I was exhausted trying to figure it out”: The experi-
Disorders, 47(8), 2519–2534. https://doi.org/10.1007/ ences of females receiving an autism diagnosis in mid-
s10803-017-3166-5 dle to late adulthood. Autism, 24, 135–146. https://doi.
Hurlbutt, K., & Chalmers, L. (2002). Adults with autism speak org/10.1177/1362361319853442
out: Perceptions of their life experiences. Focus on Autism Leigh-Hunt, N., Bagguley, D., Bash, K., Turner, V., Turnbull, S.,
and Other Developmental Disabilities, 17(2), 103–111. Valtorta, N., & Caan, W. (2017). An overview of systematic
https://doi.org/10.1177/10883576020170020501 reviews on the public health consequences of social isola-
Hurley, R. S. E., Losh, M., Parlier, M., Reznick, J. S., & Piven, tion and loneliness. Public Health, 152, 157–171.
J. (2007). The Broad Autism Phenotype Questionnaire. Leitner, Y. (2014). The co-occurrence of autism and attention
Journal of Autism and Developmental Disorders, 37(9), deficit hyperactivity disorder in children—What do we
1679–1690. https://doi.org/10.1007/s10803-006-0299-3 know? Frontiers in Human Neuroscience, 8, Article 268.
Hus, V., & Lord, C. (2014). The Autism Diagnostic Observation https://doi.org/10.3389/fnhum.2014.00268
Schedule, Module 4: Revised algorithm and standardized Lever, A. G., & Geurts, H. M. (2016). Psychiatric co-occurring
severity scores. Journal of Autism and Developmental symptoms and disorders in young, middle-aged, and older
Disorders, 44(8), 1996–2012. https://doi.org/10.1007/s10803 adults with autism spectrum disorder. Journal of Autism
-014-2080-3 and Developmental Disorders, 46(6), 1916–1930. https://
Ingersoll, B., & Hambrick, D. Z. (2011). The relationship between doi.org/10.1007/s10803-016-2722-8
the broader autism phenotype, child severity, and stress and Livingston, L. A., & Happé, F. (2017). Conceptualising compen-
depression in parents of children with autism spectrum dis- sation in neurodevelopmental disorders: Reflections from
orders. Research in Autism Spectrum Disorders, 5(1), 337– autism spectrum disorder. Neuroscience & Biobehavioral
344. https://doi.org/10.1016/j.rasd.2010.04.017 Reviews, 80, 729–742. https://doi.org/10.1016/j.neubi-
Jones, L., Goddard, L., Hill, E. L., Henry, L. A., & Crane, L. orev.2017.06.005
(2014). Experiences of receiving a diagnosis of autism spec- Livingston, L. A., Shah, P., & Happé, F. (2019). Compensatory
trum disorder: A survey of adults in the United Kingdom. strategies below the behavioural surface in autism: A quali-
Journal of Autism and Developmental Disorders, 44(12), tative study. Lancet Psychiatry, 6(9), 766–777.
3033–3044. https://doi.org/10.1007/s10803-014-2161-3 Lord, C., Rutter, M., DiLavore, P., Risi, S., Gotham, K., & Bishop,
Kanfiszer, L., Davies, F., & Collins, S. (2017). “I was just so S. (2012). Autism Diagnostic Observation Schedule–Second
different”: The experiences of women diagnosed with an Edition (ADOS-2). Western Psychological Corporation.
autism spectrum disorder in adulthood in relation to gender Lovibond, S. H., & Lovibond, P. F. (1995). Manual for the
and social relationships. Autism, 21(6), 661–669. https://doi. Depression Anxiety Stress Scales (2nd ed.). https://trove.
org/10.1177/1362361316687987 nla.gov.au/work/30421447
Katon, W. (2009). The impact of depression on workplace Mandell, D. S., Lawer, L. J., Branch, K., Brodkin, E. S., Healey,
functioning and disability costs. The American Journal of K., Witalec, R., . . . Gur, R. E. (2012). Prevalence and corre-
Managed Care, 15(11, Suppl.), S322–S327. lates of autism in a state psychiatric hospital. Autism, 16(6),
Kessler, R. C., Berglund, P., Demler, O., Jin, R., Merikangas, K. 557–567. https://doi.org/10.1177/1362361311412058
R., & Walters, E. E. (2005). Lifetime prevalence and age- Mandy, W., Chilvers, R., Chowdhury, U., Salter, G., Seigal, A.,
of-onset distributions of DSM-IV disorders in the national & Skuse, D. (2012). Sex differences in autism spectrum dis-
comorbidity survey replication. Archives of General Psy- order: Evidence from a large sample of children and ado-
chiatry, 62(6), 593–602. https://doi.org/10.1001/archpsyc lescents. Journal of Autism and Developmental Disorders,
.62.6.593 42(7), 1304–1313. https://doi.org/10.1007/s10803-011-
Kirby, A. V., Bakian, A. V., Zhang, Y., Bilder, D. A., Keeshin, 1356-0
B. R., & Coon, H. (2019). A 20-year study of suicide death Milner, V., McIntosh, H., Colvert, E., & Happé, F. (2019). A qual-
in a statewide autism population. Autism Research, 12(4), itative exploration of the female experience of autism spec-
658–666. https://doi.org/10.1002/aur.2076 trum disorder (ASD). Journal of Autism and Developmental
Kirby, A. V., Baranek, G. T., & Fox, L. (2016). Longitudinal Disorders, 49, 2389–2402. https://doi.org/10.1007/s10803-
predictors of outcomes for adults with autism spectrum dis- 019-03906-4
order: Systematic review. OTJR: Occupation, Participation Nylander, L., Axmon, A., Björne, P., Ahlström, G., & Gillberg,
and Health, 36(2), 55–64. C. (2018). Older adults with autism spectrum disorders in
Lai, M.-C., Kassee, C., Besney, R., Bonato, S., Hull, L., Mandy, Sweden: A register study of diagnoses, psychiatric care
W., Szatmari, P., & Ameis, S. H. (2019). Prevalence of utilization and psychotropic medication of 601 individu-
co-occurring mental health diagnoses in the autism popu- als. Journal of Autism and Developmental Disorders, 48(9),
lation: A systematic review and meta-analysis. Lancet 3076–3085. https://doi.org/10.1007/s10803-018-3567-0
Beck et al. 821

O’Brien, R. M. (2007). A caution regarding rules of thumb for in adults with autism spectrum disorder. Journal of Autism
variance inflation factors. Quality & Quantity, 41(5), 673– and Developmental Disorders, 49, 2597–2604. https://doi.
690. https://doi.org/10.1007/s11135-006-9018-6 org/10.1007/s10803-019-03998-y
Osman, A., Bagge, C. L., Gutierrez, P. M., Konick, L. C., Kopper, Solomon, M., Miller, M., Taylor, S. L., Hinshaw, S. P., &
B. A., & Barrios, F. X. (2001). The Suicidal Behaviors Carter, C. S. (2012). Autism symptoms and internaliz-
Questionnaire-Revised (SBQ-R): Validation with clinical ing psychopathology in girls and boys with autism spec-
and nonclinical samples. Assessment, 8(4), 443–454. trum disorders. Journal of Autism and Developmental
Oswald, T. M., Winter-Messiers, M. A., Gibson, B., Schmidt, A. Disorders, 42(1), 48–59. https://doi.org/10.1007/s10803-
M., Herr, C. M., & Solomon, M. (2016). Sex differences in 011-1215-z
internalizing problems during adolescence in autism spectrum StataCorp. (2017). Stata statistical software (Version Release 15).
disorder. Journal of Autism and Developmental Disorders, Stewart, M. E., Barnard, L., Pearson, J., Hasan, R., & O’Brien, G.
46(2), 624–636. https://doi.org/10.1007/s10803-015-2608-1 (2006). Presentation of depression in autism and Asperger
Park, S. H., Demetriou, E. A., Pepper, K. L., Song, Y. J. C., syndrome: A review. Autism, 10(1), 103–116. https://doi.
Thomas, E. E., Hickie, I. B., . . . Guastella, A. J. (2019). org/10.1177/1362361306062013
Validation of the 36-item and 12-item self-report World Taylor, J. L., Henninger, N. A., & Mailick, M. R. (2015). Lon-
Health Organization Disability Assessment Schedule II gitudinal patterns of employment and postsecond-
(WHODAS-II) in individuals with autism spectrum disor- ary education for adults with autism and average-range
der. Autism Research: Official Journal of the International IQ. Autism, 19(7), 785–793. https://doi.org/10.1177/1362
Society for Autism Research, 12, 1101–1111. https://doi. 361315585643
org/10.1002/aur.2115 Thode, H. C. (2002). Testing for normality. CRC Press.
Pine, D. S., Guyer, A. E., Goldwin, M., Towbin, K. A., & Tint, A., Weiss, J. A., & Lunsky, Y. (2017). Identifying the clini-
Leibenluft, E. (2008). Autism spectrum disorder scale scores cal needs and patterns of health service use of adolescent
in pediatric mood and anxiety disorders. Journal of the girls and women with autism spectrum disorder. Autism
American Academy of Child & Adolescent Psychiatry, 47(6), Research, 10, 1558–1566.
652–661. https://doi.org/10.1097/CHI.0b013e31816bffa5 Üstün, T. B. (2010). Measuring health and disability: Manual
Piven, J., & Palmer, P. (1999). Psychiatric disorder and the broad for WHO Disability Assessment Schedule (WHODAS 2.0).
autism phenotype: Evidence from a family study of multiple- World Health Organization.
incidence autism families. American Journal of Psychiatry, Wallace, G. L., Kenworthy, L., Pugliese, C. E., Popal, H. S.,
156(4), 557–563. https://doi.org/10.1176/ajp.156.4.557 White, E. I., Brodsky, E., & Martin, A. (2016). Real-world
Piven, J., Palmer, P., Jacobi, D., Childress, D., & Arndt, S. executive functions in adults with autism spectrum disor-
(1997). Broader autism phenotype: Evidence from a fam- der: Profiles of impairment and associations with adaptive
ily history study of multiple-incidence autism families. functioning and co-morbid anxiety and depression. Journal
American Journal of Psychiatry, 154(2), 185–190. https:// of Autism and Developmental Disorders, 46(3), 1071–1083.
doi.org/10.1176/ajp.154.2.185 https://doi.org/10.1007/s10803-015-2655-7
Portway, S. M., & Johnson, B. (2005). Do you know I have Warner, R. M. (2013). Applied statistics: From bivariate through
Asperger’s syndrome? Risks of a non-obvious dis- multivariate techniques. Los Angeles, CA: SAGE.
ability. Health, Risk & Society, 7(1), 73–83. https://doi. Webster, A. A., & Garvis, S. (2017). The importance of critical
org/10.1080/09500830500042086 life moments: An explorative study of successful women
Sasson, N. J., Lam, K. S. L., Childress, D., Parlier, M., Daniels, with autism spectrum disorder. Autism, 21(6), 670–677.
J. L., & Piven, J. (2013). The Broad Autism Phenotype https://doi.org/10.1177/1362361316677719
Questionnaire: Prevalence and diagnostic classification. Wechsler, D., & Hsiao-Pin, C. (2011). WASI-II: Wechsler
Autism Research: Official Journal of the International Abbreviated Scale of Intelligence (2nd ed.). Pearson.
Society for Autism Research, 6(2), 134–143. https://doi. White, S. W., Oswald, D., Ollendick, T., & Scahill, L. (2009).
org/10.1002/aur.1272 Anxiety in children and adolescents with autism spectrum
Schuck, R. K., Flores, R. E., & Fung, L. K. (2019). Brief report: disorders. Clinical Psychology Review, 29(3), 216–229.
Sex/gender differences in symptomology and camouflaging https://doi.org/10.1016/j.cpr.2009.01.003