Behavioural and Cognitive

669087
review-article2016
AUT0010.1177/1362361316669087AutismHull et al.
Review Article
Autism
Behavioural and cognitive 2017, Vol. 21(6) 706–727

© The Author(s) 2016
Reprints and permissions:
sex/gender differences in autism sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/1362361316669087
https://doi.org/10.1177/1362361316669087
spectrum condition and typically journals.sagepub.com/home/aut
developing males and females
Laura Hull, William Mandy and KV Petrides
Abstract
Studies assessing sex/gender differences in autism spectrum conditions often fail to include typically developing control
groups. It is, therefore, unclear whether observed sex/gender differences reflect those found in the general population or
are particular to autism spectrum conditions. A systematic search identified articles comparing behavioural and cognitive
characteristics in males and females with and without an autism spectrum condition diagnosis. A total of 13 studies were
included in meta-analyses of sex/gender differences in core autism spectrum condition symptoms (social/communication
impairments and restricted/repetitive behaviours and interests) and intelligence quotient. A total of 20 studies were
included in a qualitative review of sex/gender differences in additional autism spectrum condition symptoms. For core
traits and intelligence quotient, sex/gender differences were comparable in autism spectrum conditions and typical
samples. Some additional autism spectrum condition symptoms displayed different patterns of sex/gender differences
in autism spectrum conditions and typically developing groups, including measures of executive function, empathising
and systemising traits, internalising and externalising problems and play behaviours. Individuals with autism spectrum
conditions display typical sex/gender differences in core autism spectrum condition traits, suggesting that diagnostic
criteria based on these symptoms should take into account typical sex/gender differences. However, awareness of
associated autism spectrum condition symptoms should include the possibility of different male and female phenotypes,
to ensure those who do not fit the ‘typical’ autism spectrum condition presentation are not missed.
Keywords
autism spectrum conditions, diagnosis, gender differences, sex differences
Introduction
Autism spectrum conditions (ASCs) are more commonly et al., 2012; Thompson et al., 2003; Werling and
diagnosed in males than in females across age groups Geschwind, 2013). If females with ASC tend to demon-
(Fombonne, 2009; Russell et al., 2011).1 Reliable genetic strate different symptom patterns to the majority of ASC
and/or physiological markers of ASC have not yet been males, they may be at greater risk of being missed by clini-
identified; therefore, diagnostic criteria rely on behav- cal services and support options than males (Dworzynski
ioural descriptions of the disorder. These criteria have et al., 2012). It is, therefore, important to assess whether
been developed based on the predominantly male popula- there is a need for a broader conceptualisation of ASC to
tions previously diagnosed or identified as having ASC
(Kirkovski et al., 2013; Kopp and Gillberg, 2011; Mattila
et al., 2011). However, researchers are increasingly University College London, UK
focused on the experiences and characteristics of females
Corresponding author:
with autism to determine whether males and females with
Laura Hull, Research Department of Clinical, Educational and Health
ASC display similar behavioural and cognitive profiles Psychology, University College London, 26 Bedford Way, London
(Dworzynski et al., 2012; Gould and Ashton-Smith, 2011; WC1H 0AP, UK.
Kopp and Gillberg, 1992; Lehnhardt et al., 2015; Mandy Email: [email protected]
Hull et al. 707
include typically female patterns of this condition. If so, symptoms experienced by males and females may partly
this could have implications for the diagnostic criteria of account for this variation in diagnosis, as the female phe-
ASC. notype may not be viewed as ‘typical’ ASC symptoms and
There have been several reviews of the literature on so may not immediately point towards an ASC diagnosis.
sex/gender differences in the ASC core symptoms of Reviews have also emphasised that sex/gender differ-
social/communication impairments and restricted/repeti- ences in ASC are influenced by individual differences.
tive behaviours and interests (RRBIs). (Note: following Females with low intelligence quotient (IQ) are more
Lai et al. (2015), we use the term ‘sex/gender’ to reflect the likely to receive a diagnosis than females with high IQ
awareness that the effects of biological ‘sex’ and socially (Rivet and Matson, 2011; Van Wijngaarden-Cremers et al.,
constructed ‘gender’ cannot be easily separated and that 2014), suggesting that there are additional factors interact-
most individuals’ identities are informed by both sex and ing with sex/gender to produce differences in diagnostic
gender.) These generally conclude that females with ASC rates. Individuals’ ages were also found to influence sex/
may display a different phenotype, or different patterns of gender differences in core ASC symptoms; for instance,
ASC characteristics, than males with ASC (Kirkovski Van Wijngaarden-Cremers et al. (2014) found that sex dif-
et al., 2013; Van Wijngaarden-Cremers et al., 2014). While ferences in RRBIs only occur from the age of 6 years.
specific sex/gender differences in the severity of social and There are also likely to be interacting influences from both
communication impairments have not been conclusively social and biological factors, such as genetic influences
demonstrated (Koenig and Tsatsanis, 2005; Lai et al., and social/cultural environment, which will contribute to
2011, 2012, 2015; Van Wijngaarden-Cremers et al., 2014), different developmental outcomes for males and females
some have found that girls and women with ASC, on aver- with ASC (Kreiser and White, 2014; Lai et al., 2015). It is,
age, display fewer RRBIs (Koenig and Tsatsanis, 2005; therefore, concluded that future research into sex/gender
Kreiser and White, 2014; Rivet and Matson, 2011). differences in ASC needs to take into account IQ, age and
However, it has been argued that RRBI diagnostic criteria other characteristics in order to fully understand how
fail to reflect the true range of areas under which RRBIs males and females with ASC develop.
can fall (Mandy et al., 2012). It is possible that many Based on the above, males and females with ASC
females with ASC experience very extreme interests or appear to demonstrate somewhat different characteristics
behavioural tendencies but in areas outside the ‘typical’ and have different clinical and diagnostic experiences.
ASC interests of systems and machines, therefore exclud- This would suggest that ASC diagnostic criteria and
ing them from meeting diagnostic criteria for RRBIs in thresholds should vary for males and females, to ensure
ASC. that all individuals are able to access the services and sup-
Reviews have also addressed sex/gender differences port they require. However, the precise ways in which
in additional symptoms associated with ASC, such as diagnostic criteria might be adapted depend on exactly
internalising and externalising problems and the co-diag- how and why males and females with ASC differ. One
noses that may result from these. Males with ASC and issue with previous research into sex/gender differences in
mulheres
costumam ter typically developing (TD) males with high autism traits ASC is that typical sex/gender differences have rarely
mais problemas
com a
are more likely to experience externalising problems been taken into account. This means that we cannot be cer-
internalização de such as behavioural problems and hyperactivity, while tain whether males and females with ASC differ in the
comportamentos
levando à females are more likely to experience internalising prob- same ways that TD males and females differ or whether
ansiedade e
depressão lems such as depression and anxiety hyperactivity having ASC has a differential impact on males and females,
(Koenig and Tsatsanis, 2005; Kreiser and White, 2014; and it is this that produces the sex/gender differences
Rivet and Matson, 2011). This suggests that the pattern of described above.
behaviours associated with ASC symptoms varies If the first prediction is borne out, then the performance
between males and females, which may require adjust- of ASC males and females on diagnostic criteria should
ment of current diagnostic criteria. also be compared to that of TD males and females, respec-
In addition to sex/gender differences in ASC symp- tively. Sex/gender differences (or lack thereof) in TD pop-
toms, differences in the diagnostic experiences of males ulations have been established for a wide range of
and females with ASC have also been observed. Females behaviours related to ASC; therefore, it stands to reason
with similar levels of ASC symptoms are less likely to be that ASC males should be compared to TD males and ASC
diagnosed with ASC than males (Dworzynski et al., 2012), females to TD females when assessing strengths and
and it is suggested that females are more likely to be mis- impairments.
diagnosed with other conditions, especially internalising However, if ASC does produce different outcomes for
and eating disorders (Kopp and Gillberg, 1992; Mandy males and females beyond those attributed to typical sex/
and Tchanturia, 2015). Females who do receive an ASC gender differences, adjustments to diagnostic criteria are
diagnosis do so at a later age than males on average less straightforward. One outcome might be the develop-
(Kirkovski et al., 2013). A difference in the ASC ment of separate diagnostic criteria for males and females,
708 Autism 21(6)
reflecting differential presentations of ASC in each sex/ were contacted to request data on control groups for inclu-
gender in at least some areas. It has also been suggested sion in the analysis, but none were able to provide com-
that females with ASC may compensate for or mask their plete data sets. Due to the limited number of eligible
ASC-related behaviours to a greater extent than males with studies, meta-analysis of ASC and typical sex/gender dif-
ASC, resulting in underestimations of the true extent of ferences was only possible for 6 studies measuring social/
ASC and its symptomatology in females (Dworzynski communication impairments, 5 studies measuring RRBIs
et al., 2012; Lai et al., 2012). Including these behaviours in and 13 studies measuring IQ.
a female phenotype of ASC would increase identification
of females and enable them to access the services and sup-
port they need. Statistical analysis
Thus, it is important to compare sex/gender differences Random-effects meta-analyses were performed using the
in the ASC population with those in TD groups, in order to ‘metafor’ package in R (R Core Team, 2013; Viechtbauer,
establish whether ASC interacts with an individual’s sex/ 2010) for measures of the core ASC symptoms and IQ.
gender to produce different outcomes or whether typical Using a random-effects model accounts for variance
sex/gender differences also exist within the ASC popula- between studies caused by sampling error and other arte-
tion. This then has implications for adjustments to diag- facts (Hunter and Schmidt, 2004). Mean sex/gender differ-
nostic criteria and for a broader conceptualisation of ASC ences in social/communication impairments (see Table 3),
in males and females. RRBIs (see Table 4) and IQ (see Table 5) were calculated
This research therefore aims to address the following for ASC and TD groups and then standardised mean differ-
questions: ences (SMDs) between these differences were calculated,
to take into account the variety of test instruments used.
What are the sex/gender differences in ASC core and Social and communication impairments were analysed
associated symptoms (if any), for ASC and TD groups? separately due to some studies testing these separately or
Do these sex/gender differences vary between ASC and only testing one of these but are presented and discussed
TD groups? In other words, is there an interaction together, to reflect the fact that these autistic symptoms are
between sex/gender and ASC diagnosis? treated as a unitary domain in Diagnostic and Statistical
Manual of Mental Disorders (5th ed.; DSM-5). Where
tests for heterogeneity were significant, a mixed-effects
Methods model was used to test for the effect of the moderator
Literature review ‘Age’ (age of participants). ASC groups were entered into
the analysis first; therefore, positive effect sizes would
A search of the PsycINFO, PubMed and Web of Science mean greater sex/gender differences in ASC groups than
directories in September 2015 for the terms ‘autism + sex typical groups, and negative effect sizes would mean
differences’ and ‘autism + gender differences’ produced smaller sex/gender differences in ASC groups. Where
3290 initial results. Figure 1 describes the logic used to multiple measures of the same symptom were used within
select studies for inclusion. Eligibility criteria were peer- one study, the measure most similar to those used in other
reviewed articles published in English and comparing studies was selected for inclusion in this analysis. R script
males and females with and without an ASC diagnosis, for all analyses is available on request from L.H.
which matched ASC and TD groups for IQ and age.
Bibliographies of relevant articles, including those of
seven recent review and/or meta-analysis articles, were Results
manually searched to find additional articles which may
have been missed in the initial search (n = 37). Studies
Meta-analysis
were excluded (n = 3307) if they were duplicates; if they Figure 2 presents the funnel plots for each of the four
only measured biological sex/gender differences; and if meta-analyses conducted. Due to the limited number of
they did not include groups of males and females with and studies, it is difficult to draw conclusions about publica-
without an ASC diagnosis, matched on age and IQ. A total tion bias. However, three of the four plots show some
of 20 original studies were selected for inclusion in the asymmetry, with a positive skew, suggesting that there
review of variation in sex/gender differences between may have been some publication bias in favour of studies
ASC and TD groups. See Tables 1 and 2 for information reporting statistically significant sex/gender differences in
about all 20 studies, including summaries of their findings ASC populations. Despite this, Hunter and Schmidt (2004)
and characteristics of the samples used. Where multiple note that studies of sex/gender difference may be less sus-
comparison groups were included, the group most similar ceptible to availability bias (the suggestion that studies
to an unrelated, general population sample was selected with significant findings and large effect sizes are more
for inclusion in this review. Several additional authors likely to be published and, therefore, more available for
Hull et al. 709
Figure 1. PRISMA flow diagram of study identification and selection. Flow diagram showing identification and selection of studies
for inclusion in the review and meta-analysis.
ASC: autism spectrum condition; TD: typically developing.
710
Table 1. Sample characteristics of studies included in review.

Article Authors (date) ASC symptom(s) assessed ASC group diagnoses at ASC diagnostic criteria How diagnosis confirmed ASC group TD group
the time of study used
Males Females Mean age Males Females Mean age
(n) (n) (years) (n) (n) (years)
1 Baron-Cohen et al. (2014) Empathising traits AS (62%) Not reported Not reported 357 454 34.7 1344 2562 34.4
Systemising traits ASD (29%)
Autism traits HFA (5%)
PDD (2%)
Autism (1%)
2 Baron-Cohen et al. (2003) Empathising traits AS/HFA (proportions not DSM-IV criteria for Not reported 33 14 38.1 114 164 30.9
Systemising traits reported) autism/AS
3 Baron-Cohen and Friendships AS/HFA (proportions not DSM-IV criteria for Not reported 51 17 34.4 27 49 40.5
Wheelwright (2005) reported) autism/AS
4 Baron-Cohen et al. (2001) Autism traits AS/HFA (proportions not DSM-IV criteria for Not reported 45 13 31.6 76 98 37
reported) autism/AS
5 Bölte et al. (2011) IQ Autism (68%) ICD-10 criteria for Clinician assessment and 35 21 14.2 23 35 14.6
Executive functions AS (20%) autism/AS/PDD-NOS ADI-R/ADOS
PDD-NOS (13%)
6 Dean et al. (2014) Friendships ASD (100%) Not reported Clinician assessment and ADOS 25 25 7.5 25 25 7.8
7 Goddard et al. (2014) IQ ASD (100%) DSM-IV-TR criteria Clinician assessment and SCQ 12 12 12.9 12 12 12.6
Memory for ASD
8 Harrop et al. (2016) IQ ASD (100%) Not reported ADI-R/ADOS 14 14 3.8 14 12 2.0
Play behaviours
9 Head et al. (2014) Friendships ASD (not including LFA Not reported Not reported 25 25 13.7 26 25 12
or PDD-NOS; 100%)
10 Holt et al. (2014) IQ AS/HFA (proportions not Not reported ADI-R/ADOS 33 16 14.6 20 20 15.1
Theory of mind reported)
11 Kirkovski et al. (2016) IQ AS (85%) DSM-IV-TR criteria Viewing of clinician diagnostic 13 14 30.7 11 12 30.7
Social impairments HFA (15%) for ASD report
RRBIs
Empathising traits
Autism traits
12 Knickmeyer et al. (2008) Play behaviours Of those available (91% of ICD-10 or DSM-IV Not reported 46 20 10.2 31 24 5.2
total sample): criteria for ASC
AS (32%)
Autism (58%)
HFA (3%)
PDD-NOS (3%)
Atypical autism (2%)
13 Lai et al. (2012) IQ AD/AS (proportions not ICD-10 or DSM-IV ADI-R/ADOS/AAA 32 32 27.6 32 32 28.2
Executive functions reported) criteria for ASC
Theory of mind
Memory
(Continued)
Autism 21(6)
Hull et al.
Table 1. (Continued)
Article Authors (date) ASC symptom(s) assessed ASC group diagnoses at ASC diagnostic criteria How diagnosis confirmed ASC group TD group
the time of study used
Males Females Mean age Males Females Mean age
(n) (n) (years) (n) (n) (years)
14 Lemon et al. (2011) IQ HFA (70%) DSM-IV criteria for Clinician assessment 10 13 11 8 14 11.4
Executive functions AS (30%) HFA/AS
15 May et al. (2012) Social impairment AS (64%) DSM-IV-TR criteria for Viewing of clinician diagnostic 32 32 9.9 30 30 9.3
Communication impairment AD (36%) AS/AD report
RRBIs
IQ
Inattention/hyperactivity
16 Oswald et al. (2016) IQ ASD (100%) Not reported AQ/ASDS 18 14 14.8 18 14 14.9
Internalising problems
17 Park et al. (2012) Social impairment ASD (100%) DSM-IV-R criteria for Clinician assessment 91 20 8.3 26 25 8.6
Communication impairment PDDs
RRBIs
IQ
Externalising problems
18 Sedgewick et al. (2015) Social impairment Autism (83%) DSM-IV-TR or ICD-10 Clinician assessment and 10 13 13.9 10 13 13.8
IQ AS (17%) criteria for autism/AS statement of special educational
Friendships needs indicating autism
19 Solomon et al. (2012) Social impairment ASD including HFA, DSM-IV-TR criteria for ADOS/SDQ 20 20 12.2 17 19 12
Communication impairment AS and PDD-NOS AD/AS/PDD-NOS
RRBIs (proportions not
IQ reported)
20 Zwaigenbaum et al. Social impairment Not reported DSM-IV-TR criteria Viewing of clinician diagnostic 57 28 3.3 64 64 3.3
(2012) Communication impairment for ASD report and ADI-R/ADOS
RRBIs
IQ 55 25 3.3 63 63 3.3
ASC: autism spectrum condition; TD: typically developing; RRBIs: restricted–repetitive behaviours and interests; AS: Asperger’s syndrome; ASD: autism spectrum disorder; HFA: high functioning autism; PDD: pervasive developmen-
tal disorder; PDD-NOS: pervasive developmental disorder, not otherwise specified; LFA: low functioning autism; AD: autistic disorder; ADI-R: Autism Diagnostic Interview–Revised; ADOS: autism diagnostic observation schedule;
SCQ: Social Communication Questionnaire; AAA: Adult Asperger’s Assessment; DSM-IV: Diagnostic and Statistical Manual of Mental Disorders (4th ed.); ASDS: Asperger’s Syndrome Diagnostic Scale; AQ: autism quotient; DSM-IV-TR:
Diagnostic and Statistical Manual of Mental Disorders (4th ed., text rev.); ICD-10: International Classification of Diseases (10th ed.).
711
712
Table 2. Measures and key findings of articles included in review.

Article Authors (date) Outcome measures Key findings
1 Baron-Cohen et al. (2014) EQ EQ: significant interaction between sex and diagnosis found smaller sex differences in ASC than TD group (F(df = 1, 4351) = 14, p < 0.001, ω = 0.06); ASC
female > ASC male (F(df = 1, 4351) = 33.4, p < 0.001, d = 0.40); TD female > TD male (F(df = 1, 4351) = 455, p < 0.001, d = 0.76)
SQ SQ: significant interaction between sex and diagnosis found smaller sex differences in ASC than TD group (F(df = 1, 4146) = 11.6, p < 0.001, ω = 0.06); ASC
male > ASC female (F(df = 1, 4146) = 15.6, p < 0.001, d = 0.27); TD male > TD female (F(df = 1, 4146) = 275.36, p < 0.001, d = 0.61)
AQ AQ: significant interaction between sex and diagnosis found smaller sex differences in ASC than TD group (F(df = 1, 4713) = 3.94, p = 0.047, ω = 0.02); ASC
male > ASC female (F(df = 1, 4713) = 10.97, p < 0.001, d = 0.18); TD male > TD female (F(df = 1, 4713) = 133, p < 0.001, d = 0.41)
2 Baron-Cohen et al. (2003) EQ EQ: no significant difference between ASC female and ASC male (t(df = 18.68) = 1.09, p = 0.22); TD female > TD male (F(df = 1, 269) = 38.6, p < 0.001)
SQ SQ: no significant difference between ASC female and ASC male (t(df = 45) = −0.46, p > 0.65); TD male > TD female (F(df = 1, 270) = 18.1, p < 0.001)
3 Baron-Cohen and FQ FQ: no significant interaction between sex and diagnosis (F(df = 1, 139) = 3.5, p = 0.06); TD > ASC (F(df = 1, 139) = 51.6, p < 0.001); female > male (F(df = 1,
Wheelwright (2005) 139) = 16.8, p < 0.001)
4 Baron-Cohen et al. (2001) AQ AQ: significant interaction between sex and diagnosis (F(df = 1, 228) = 6.01, p = 0.02); no significant difference between ASC female and ASC male (statistical
tests not reported); TD male > TD female (t = 2.56, p < 0.01)
5 Bölte et al. (2011) WISC WISC: no significant interaction between sex and diagnosis (F = 0.07, p = 0.79, partial η2 = 0.00)
WCST WCST: no significant interaction between sex and diagnosis (F = 0.09, p = 0.75, partial η2 = 0.00)
ToH ToH: no significant interaction between sex and diagnosis for number of moves (F = 2.22, p = 0.07, partial η2 = 0.03) or completion time (F = 0.00, p = 0.96,
partial η2 = 0.00)
TMT TMT: significant interaction between sex and diagnosis (F = 3.91, p = 0.04, partial η2 = 0.04); ASC females were faster than ASC males (statistical tests not
reported); TD males were faster than TD females (statistical tests not reported)
EFT EFT: no significant interaction between sex and diagnosis (F = 0.02, p = 0.88, partial η2 = 0.00)
BDT BDT: significant interaction between sex and diagnosis (F = 5.56, p = 0.02, partial η2 = 0.05); ASC males performed better than ASC females (statistical tests
not reported); TD females performed better than TD males (statistical tests not reported)
6 Dean et al. (2014) Friendships Survey Friendships Survey: no significant interaction between sex and diagnosis for social preferences (F(df = 1, 96) = 1.09, p = 0.30, ω2 = 0.2), social acceptance
(F(df = 4, 95) = 0.41, p = 0.53, ω2 = 1.35) or social connections (F(df = 3, 96) = 1.35, p = 0.25, ω2 = 0.01)
7 Goddard et al. (2014) WASI WASI: no significant difference between ASC and TD scores (t = 0.12, p = 0.94)
BPVS BPVS: no significant difference between ASC and TD scores (t = 1.3, p = 0.24)
AMCT AMCT: significant interaction between sex and diagnosis (F(df = 1, 44) = 4.24, p = 0.045, η2 = 0.09); ASC females produced more autobiographical memories
than ASC males (statistical tests not reported); no difference between TD females and TD males (statistical tests not reported)
RRMT RRMT: no significant interaction between sex and diagnosis (statistical tests not reported)
8 Harrop et al. (2016) Mullen ELC Mullen: no significant difference between ASC female and ASC male (t(df = 26) = 9.15, p = 0.37); no significant difference between ASC and TD scores (t(df = 3,
50) = 0.94, p = 0.96)
Toy Engagement Toy Engagement: significant interaction between sex and diagnosis for garage and cars (F(df = 3, 50) = 20.21, p < 0.001); TD males played more than ASC
males (p = 0.04) or TD females and ASC females (p < 0.001); no significant interactions for other types of play
9 Head et al. (2014) FQ FQ: no significant interaction between sex and diagnosis (F(df = 1, 101) = 1.00, p > 0.05, η2 = 0.01); ASC females > ASC males (t(df = 48) = −3.64, p < 0.05)
10 Holt et al. (2014) WASI WASI: no significant difference between ASC male and TD male (statistical tests not reported); ASC female < TD female (p = 0.001)
RMET RMET: ASC < TD (p = 0.002); ASC male < TD male (F(df = 2, 61) = 3.39, p = 0.004); no significant difference between ASC female and TD female (F(df = 2,
60) = 2.02, p = 0.141); no significant interaction between sex and diagnosis (statistical tests not reported)
11 Kirkovski et al. (2016) KBIT-2 KBIT-2: no significant difference between ASC and TD (statistical tests not reported)
RAADS-R RAADS-R: ASC > TD on all subscales (statistical tests not reported)
EQ EQ: ASC < TD (statistical tests not reported)
AQ AQ: ASC > TD (statistical tests not reported)
12 Knickmeyer et al. (2008) CPQ CPQ: sex-typical play shown by TD females (t(df = 42) = 11.58, p < 0.001), TD males (t(df = 60) = 13.55, p < 0.001) and ASC males (t(df = 45) = 11.8, p < 0.001);
sex-typical play not shown by ASC females (t(df = 19) = −1.30, p = 0.21)
(Continued)
Autism 21(6)
Hull et al.
13 Lai et al. (2012) WASI WASI: no significant difference between female ASC, male ASC, female TD or male TD groups (statistical tests not reported)
Go–No-Go Task Go–No-Go Task: no significant interaction between sex and diagnosis (F(df = 2, 120) = 0.173, p = 0.842, Pillai’s trace V = 0.003)
EFT EFT: marginally significant interaction between sex and diagnosis (F(df = 1, 122) = 137.40, p < 0.001); ASC male < TD male (p = 0.001); no significant difference
between ASC female and TD male (p = 0.83); no significant difference between ASC female and ASC male (p = 0.04); TD male > TD female (p < 0.001)
RMET RMET: no significant interaction between sex and diagnosis (F(df = 1, 122) = 0.42, p = 0.521, partial η2 = 0.003)
KDEFT KDEFT: no significant interaction between sex and diagnosis on any emotion (see article for test results)
NWRT NWRT: no significant interaction between sex and diagnosis (F(df = 1, 122) = 0.23, p = 0.635, Pillai’s trace V = 0.002)
14 Lemon et al. (2011) WISC WISC: no statistical tests reported
Stop Task Stop Task: significant effect of group (F(df = 3, 19) = 3.87, p = 0.026); ASC females were slower than TD females (p = 0.002, d = 1.30) and TD males (p = 0.025,
d = 0.86); no significant difference between ASC males and TD males (p = 0.919, d = 0.05)
15 May et al. (2012) SRS SRS: no significant interaction between sex and diagnosis (statistical tests not reported); ASC group > TD group (F = 229.871, p < 0.001); no sex differences
(F = 0.996, p not reported)
CCC CCC: no significant interaction between sex and diagnosis (statistical tests not reported); ASC group < TD group on all subscales (see article for test
results); males < females for some subscales (see article for test results)
RBQ RBQ: no significant interaction between sex and diagnosis (statistical tests not reported); ASC group > TD group (F = 85.397, p < 0.001); males > females for
one subscale (see article for test results)
WISC/WASI WISC/WASI Full-Scale IQ: TD group > ASC group (F = 7.716, p < 0.001)
SWAN SWAN: ASC group > TD group for hyperactivity (F = 60.08, p < 0.001) and inattention (F = 83.08, p < 0.001); males > females for hyperactivity (F = 4.51,
p < 0.05) and inattention (F = 4.28, p < 0.05)
Conners 3 Parent Conners 3: significant interaction between sex, age and diagnosis for hyperactivity (F(df = 1, 122) = 4.279, p = 0.041); no sex differences for inattention
Short Form (F = 2.981, p not reported); ASC group > TD group for inattention (F = 80.089, p < 0.001)
16 Oswald et al. (2016) KBIT-2 KBIT-2: no significant difference between ASC and TD scores (F < 0.01, p not reported)
RCADS RCADS: significant interaction between sex, diagnosis and developmental stage (F(df = 2, 54) = 3.30, p = 0.04, partial η2 = 0.11); ASC female > ASC male and
TD female in early adolescence but no difference between ASC female and ASC male by late adolescence (all ps < 0.01)
MASC MASC: marginally significant interaction between sex, diagnosis and developmental stage (F(df = 1, 55) = 3.79, p = 0.06, partial η2 = 0.06; ASC female > ASC
male and TD female in early adolescence (all ps < 0.01)
CES-D CES-D: marginally significant interaction between sex, diagnosis and developmental stage (F(df = 4, 51) = 2.17, p = 0.09, partial η2 = 0.15; ASC female and TD
male > ASC male and TD female in early adolescence but ASD female and male > TD female and male by late adolescence (all ps < 0.05)
17 Park et al. (2012) SCQ SCQ: ASC male > ASC female (t = 2.27, p < 0.001); no significant difference between TD male and TD female (t = 0.62, p = 0.54)
ASDS ASDS: no significant difference between ASC female and ASC male on any subscale (see article for test results); no significant difference between TD female
and TD male on any subscale (see article for test results)
ADI-R ADI-R: ASC male > ASC female for communication impairments (t = 2.34, p = 0.028) and repetitive, stereotyped behaviours (t = 2.03, p = 0.045); no significant
difference between TD female and TD male for any core ASC symptom (see article for test results)
LIPS LIPS: TD group > ASC group (F = 26.80, p ⩽ 0.001)
CBC CBC: no significant difference between ASC female and ASC male on any subscale (see article for test results); no significant difference between TD female
and TD male on any subscale (see article for test results)
AQ AQ: ASC male > ASC female (t = 2.19, p = 0.031); no significant difference between TD male and TD female (t = 1.76, p = 0.085)
EQ EQ: no significant difference between ASC female and ASC male (t = −0.53, p = 0.605); no significant difference between TD female and TD male (t = −1.67,
p = 0.104)
SQ SQ: no significant difference between ASC female and ASC male (t = 0.87, p = 0.388); TD male > TD female (t = 2.52, p = 0.016)
713
714
18 Sedgewick et al. (2015) SRS SRS: significant interaction between sex and diagnosis (F(df = 1, 42) = 4.79, p = 0.03, partial η2 = 0.10); ASC male > ASC female (t(df = 21) = 0.242, p = 0.03,
d = 1.03); no significant difference between TD male and TD female (t (df = 21) = 0.26, p = 0.12)
WAIS WAIS: no significant effect of sex (p > 0.33) or diagnosis (p > 0.18); no significant interaction between sex and diagnosis (p > 0.33)
FQS FQS: significant interaction between sex and diagnosis for help (F(df = 1, 42) = 6.21, p = 0.01, partial η2 = 0.13) and closeness (F(df = 1, 42) = 6.26, p = 0.01, partial
η2 = 0.13) subscales; no significant interactions found for other subscales (see article for test results)
19 Solomon et al. (2012) SRS SRS: no significant difference between ASC female and ASC male on any subscale (statistical tests not reported); ASC female > TD female on all subscales
(statistical tests not reported); no difference between ASC male and TD male on any subscale (statistical tests not reported); no difference between TD
female and TD male on any subscale (statistical tests not reported)
CCC CCC: no significant difference between ASC female and ASC male on any subscale (statistical tests not reported); TD female > ASC female on all subscales
(statistical tests not reported); TD male > ASC male on all subscales (statistical tests not reported); no significant difference between TD female and TD male
(statistical tests not reported)
RBS RBQ: no significant difference between ASC female and ASC male on any subscale (statistical tests not reported); ASC female > TD female for all subscales
but one (statistical tests not reported); ASC male > TD male for all subscales (statistical tests not reported); no significant difference between TD female and
TD male (statistical tests not reported)
WASI WASI: TD group > ASC group (statistical tests not reported)
BASC BASC: no significant difference between ASC female and ASC male on any subscale (statistical tests not reported); ASC female > TD female on all subscales
(statistical tests not reported); ASC male > TD male on depression only (statistical tests not reported); no significant difference between TD female and TD
male (statistical tests not reported)
20 Zwaigenbaum et al. (2012) ADI-R ADI-R: no significant interaction between sex and diagnosis for any subscale (see article for test results); males > females for communication (F = 19.5,
p < 0.001) and social impairments (F = 3.95, p = 0.049); ASC group > TD group for all subscales (see article for test results)
Mullen ELC Mullen ELC: no significant interaction between sex and diagnosis for any subscale (see article for test results)
ASC: autism spectrum condition; TD: typically developing; EQ: empathising quotient; SQ: systemising quotient; AQ: autism quotient; FQ: friendship quotient; WISC: Wechsler Intelligence Scales for Children; WCST: Wisconsin Card
Sort Test; ToH: Tower of Hanoi; TMT: Trail-Making Test; EFT: Embedded Figures Task; BDT: Block Design Task; WASI: Wechsler Abbreviated Scale of Intelligence; BPVS: British Picture Vocabulary Scale; AMCT: Autobiographical
Memory Cueing Task; RRMT: Recent–Remote Memory Task; CPQ: Child Play Questionnaire; RMET: Reading the Mind in the Eyes Task; KDEFT: Karolinska Directed Emotional Faces Task; NWRT: Non-Word Repetition Task; SRS:
Social Responsiveness Scale; CCC: Children’s Communication Checklist; RBS: Repetitive Behaviours Scale; SWAN: strengths and weaknesses in attention-deficit hyperactivity symptoms; SCQ: Social Communication Questionnaire;
ASDS: Asperger’s Syndrome Diagnostic Scale; RBQ: Repetitive Behaviours Questionnaire; ADI-R: Autism Diagnostic Interview–Revised; LIPS: Leiter International Performance Scale; CBC: Child Behavioural Checklist; BASC: Behav-
iour Assessment System for Children; Mullen ELC: Mullen Early Learning Composite; KBIT-2: Kaufman Brief Intelligence Test–2nd ed.; RAADS-R: Ritvo Autism and Asperger’s Diagnostic Scale–Revised; FQS: Friendship Qualities
Scale; RCADS: Revised Child Anxiety and Depression Scale; MASC: Multidimensional Anxiety Scale for Children; CES-D: Centre for Epidemiological Studies Depression Scale.
Degrees of freedom (df) for tests are included where reported in original articles.
Autism 21(6)
Hull et al.
Table 3. Sex/gender differences in social and communication impairments for autism spectrum condition (ASC) and typically developing (TD) groups.
Authors (date) Test used Age of ASC TD

participants
Female (SD) Male (SD) SMD (95% CI) Female (SD) Male (SD) SMD (95% CI)
Social impairments
Kirkovski et al. (2016) RAADS-R social Adult 28.36 (13.87) 28.77 (13.80) −0.03 (−0.78, 0.73) 3.92 (4.01) 4.36 (4.61) −0.10 (−0.87, 0.67
relatedness subscale
May et al. (2012) SRS Child 97.41 (31.77) 99.97 (22.71) −0.09 (−0.58, 0.40) 23.17 (16.49) 27.30 (20.42) −0.22 (−0.73, 0.29)
Park et al. (2012) ADI-R social subscale Child 8.55 (4.43) 10.25 (3.83) −0.43 (−0.92, 0.06) 1.00 (1.22) 1.28 (1.46) −0.20 (−0.76, 0.35)
Sedgewick et al. (2015) SRS Adolescent 72.00 (32.39) 103.00 (27.76) −0.98 (−1.85, −0.11) 43.00 (13.18) 40.00 (26.16) 0.15 (−0.68, 0.97)
Solomon et al. (2012) SRS Child or 103.85 (27.64) 104.60 (32.04) −0.02 (−0.64, 0.60) 18.11 (18.79) 62.81 (60.81) −1.00 (−1.69, −0.30)
adolescent
Zwaigenbaum et al. ADI-R social subscale Child 10.75 (7.00) 11.30 (5.90) −0.09 (−0.54, 0.37) 2.25 (2.34) 2.70 (2.66) −0.18 (−0.44, 0.08)
(2012)
Communication impairments
May et al. (2012) CCC – global Child 36.75 (15.05) 33.19 (16.00) 0.23 (−0.26, 0.72) 80.60 (22.94) 78.63 (19.78) 0.09 (−0.42, 0.60)
communication subscale
Park et al. (2012) ADI-R non-verbal Child 17.75 (8.20) 22.31 (6.16) −0.69 (−1.18, −0.20) 1.80 (2.33) 1.50 (1.90) 0.14 (−0.41, 0.69)
communication subscale
Solomon et al. (2012) CCC – global Child or 76.00 (14.93) 80.95 (24.55) −0.24 (−0.86, 0.38) 113.05 (16.20) 111.00 (16.37) 0.12 (−0.53, 0.78)
communication subscale adolescent
Zwaigenbaum et al. ADI-R communication Child 8.71 (4.54) 10.09 (3.61) −0.35 (−0.80, 0.11) 1.71 (2.14) 2.85 (2.86) −0.45 (−0.71, −0.19)
(2012) subscale
SD: standard deviation; SMD: standardised mean difference; CI: confidence interval; SRS: Social Responsiveness Scale; ADI-R: Autism Diagnostic Interview–Revised; RAADS-R: Ritvo Autism and Asperger’s
Diagnostic Scale–Revised; CCC: Child Communication Checklist.
715
716
Table 4. Sex/gender differences in restricted/repetitive behaviours and interests (RRBIs) for ASC and TD groups.

participants
Kirkovski et al. (2016) RAADS-R Adult 53.79 (21.67) 57.00 (12.85) −0.17 (−0.93, 0.58) 8.42 (5.84) 12.09 (12.00) −0.38 (−1.21, 0.45)
circumscribed interests
May et al. (2012) RBQ Child 35.48 (31.77) 38.34 (9.01) −0.12 (−0.61, 0.37) 23.23 (4.52) 23.86 (3.42) −0.16 (−0.66, 0.35)
Park et al. (2012) ADI-R RBS subscale Child 4.10 (2.51) 5.48 (2.79) −0.50 (−0.99, −0.01) 0.36 (0.70) 0.50 (0.81) −0.18 (−0.73, 0.37)
Solomon et al. (2012) RBS Child or 2.47 (1.77) 5.00 (3.16) −0.97 (−1.62, −0.31) 0.00 (0.00) 0.41 (1.23) −0.48 (−1.14, 0.19)
adolescent
Zwaigenbaum et al. (2012) ADI-R RBS subscale Child 4.43 (2.60) 4.07 (2.68) 0.13 (−0.32, 0.59) 0.74 (1.26) 1.21 (1.74) −0.31 (−0.57, −0.05)
ASC: autism spectrum condition; TD: typically developing; SD: standard deviation; SMD: standardised mean difference; CI: confidence interval; RBQ: Repetitive Behaviours Questionnaire; RBS: Repetitive
Behaviours Scale; ADI-R: Autism Diagnostic Interview–Revised; RAADS-R: Ritvo Autism and Asperger’s Diagnostic Scale–Revised.
Autism 21(6)
Hull et al.
Table 5. Sex/gender differences in IQ for ASC and TD groups.

participants
Bölte et al. (2011) WISC Non-Verbal Child or 98.60 (9.80) 99.80 (11.30) −0.11 (−0.65, 0.43) 102.30 (12.80) 104.70 (13.30) −0.18 (−0.71, 0.35)
IQ adolescent
Goddard et al. (2014) Wechsler Full-Scale Child or 107.40 (13.50) 104.30 (12.40) 0.23 (−0.57, 1.03) 106.00 (11.10) 106.60 (11.20) −0.05 (−0.85, 0.75)
IQ adolescent
Harrop et al. (2016) MSEL Child 27.12 (10.27) 27.20 (10.92) 0.91 (0.13, 1.69) 23.35 (7.86) 22.50 (8.05) 0.10 (−0.67, 0.87)
Holt et al. (2014) Wechsler Full-Scale Adolescent 96.44 (11.68) 108.42 (19.47) −0.68 (−1.29, −0.07) 110.55 (12.66) 112.3 (11.57) −0.14 (−0.76, 0.48)
IQ
Kirkovski et al. (2016) KBIT-2 Adult 107 (14.48) 112.08 (14.37) −0.34 (−1.10, 0.42) 113 (11.71) 112.45 (16.20) 0.04 (−0.78, 0.86)
Lai et al. (2012) Wechsler Full-Scale Adult 114.10 (15.50) 113.70 (15.10) 0.03 (−0.46, 0.52) 119.70 (8.40) 116.30 (11.80) 0.33 (−0.17, 0.82)
IQ
Lemon et al. (2011) Wechsler Full-Scale Child or 97.30 (16.74) 91.68 (18.40) 0.31 (−0.52, 1.14) 107.00 (10.72) 108.00 (11.00) −0.09 (−0.96, 0.78)
IQ adolescent
May et al. (2012) Wechsler Full-Scale Child 96.19 (12.62) 97.38 (9.01) −0.11 (−0.60, 0.38) 106.50 (11.25) 108.43 (11.99) −0.16 (−0.67, 0.34)
IQ
Oswald et al. (2016) KBIT-2 Adolescent 107.64 (18.13 112.11 (11.67) −0.29 (−1.00, 0.41) 108.79 (16.80) 110.61 (10.60) −0.13 (−0.83, 0.57)
Park et al. (2012) LIPS Child 92.00 (25.61) 93.34 (23.67) −0.06 (−0.54, 0.43) 123.96 (11.37) 121.68 (9.11) 0.22 (−0.33, 0.77)
Sedgewick et al. (2015) Wechsler Full-Scale Adolescent 81.17 (11.50) 78.40 (11.26) 0.31 (−0.52, 1.13) 76.54 (10.25) 76.54 (10.25) 0.00 (−0.82, 0.82)
IQ
Solomon et al. (2012) Wechsler Full-Scale Child or 104.20 (15.29) 103.95 (16.87) 0.02 (−0.60, 0.64) 113.26 (10.23) 121.65 (11.01) −0.77 (−1.45, −0.10)
IQ adolescent
Zwaigenbaum et al. Mullen – Receptive Child 40.20 (13.00) 41.90 (13.40) −0.13 (−0.60, 0.35) 55.00 (9.60) 51.50 (10.50) 0.35 (0.09, 0.61)
(2012) Language Subscale
ASC: autism spectrum condition; TD: typically developing; SD: standard deviation; SMD: standardised mean difference; CI: confidence interval; WISC: Wechsler Intelligence Scales for Children; LIPS:
Leiter International Performance Scale; KBIT-2: Kaufman Brief Intelligence Test–2nd ed.; MSEL: Mullen Scales of Early Learning.
717
718 Autism 21(6)
Figure 2. Funnel plots of studies included in meta-analysis of sex/gender differences in ASC and typically developing populations.
Studies compared social impairments (n = 6), communication impairments (n = 4), restricted/repetitive behaviours and interests
(n = 5) and IQ (n = 13).
inclusion in meta-analyses) than other studies. This is finding smaller sex/gender differences in ASC adolescents
because the sex/gender difference is usually a supplemen- than TD, and the study by Solomon et al. (2012) finding
tary analysis to the research question of interest, and so, the opposite effect. In those studies which only included
publication is less likely to be dependent on satisfactory children or adults (n = 4), no significant variation in sex/
sex/gender difference results. gender differences between ASC and TD groups was
found. However, the test for residual heterogeneity was
Social and communication impairments. Table 3 displays the significant, QE (df = 2) = 43.19, p < 0.001, indicating that
mean scores, test used and sex/gender differences in social other moderators, not included in the model, may still be
and communication impairments for ASC and TD groups. influencing the effect of sex/gender.
Random-effects meta-analysis found no significant differ- No significant difference in communication impair-
ences between social impairments for ASC males or ments was found using meta-analysis for ASC males and
females across studies, SMD = −0.21, 95% confidence females, SMD = −0.26, 95% CI = (−0.65, 0.12), or TD
interval (CI) = (−0.44, 0.02). TD females were found to males and females, SMD = −0.09, 95% CI = (−0.44, 0.26).
have significantly lower levels of social impairments than A random-effects meta-analysis (n = 4) revealed no sig-
TD males, SMD = −0.23, 95% CI = (−0.42, −0.04). Never- nificant difference in the effect of sex/gender for the
theless, a random-effects meta-analysis revealed no sig- ASC-TD groups, SMD = −0.90, 95% CI = (−2.52, 0.72)
nificant difference in the effect of sex/gender between the (Figure 4). Significant heterogeneity was found in this
ASC and TD groups. analysis (Q = 174.91, p < 0.001); therefore, the modera-
Significant heterogeneity was found in this analysis tor Age was included in the model but was not found to
(Q = 158.76, p < 0.001); therefore, the moderator Age was be significant, QM (df = 1) = 0.01, p = 0.91. In contrast,
included in the model and found to be significant, QM the test for residual heterogeneity was significant, QE
(df = 4) = 20.53, p < 0.001. The resulting mixed-effects (df = 2) = 166.56, p < 0.001, indicating that other modera-
meta-analysis (see Figure 3) found significant variation in tors not included in the model may still be influencing the
sex/gender differences for social impairment between effect of sex/gender.
ASC and TD groups for studies including adolescents
(n = 2). However, these two studies found different patterns RRBIs. Table 4 displays the mean scores, test used and
of variation, with the study by Sedgewick et al. (2015) sex/gender differences for ASC and TD groups. The
Hull et al. 719
Figure 3. Meta-analysis of studies comparing differences in sex/gender variation in social impairment between ASC and TD groups.
Forest plot of standardised mean differences (SMDs) for social impairment in each study and total SMD at each level of moderator
‘Age’, drawn in R using ‘metafor’ package (Viechtbauer, 2010; R Foundation for Statistical Computing, Vienna, Austria).
ASC: autism spectrum condition; TD: typically developing; CI: confidence interval.
Central rectangle indicates mean effect, lines indicate 95% confidence intervals, negative effects indicate smaller sex/gender differences in ASC
groups than in TD groups and positive effects indicate larger sex/gender differences in ASC groups than in TD groups. If lines cross the y-axis, the
effect is not significant. Rectangles indicate the effect size (SMD) in each study, with the size of the rectangle indicating the ‘weight’ of the study
(determined by the sample size and the precision of the confidence intervals); diamonds indicate the average effect size in each group of studies; with
wider diamonds indicating wider confidence intervals of the effect.
Figure 4. Meta-analysis of studies comparing differences in sex/gender variation in communication impairment between ASC and
TD groups. Forest plot of standardised mean differences (SMDs) for communication impairment in each study and total SMD from
all studies, drawn in R using ‘metafor’ package (Viechtbauer, 2010; R Foundation for Statistical Computing, Vienna, Austria).
effect is not significant. Rectangles indicate the effect size (SMD) in each study, with the width of the rectangle indicating the ‘weight’ of the study
(determined by the sample size and the precision of the confidence intervals), and the diamond indicates the average effect across all studies, with
the width of the diamond indicating the confidence intervals of the effect.
720 Autism 21(6)
extent of RRBIs was not significantly different between therefore, it is possible that significant interactions were,
males and females with ASC, SMD = −0.30, 95% in fact, undetected in these studies.
CI = (−0.66, 0.07). TD females had significantly lower In contrast, the Trail-Making Test was found to produce
levels of RRBIs than TD males, SMD = −0.29, 95% significantly different sex/gender relative performances
CI = (−0.49, −0.09). A random-effects meta-analysis depending on diagnostic status. In the ASC group, males had
(n = 5) revealed no significant difference in the effect of significantly longer reaction times than females, but in the
sex/gender for the ASC-TD groups, SMD = 0.09, 95% TD group, females took longer to complete the task than
CI = (−1.30, 1.48) (see Figure 5). Significant heterogene- males (Bölte et al., 2011). With regard to the Stop Task,
ity was found in this analysis (Q = 255.24, p < 0.001); Lemon et al. (2011) found that ASC females demonstrated
therefore, the moderator Age was included in the analy- significantly longer reaction times than ASC males or TD
sis. However, omnibus testing revealed no significant females, while no differences were found between ASC
effect of Age, QM (df = 1) = 0.71, p = 0.40. In contrast, the males’ and TD males’ performance on this task (see Table 2).
test for residual heterogeneity was significant, QE
(df = 3) = 225.11, p < 0.001, indicating that other modera- Attention to detail. Some theories of ASC propose that indi-
tors not included in the model may still be influencing the viduals with ASC have a bottom-up, centrally focused pro-
effect of sex/gender. cessing style as opposed to the TD top-down, holistic
processing style (Happé and Frith, 2006). Bölte et al.
IQ. Table 5 displays the mean scores, test used and sex/ (2011) found no significant interaction between sex/gen-
gender differences for ASC and TD groups. There were no der and diagnosis for the Embedded Figures Task (EFT),
significant differences between ASC male and ASC female although a marginal interaction was found by Lai et al.
IQ scores, SMD = −0.05, 95% CI = (−0.22, 0.12), or TD (2012). In the latter study, ASC males demonstrated poorer
male and female IQ scores, SMD = 0.02, 95% CI = (−0.17, performance on the EFT than TD males, while no differ-
0.21). A random-effects meta-analysis (n = 13) revealed no ences were found between ASC and TD females. As above,
significant difference in the effect of sex/gender for the it is possible that small effect sizes went undetected in
ASC versus TD groups, SMD = −0.09, 95% CI = (−0.88, these studies. However, a significant sex/gender and diag-
0.71) (see Figure 6). Significant heterogeneity was found nosis interaction was found on the Block Design Task (see
(Q = 453.68, p < 0.001); therefore, the moderator Age was Table 2). ASC males performed better than ASC females,
included in a mixed-effects meta-analysis but was not whereas the reverse pattern was found for TD individuals
found to be a significant moderator, QM (df = 1) = 2.45, (Bölte et al., 2011).
p = 0.12. The test for residual heterogeneity was signifi-
cant, QE (df = 11) = 399.72, p < 0.001, indicating that other Theory of mind/emotion recognition. The ability to infer the
moderators not included in the model may still be influ- content of others’ mental and emotional states, regardless
encing the effect of sex/gender. of whether they are different from one’s own, is known as
theory of mind. Late or incomplete development of theory
of mind abilities is considered a hallmark of ASC (Baron-
Systematic qualitative review
Cohen et al., 1985), with some individuals failing to
Executive functioning. Executive functions are a set of abili- achieve ‘simple’ theory of mind abilities, such as recognis-
ties which facilitate higher level cognitive control of ing emotional expressions, and others struggling only with
behaviour, self-monitoring and future planning, among more complex tests, such as dynamic interactions (Baron-
other tasks (Ozonoff and Jensen, 1999). Individuals with Cohen et al., 1997).
ASC are often reported to have lower levels of executive No significant sex/gender and diagnosis interaction was
functions than TD individuals (Happé et al., 2006). There found for either the Reading the Mind in the Eyes Task
are contradictions in the literature when it comes to perfor- (RMET) or the Karolinska Directed Emotional Faces Task
mance on specific tasks of executive functioning. All stud- (Holt et al., 2014; Lai et al., 2012; see Table 2). However,
ies examined here found that as a group, individuals with post hoc analyses by Holt et al. (2014) revealed that ASC
ASC performed more poorly than TD individuals. No sta- males performed more poorly than TD males on the
tistically significant interaction between sex/gender and RMET, whereas no significant differences were found
diagnosis was found in the Wisconsin Card Sorting Test or between ASC and TD females. Again, it should be noted
the Tower of Hanoi (Bölte et al., 2011) or the Go/No-Go that both studies had limited power to detect small effect
Task (Lai et al., 2012), suggesting that sex/gender differ- sizes. In line with previous research, these studies found
ences may not vary between ASC and TD groups (see that individuals with ASC generally demonstrated poorer
Table 2). Both studies had medium sample sizes with rela- theory of mind abilities than TD individuals.
tively high proportions of females in each group (com-
pared to many studies examining sex/gender differences in Memory. No significant sex/gender and diagnosis interac-
ASC) but had limited power to detect small effect sizes; tion was found on the Non-Word Repetition Task (Lai
Hull et al. 721
Figure 5. Meta-analysis of studies comparing differences in sex/gender variation in restricted/repetitive behaviours and interests
(RRBIs) between ASC and TD groups. Forest plot of standardised mean differences (SMDs) for RRBIs in each study and total SMD
from all studies, drawn in R using ‘metafor’ package (Viechtbauer, 2010; R Foundation for Statistical Computing, Vienna, Austria).
Figure 6. Meta-analysis of studies comparing differences in sex/gender variation in IQ between ASC and TD groups. Forest plot
of standardised mean differences (SMDs) for IQ in each study and total SMD at each level of moderator ‘Age’, drawn in R using
‘metafor’ package (Viechtbauer, 2010; R Foundation for Statistical Computing, Vienna, Austria).
722 Autism 21(6)
et al., 2012) or the Recent/Remote Memory Task (Goddard systemising and autistic traits than males and females
et al., 2014; see Table 2). The former task is associated without ASC.
with (non-verbal) auditory working memory, as opposed The study by Park et al. (2012) did not directly test
to verbal memory tasks, which may be influenced by indi- interactions but found no significant differences between
viduals’ language abilities. The Recent/Remote Memory ASC males and females on the SQ or EQ. TD males had
Task measures both short-term and long-term recall mem- higher SQ scores than equivalent females, while no sex/
ory and is scored based on the number of details provided gender differences were found for the TD group on the EQ.
in response to each memory cue (Goddard et al., 2014). In ASC males scored higher on the AQ than ASC females,
this task, individuals with ASC performed more poorly but no significant differences were found between TD
than TD individuals; otherwise, there were no group dif- males and females on this measure (see Table 2). Similarly,
ferences for these tasks. However, a significant sex/gender Kirkovski et al. (2016) found that ASC females and males
and diagnosis interaction was found for the Autobiograph- scored lower on the EQ and higher on the AQ than TD
ical Memory Cueing Task. Males with ASC were found to females and males. Sex/gender differences within diagno-
produce fewer autobiographical memories than females sis groups were not reported in this study.
with ASC, whereas no such difference was found between
TD males and females (Goddard et al., 2014). Friendship. No significant interactions between sex/gender
and diagnosis were found using either the Friendship
Empathising, systemising and autistic traits. These traits rep- Questionnaire (Baron-Cohen and Wheelwright, 2005;
resent a continuum of abilities reaching from the typical Head et al., 2014) or the Friendships Survey (Dean et al.,
population through those with an ASC diagnosis. System- 2014). Using both measures, ASC individuals were found
ising ability, measured by the systemising quotient (SQ), to perform more poorly than TD individuals (see Table 2).
represents an interest and understanding of the mecha- However, one study utilising the Friendship Qualities
nisms within a system. High levels of these abilities are Scale found that males with ASC reported significantly
associated with more autistic traits in sub-clinical popula- lower closeness and helping in their best friendship than
tions as well as with a diagnosis of ASC (Baron-Cohen females with ASC or TD children of either sex/gender
et al., 2003). In contrast, higher levels of empathising (Sedgewick et al., 2015).
abilities (measured by the empathising quotient (EQ)),
such as understanding others’ mental states and emotions, Internalising and externalising. As a group, children with
are generally found in individuals without an ASC diagno- ASC experienced more internalising and externalising
sis and are associated with lower levels of autistic traits in behaviours than TD children. Although interactions were
the general population (Baron-Cohen and Wheelwright, not tested, no sex/gender differences were found for exter-
2004). Autistic traits describe behaviours and cognitive nalising or internalising behaviours, as measured by the
styles associated with autism that also exist in the general Child Behaviour Checklist, in either ASC or TD groups
population and are proposed to be more common in TD (Park et al., 2012). However, this study was limited by a
males than females (Baron-Cohen et al., 2001). Autistic low proportion of females with ASC, which means more
traits are measured by the autism quotient (AQ). subtle differences may have been missed. Similarly, inter-
An earlier study using smaller samples found no sig- nalising behaviours measured using the Behavioural
nificant differences between ASC males and females on Assessment System for Children revealed no sex/gender
any of these traits but found higher SQ traits in TD males differences in either ASC or TD groups (Solomon et al.,
and higher EQ traits in TD females (Baron-Cohen et al., 2012). Although this study had a moderate sample size,
2003; see Table 2). Interactions were not directly tested in sex/gender and diagnosis interactions were not directly
this study, which was underpowered to detect small effect tested, and the results of difference tests were not reported
sizes. Baron-Cohen et al. (2001) found a significant inter- (see Table 2).
action between group and sex/gender for AQ scores, with Through parent- and self-report, an interaction between
TD males scoring higher than females, and no sex/gender sex, diagnosis and developmental stage was found for
difference between ASC males and females. In contrast, a depressive symptoms, with ASC females demonstrating
more recent study using a much larger sample and a larger higher levels of depressive symptoms than either ASC
proportion of females with ASC found significant interac- males or TD females in early adolescence (Oswald et al.,
tions between sex/gender and diagnosis on all three traits 2016). However, by late adolescence, ASC males and
(Baron-Cohen et al., 2014). Both ASC and TD groups dis- females were found to have similar levels of depressive
played higher SQ and AQ scores for males and higher EQ symptoms, with the change being explained by ASC males
scores for females (see Table 2). However, sex/gender dif- alone having a significant increase in depressive symp-
ferences in these studies were significantly smaller for toms as they became older (see Table 2). The same study
ASC individuals, suggesting that males and females with also found a marginally significant interaction between
ASC may be more similar in their empathising, sex, diagnosis and developmental stage for anxiety, with
Hull et al. 723
ASC females and TD males reporting higher levels of anx- including adolescents in their sample. One study found
iety than ASC males and TD females in early adolescence, smaller sex/gender differences in ASC than TD groups,
but both ASC males and females reporting higher levels of whereas the other study found larger sex/gender differences
anxiety than their TD peers by late adolescence. for ASC participants. No variation in sex/gender differ-
Some significant sex/gender and diagnosis interactions ences between groups was found for the other four sam-
were found when looking at hyperactivity and inattention ples, which included either children or adults only. Due to
in particular. The study by May et al. (2016) also looked at the small number of studies and contradictory findings in
the effect of age on ASC-related outcomes. They found each of these studies, a conclusion of either greater or
that sex/gender differences varied between ASC and TD smaller sex/gender differences in ASC social impairments
groups but that this variation depended on the age of the cannot be drawn. However, these findings raise the impor-
individuals (see Table 2). Younger males with ASC (aged tance of comparing sex/gender differences across all ages,
7–9 years) were more impaired than younger ASC females, as there may be age-related variation in the similarities and/
compared to typical males and females. By the time these or differences between ASC and TD groups which could
children reached the age of 10–12 years, both ASC and TD not be fully assessed in this limited sample.
groups showed similar sex/gender differences, with males These results suggest that typical sex/gender differ-
having higher levels of attention-deficit hyperactivity dis- ences in core symptoms and IQ also occur for individuals
order (ADHD)-related behaviours than females. As a with ASC and, therefore, that individuals with ASC are
group, children with ASC at all ages demonstrated higher fundamentally similar to TD individuals in regard to their
levels of inattention and hyperactivity than TD children. sex/gender variation in core ASC characteristics. This
reflects the dimensional nature of ASC, such that people
Play behaviours. The study by Knickmeyer et al. (2008) above and below the diagnostic threshold for ASC share
found that ASC females demonstrate significantly less traits which vary between sexes/genders.
sex-typical pretend play relative to their TD peers than However, the review of sex/gender differences in asso-
ASC males. No such sex/gender differences were found ciated ASC symptoms revealed some degree of variation
for non-pretend play, where both males and females with between ASC and typical populations, suggesting that
ASC demonstrate similar play preferences to TD males having an ASC may impact differently males and females.
and females, respectively. In contrast, Harrop et al. (2016) Males with ASC were found to have significantly more
found that ASC males played with sex-typical cars and impaired performance on the Trail-Making Test (one
trucks less than their TD peers, whereas no differences in measure of executive function, focusing on task switching
this play behaviour were found between ASC and TD and cognitive flexibility), to produce fewer autobiographi-
females. While typical sex differences were found for cal memories, and higher levels of hyperactivity (although
other types of play, such as playing with dolls and houses, only at a younger age) than females with ASC, taking into
there were no differences between ASC females and TD account typical sex/gender differences. In contrast, females
females or between ASC males and TD males. One possi- with ASC were found to be significantly more impaired on
ble explanation for these different findings is that the study response inhibition, as measured by the Stop Task, and
by Knickmeyer et al. (2008) utilised a sample with a visual–spatial processing, as measured by the Block
greater range of ages, who were on average older, than the Design Task. Play behaviours in both males and females
sample used by Harrop et al. (see Table 2). with ASC were found to be different from those of TD
males and females. However, the differences appear to
depend on the age of the individual, with ASC females dis-
Discussion
playing more sex-typical behaviours than males as young
This study aimed to compare sex/gender differences children, but this pattern reversing between childhood and
between individuals with ASCs and TD individuals, to early adolescence. Age-related patterns were also found
determine whether the patterns of difference vary between for internalising and externalising problems. At younger
these groups. A difference in sex/gender variation between ages, ASC females generally reported higher levels of
groups would suggest diagnostic criteria for ASC should internalising problems, while ASC males reported higher
differ for males and females, to reflect separate ASC phe- levels of externalising problems, a similar pattern to the
notypes for males and females. TD groups. However, as the ASC children became older,
Meta-analyses found no variation in the profiles of sex/ their levels of internalising and externalising problems
gender differences for ASC and TD groups for the core became more similar. In particular, males with ASC dem-
ASC symptoms of communication impairments and RRBIs onstrated increased levels of internalising problems as
or for IQ. Sex/gender differences in social impairments they developed, bringing them to a similar level as their
were found to vary depending on the age of the partici- female peers.
pants. Different patterns of variation in sex/gender differ- Although patterns of sex/gender differences in autism,
ences of social impairments were found for two studies empathising and systemising traits were the same in both
724 Autism 21(6)
groups, the differences were smaller for the ASC group, than within the general population (Glidden et al., 2016).
suggesting that males and females with ASC are more Although research into gender dysphoria in ASC is rela-
similar in these respects. While some of these findings tively limited, it has been suggested that there are different
contradict those using other measures of the same charac- mechanisms underlying this co-occurrence within males
teristics, they raise the suggestion that male and female and females (Pasterski et al., 2014), which may reflect the
performance may vary depending on the task used and differences in expression of ASC between sexes/genders
encourage further testing of sex/gender differences using found in this study.
a range of measures. The differences that have been found Although these two theories predict males and females
suggest that males and females with ASC are not a homo- with ASC will be relatively similar to each other, they offer
geneous group but may have distinct patterns of ability different predictions about how individuals with ASC are
and impairment which, so far, have not been thoroughly similar and/or different from TD males and females.
investigated. However, these comparisons were not directly tested in
In contrast, no significant interactions between sex/gen- this analysis; therefore, conclusions about whether indi-
der and diagnosis were found for the majority of executive viduals with ASC are more similar to TD males, or differ
function tasks, attention to detail, theory of mind, most from both males and females, must be left for future
measures of friendship and most memory tasks. These research.
results suggest that any sex/gender differences found in The differences found in male and female ASC symp-
ASC groups here can be attributed to typical sex/gender toms may offer some explanation for the differences in
differences, rather than the specific differences found diagnostic rates between sexes/genders in ASC. As
between males and females with ASC. When evaluating recently suggested by Lehnhardt et al. (2015), the greater
ASC sex/gender differences in these areas, typical sex/ task switching and cognitive flexibility abilities of females
gender performance should be taken into account to gain with ASC may explain why they are able to develop com-
true measures of relative ability and impairment. However, pensatory or ‘camouflaging’ techniques to ‘mask’ their
it is also possible that sex/gender variation between ASC social and communication impairments. Lehnhardt et al.
and TD groups in these areas may have differed in size (2015) also found higher processing speed in females than
rather than direction, as was found for some of the cogni- males in their adult diagnosis sample, suggesting that
tive traits associated with ASC. Sex/gender differences in females with ASC are better able to use explicit cognitive
ASC groups may, therefore, be broadly similar to those strategies to cope in complex social interactions. It is pos-
found in TD groups for these characteristics, but these dif- sible that other cognitive and behavioural abilities found in
ferences may be larger within one group than the other. TD females are also utilised by females with ASC when
The smaller sex/gender differences found for ASC camouflaging, although studies of ASC sex/gender differ-
groups in systemising, empathising and autism traits sug- ences with TD controls are still limited. Further explora-
gest that males and females with ASC are more similar to tion of the female phenotype in ASC will give us a greater
each other than TD males and females. This offers some understanding of the tools and techniques used by women
support for recent theories of sex/gender variation in ASC. and girls, which may result in their being missed by clini-
Baron-Cohen’s Extreme Male Brain theory (2002) pro- cal services.
poses that ASC individuals are more ‘masculinised’ than
TD individuals, displaying cognitive and behavioural pat-
Limitations
terns more similar to TD males than females regardless of
the ASC individual’s sex/gender. The Extreme Male Brain A key limitation of this analysis is the small number of
theory would, therefore, predict that males and females studies included, due to a dearth of research comparing
with ASC are more similar than TD males and females and ASC and TD groups. Meta-analysis based on a small num-
even that there might be no sex/gender differences within ber of studies is more susceptible to second-order sam-
the ASC population. pling errors, because variation in standard deviations is
In contrast, Bejerot and Eriksson (2014) suggest that more likely to be influenced by artefacts (Hunter and
both males and females with ASC are different from TD Schmidt, 2004). Several of the studies included in the
males and females, with individuals with ASC displaying qualitative review and meta-analyses were underpowered
gender-atypical patterns of behaviours. According to this to detect small effect sizes, and so, it is possible that sig-
hypothesis, sex/gender differences in ASC might not be nificant variation in sex/gender differences between
significant, but males with ASC would be different from groups was not picked up in our analysis. Consequently,
TD males, and females with ASC would be different from we echo the calls by many others (e.g. Lai et al., 2015) for
TD females. Gender-atypicality in ASC would also future studies to include large enough numbers of males
account for the higher levels of gender dysphoria (incon- and females from both typical and ASC populations, in
gruence between one’s natal sex and experienced gender) order to draw stronger and more consistent conclusions
found within children, adolescents and adults with ASC about sex/gender differences.
Hull et al. 725
Another consequence of the limited number of studies males and females, beyond typical sex/gender variation.
is that few potential confounding variables were identified This supports the conclusions of several previous reports
or controlled for. Age was included as a moderator in the that females with ASC may present different cognitive
meta-analyses and in some of the reviewed studies and and/or behavioural phenotypes than most males with ASC
was found to influence sex/gender variation between and that clinicians should be mindful of these differences
groups in some areas. Previous studies have also identified during assessment and diagnosis. We also suggest that
IQ, ethnicity, co-morbidities and characteristics of ASC there are many individual factors, including age, IQ and
diagnosis, among other factors, as interacting with both social background, which may interact with an ASC to
sex/gender and ASC to produce differential outcomes over produce variations in development and so should not be
time (Brugha et al., 2011; Croen et al., 2002; Farley et al., disregarded in favour of the ‘typical’ ASC presentation. A
2009; Holtmann et al., 2007). Consistent measurement and significant limitation of this study was the small number of
reporting of these characteristics would enable better inter- studies available for review. Future research can address
pretation of these studies’ heterogeneity, which is a signifi- this by ensuring all sex/gender comparisons within ASC
cant limitation of the present meta-analyses. Our results individuals include a comparison group of TD males and
should be interpreted with these limitations in mind, females, to guarantee that sex/gender differences in the
although we conclude that the finding of some significant general population are accounted for in analyses.
variation in sex/gender differences, despite these limita-
tions, is robust and meaningful. Acknowledgements
Although the most recent DSM-5 diagnostic criteria The authors would like to thank Andy Fugard for his advice on
have combined social and communication impairments statistical analyses.
into one symptom, we analysed them separately. This is
because some of the studies included in this analysis only Funding
measured either social or communication impairments;
The author(s) received no financial support for the research,
therefore, scores for both could not be combined for all
authorship and/or publication of this article.
studies. In addition, hypo- or hyper-reactivity to sensory
stimulation is a criterion in DSM-5, but was not measured
Note
in many of the studies included here. Therefore, conclu-
sions cannot be directly applied to the most recent DSM-5 1. Individuals on the autism spectrum have reported that the
criteria but still apply to the International Classification of term ‘disorder’ is stigmatising and does not reflect the
Diseases (10th ed.; ICD-10) diagnostic criteria. range of strengths and difficulties experienced by those on
the spectrum. Following previous researchers (e.g. Lai and
A final limitation is that this study was focused on
Baron-Cohen, 2015), we use the term ‘autism spectrum con-
behavioural and cognitive characteristics of ASC only. dition’ (ASC) in reference to those diagnosed with an autism
While these characteristics are of the most relevance to spectrum disorder.
diagnostic criteria (as physiological markers of ASC have
not been identified, and therefore, diagnosis relies on References
behavioural information solely), there are many other
characteristics of ASC which also display sex/gender vari- Baron-Cohen S (2002) The extreme male brain theory of autism.
Trends in Cognitive Sciences 6(6): 248–254.
ation. This article lacks the space to offer a full review of
Baron-Cohen S and Wheelwright S (2004) The empathy quotient:
sex/gender differences in all areas of research relating to an investigation of adults with Asperger syndrome or high
ASC. However, see recent reviews by Kirkovski et al. functioning autism, and normal sex differences. Journal of
(2013), Lai et al. (2015) and Werling and Geschwind Autism and Developmental Disorders 34(2): 163–175.
(2013) for more information on sex/gender differences in Baron-Cohen S and Wheelwright S (2005) The Friendship
neurodevelopmental, biological and genetic factors among Questionnaire: an investigation of adults with Asperger
other characteristics. A comparison of sex/gender differ- syndrome or high-functioning autism, and normal sex dif-
ences between ASC and TD groups in these characteristics ferences. Journal of Autism and Developmental Disorders
would further broaden our understanding of the expression 33(5): 509–517.
of ASC in both males and females. Baron-Cohen S, Cassidy S, Auyeung B, et al. (2014) Attenuation
of typical sex differences in 800 adults with autism vs. 3,900
controls. PLoS ONE 9(7): e102251.
Conclusion Baron-Cohen S, Jolliffe T, Mortimore C, et al. (1997) Another
advanced test of theory of mind: evidence from very high
The present results suggest that ASC may have differential functioning adults with autism or Asperger syndrome.
impacts on individuals depending on their sex/gender. Journal of Child Psychology and Psychiatry 38(7):
While differences in core symptoms and IQ reflect typical 813–822.
sex/gender patterns of ability, ASC appears to produce dif- Baron-Cohen S, Leslie AM and Frith U (1985) Does the autistic
ferent patterns of some associated ASC characteristics for child have a ‘theory of mind‘? Cognition 21(1): 37–46.
726 Autism 21(6)
Baron-Cohen S, Richler J, Bisarya D, et al. (2003) The system- Journal of Autism and Developmental Disorders 36(1):
izing quotient: an investigation of adults with Asperger 5–25.
syndrome or high-functioning autism, and normal sex dif- Happé F, Booth R, Charlton R, et al. (2006) Executive function
ferences. Philosophical Transactions of the Royal Society deficits in autism spectrum disorders and attention-deficit/
of London, Series B: Biological Sciences 3548: 361–374. hyperactivity disorder: examining profiles across domains
Retrieved from http://rstb.royalsocietypublishing.org/con- and ages. Brain and Cognition 61(1): 25–39.
tent/royptb/358/1430/361.full.pdf Harrop C, Green J and Hudry K (2016) Play complexity and toy
Baron-Cohen S, Wheelwright S, Skinner R, et al. (2001) The engagement in preschoolers with autism spectrum disor-
Autism-Spectrum Quotient (AQ): evidence from Asperger der: do girls and boys differ? Autism. Epub ahead of print 2
syndrome/high-functioning autism, males and females, March. DOI: 10.1177/1362361315622410.
scientists and mathematicians. Journal of Autism and Head AM, McGillivray JA and Stokes MA (2014) Gender differ-
Developmental Disorders 31(1): 5–17. ences in emotionality and sociability in children with autism
Bejerot S and Eriksson JM (2014) Sexuality and gender role in spectrum disorders. Molecular Autism 5: 19. Available at:
autism spectrum disorder: a case control study. PLoS ONE http://www.biomedcentral.com/content/pdf/2040-2392-5-
9(1): e87961. 19.pdf
Bölte S, Duketis E, Poustka F, et al. (2011) Sex differences in Holt RJ, Chura LR, Lai M-C, et al. (2014) ‘Reading the mind
cognitive domains and their clinical correlates in higher- in the eyes’: an fMRI study of adolescents with autism and
functioning autism spectrum disorders. Autism: The their siblings. Psychological Medicine 44(15): 3215–3227.
International Journal of Research and Practice 15(4): Holtmann M, Bölte S and Poustka F (2007) Autism spec-
497–511. trum disorders: sex differences in autistic behaviour
Brugha TS, McManus S, Bankart J, et al. (2011) Epidemiology domains and coexisting psychopathology – ProQuest.
of autism spectrum disorders in adults in the community in Developmental Medicine & Child Neurology 49(5):
England. Archives of General Psychiatry 68(5): 459–466. 361–366. Available at: http://search.proquest.com/
Available at: http://people.uncw.edu/imperialm/UNCW/ docview/195615734/fulltextPDF?accountid=14511
PLS_506/yoa05097_459_466[1].pdf Hunter JE and Schmidt FL (2004) Study artifacts and their
Croen LA, Grether JK and Selvin S (2002) Descriptive epidemi- impact on study outcomes. In: Hunter JE and Schmidt FL
ology of autism in a California population: who is at risk? (eds) Methods of Meta-Analysis. 2nd ed. Thousand Oaks,
Journal of Autism and Developmental Disorders 32(3): CA: SAGE, pp.33–73. Available at: http://srmo.sagepub.
217–224. com/view/methods-of-meta-analysis/n2.xml
Dean M, Kasari C, Shih W, et al. (2014) The peer relationships Kirkovski M, Enticott PG and Fitzgerald PB (2013) A review
of girls with ASD at school: comparison to boys and girls of the role of female gender in autism spectrum disorders.
with and without ASD. Journal of Child Psychology and Journal of Autism and Developmental Disorders 43(11):
Psychiatry, and Allied Disciplines 55(11): 1218–1225. 2584–2603.
Dworzynski K, Ronald A, Bolton P, et al. (2012) How different Kirkovski M, Enticott PG, Hughes ME, et al. (2016) Atypical
are girls and boys above and below the diagnostic thresh- neural activity in males but not females with autism spec-
old for autism spectrum disorders? Journal of the American trum disorder. Journal of Autism and Developmental
Academy of Child and Adolescent Psychiatry 51(8): 788– Disorders 46(3): 954–963.
797. Knickmeyer RC, Wheelwright S and Baron-Cohen SB (2008)
Farley MA, McMahon WM, Fombonne E, et al. (2009) Twenty- Sex-typical play: masculinization/defeminization in girls
year outcome for individuals with autism and average or with an autism spectrum condition. Journal of Autism and
near-average cognitive abilities. Autism Research 2(2): Developmental Disorders 38(6): 1028–1035.
109–118. Koenig K and Tsatsanis KD (2005) Pervasive develop-
Fombonne E (2009) Epidemiology of pervasive developmental mental disorders in girls. In: Bell DJ, Foster SL and
disorders. Pediatric Research 65(6): 591–598. Mash EJ (eds) Handbook of Behavioral and Emotional
Glidden D, Bouman WP and Jones BA (2016) Gender dysphoria Problems in Girls. New York: Kluwer Academic/Plenum
and autism spectrum disorder: a systematic review of the Publishers, pp.211–237. Available at: http://download.
literature. Sexual Medicine Review 4(1): 3–14. springer.com/static/pdf/60/chp:10.1007/0-306-48674-1_7.
Goddard L, Dritschel B and Howlin P (2014) A preliminary pdf?originUrl=http://link.springer.com/chapter/10.1007/0-
study of gender differences in autobiographical memory 306-48674-1_7&token2=exp=1444990643~acl=/static/
in children with an autism spectrum disorder. Journal of pdf/60/chp%3A10.1007%2F0-306-48674-1_7.pdf%
Autism and Developmental Disorders 44(9): 2087–2095. Kopp S and Gillberg C (1992) Girls with social deficits and
Goddard L, Dritschel B, Robinson S, et al. (2014) Development learning problems: autism, atypical Asperger syndrome or
of autobiographical memory in children with autism spec- a variant of these conditions. European Child & Adolescent
trum disorders: deficits, gains, and predictors of perfor- Psychiatry 1(2): 89–99.
mance. Development and Psychopathology 26(1): 215–228. Kopp S and Gillberg C (2011) The Autism Spectrum Screening
Gould J and Ashton-Smith J (2011) Missed diagnosis or misdi- Questionnaire (ASSQ)-Revised Extended Version (ASSQ-
agnosis? Girls and women on the autism spectrum. Good REV): an instrument for better capturing the autism phe-
Autism Practice 12(1): 34–41. notype in girls? A preliminary study involving 191 clinical
Happé F and Frith U (2006) The weak coherence account: cases and community controls. Research in Developmental
detail-focused cognitive style in autism spectrum disorders. Disabilities 32(6): 2875–2888.
Hull et al. 727
Kreiser NL and White SW (2014) ASD in females: are we over- Park S, Cho S-C, Cho IH, et al. (2012) Sex differences in chil-
stating the gender difference in diagnosis? Clinical Child dren with autism spectrum disorders compared with their
and Family Psychology Review 17(1): 67–84. unaffected siblings and typically developing children.
Lai M-C and Baron-Cohen SB (2015) Identifying the lost gener- Research in Autism Spectrum Disorders 6(2): 861–870.
ation of adults with autism spectrum conditions. The Lancet Available at: http://ac.els-cdn.com/S1750946711001978/1-
Psychiatry 2(11): 1013–1027. s2.0-S1750946711001978-main.pdf?_tid=4648c746-77d7-
Lai M-C, Lombardo MV, Auyeung B, et al. (2015) Sex/gender 11e5-8206-00000aacb360&acdnat=1445420398_3b7b24d
differences and autism: setting the scene for future research. 5ae6ee9104067cb193de75725
Journal of the American Academy of Child & Adolescent Pasterski V, Gilligan L and Curtis R (2014) Traits of autism spec-
Psychiatry 54(1): 11–24. trum disorders in adults with gender dysphoria. Archives of
Lai M-C, Lombardo MV, Pasco G, et al. (2011) A behavioral Sexual Behavior 43(2): 387–393.
comparison of male and female adults with high functioning R Core Team (2013) R: A Language and Environment for
autism spectrum conditions. PLoS ONE 6(6): e20835. Statistical Processing. Vienna, Austria: R Foundation for
Lai M-C, Lombardo MV, Ruigrok ANV, et al. (2012) Cognition Statistical Computing. Available at: http://www.r-project.
in males and females with autism: similarities and differ- org/
ences. PLoS ONE 7(10): e47198. Rivet TT and Matson JL (2011) Review of gender differences
Lehnhardt F-G, Falter CM, Gawronski A, et al. (2015) Sex- in core symptomatology in autism spectrum disorders.
related cognitive profile in autism spectrum disorders Research in Autism Spectrum Disorders 23(3): 957–976.
diagnosed late in life: implications for the female autistic Russell G, Steer C and Golding J (2011) Social and demographic
phenotype. Journal of Autism and Developmental Disorders factors that influence the diagnosis of autistic spectrum
46(1): 139–154. disorders. Social Psychiatry and Psychiatric Epidemiology
Lemon JM, Gargaro B, Enticott PG, et al. (2011) Executive func- 46(12): 1283–1293.
tioning in autism spectrum disorders: a gender comparison Sedgewick F, Hill V, Yates R, et al. (2015) Gender differences
of response inhibition. Journal of Autism and Developmental in the social motivation and friendship experiences of autis-
Disorders 41(3): 352–356. tic and non-autistic adolescents. Journal of Autism and
Mandy W and Tchanturia K (2015) Do women with eating disor- Developmental Disorders 46(4): 1297–1306.
ders who have social and flexibility difficulties really have Solomon M, Miller M, Taylor SL, et al. (2012) Autism symp-
autism? A case series. Molecular Autism 6(1): 6. toms and internalizing psychopathology in girls and boys
Mandy W, Chilvers R, Chowdhury U, et al. (2012) Sex differ- with autism spectrum disorders. Journal of Autism and
ences in autism spectrum disorder: evidence from a large Developmental Disorders 42(1): 48–59.
sample of children and adolescents. Journal of Autism and Thompson T, Caruso M and Ellerbeck K (2003) Sex matters
Developmental Disorders 42(7): 1304–1313. in autism and other developmental disabilities. Journal of
Mattila M-L, Kielinen M, Linna S-L, et al. (2011) Autism spec- Intellectual Disabilities 7(4): 345–362.
trum disorders according to DSM-IV-TR and compari- Van Wijngaarden-Cremers PJM, van Eeten E, Groen WB, et al.
son with DSM-5 draft criteria: an epidemiological study. (2014) Gender and age differences in the core triad of impair-
Journal of the American Academy of Child & Adolescent ments in autism spectrum disorders: a systematic review
Psychiatry 50(6): 583–592.e11. and meta-analysis. Journal of Autism and Developmental
May T, Cornish K and Rinehart NJ (2012) Gender profiles of Disorders 44(3): 627–635.
behavioral attention in children with autism spectrum disor- Viechtbauer W (2010) Conducting meta-analyses in R with the
der. Journal of Attention Disorders 20: 627–35. metafor package. Journal of Statistical Software 36(3): 1–
Oswald TM, Winter-Messiers MA, Gibson B, et al. (2016) 48. Available at: http://www.jstatsoft.org/v36/i03/
Sex differences in internalizing problems during adoles- Werling DM and Geschwind DH (2013) Sex differences in
cence in autism spectrum disorder. Journal of Autism and autism spectrum disorders. Current Opinion in Neurology
Developmental Disorders 46(2): 624–636. 26(2): 146–153.
Ozonoff S and Jensen J (1999) Brief report: specific executive Zwaigenbaum L, Bryson SE, Szatmari P, et al. (2012) Sex dif-
function profiles in three neurodevelopmental disorders. ferences in children with autism spectrum disorder identi-
Journal of Autism and Developmental Disorders 29(2): fied within a high-risk infant cohort. Journal of Autism and
171–177. Developmental Disorders 42(12): 2585–2596.

Behavioural and Cognitive

Uploaded by

Copyright:

Available Formats

Behavioural and Cognitive

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Behavioural and Cognitive

Uploaded by

Copyright:

Available Formats

669087