668653

research-article2016
AUT0010.1177/1362361316668653AutismSutherland et al.

Special Issue Article

Autism

Parent-reported differences 1­–10

© The Author(s) 2017
Reprints and permissions:
between school-aged girls and sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/1362361316668653

boys on the autism spectrum journals.sagepub.com/home/aut

Rebecca Sutherland1,2,3, Antoinette Hodge2, Susan Bruck1,

Debra Costley1 and Helen Klieve3

Abstract
More boys than girls are diagnosed with autism spectrum disorder; however, there are conflicting findings about whether
they differ in their presentation. This study involved a survey of parents of school-aged children on the autism spectrum
(171 parents of girls and 163 parents of boys) that was distributed via social media. The surveys provided insights
regarding the characteristics of boys and girls (as perceived by parents) as well as some demographic information. There
were very few differences reported regarding communication and social strengths and difficulties of boys and girls with
autism. No differences were reported in the number of boys and girls on the autism spectrum with special interests or
repetitive behaviours; however, significant differences were found in the types of special interests with boys and girls
showing generally interests along traditional gender lines. Qualitative analysis of open comments indicated that some
parents of girls on the autism spectrum described their daughter as trying to hide or mask her difficulties more but no
parents of boys on the spectrum described this phenomenon.

Keywords
autism spectrum disorder, gender, school-aged children

Introduction
More boys than girls are diagnosed with autism spectrum and Sikora, 2009; Kirkovski et al., 2013). Others have sug-
disorder (ASD). The ratio is commonly reported as 4:1 for gested similarities in core autism symptomatology, but that
all children with ASD with even fewer girls diagnosed differences exist in associated features such as internalis-
with autism among children without intellectual disability ing difficulties of anxiety and depression (Hartley and
(Baoi, 2014; Shefcyk, 2015) and higher rates of girls Sikora, 2009; Holtmann et al., 2007). More recently,
among children who also have an intellectual disability Mandy et al. (2012) reported no differences between
(Fombonne, 2003, 2009). While these ratios appear to be girls and boys on communication or social interaction
consistent across much of the research in the area, it is domains according to parental report, but that there were
unclear why this is the case. Ascertainment bias may differences between groups on repetitive behaviours.
account for some of the difference in rates of diagnosis Importantly, parents reported higher levels of internalising
(Halladay et al., 2015) although the possibility that bio- difficulties in girls, but these were not noted by teachers.
logical differences underpin the diagnosis ratios still needs A recent review of 22 studies found few differences in
to be considered (Werling and Geschwind, 2013). One the core deficits of autism between girls and boys (Van
suggestion is that boys and girls may differ in their presen-
tation, making the detection of girls more difficult 1Autism Spectrum Australia, Australia
(Dworzynski et al., 2012; Kirkovski et al., 2013; Lai et al., 2The Children’s Hospital at Westmead, Australia
2012). Research, however, has not found a consistent pat- 3Griffith University, Australia

tern of differences.
Corresponding author:
Studies have suggested that boys may show higher lev- Rebecca Sutherland, Autism Spectrum Australia, PO Box 87,
els of repetitive behaviour and restricted interests while Forestville, NSW 2087, Australia.
girls showed greater communication difficulties (Hartley Email: rsutherland@autismspectrum.org.au
2 Autism

Wijngaarden-Cremers et al., 2014), but on the whole, skills while further demonstrating the differences in
lower rates of restricted interests and repetitive behaviours interests in boys and girls on the spectrum.
were noted in girls and higher rates of stereotyped move-
ments were seen in boys. Recent studies have suggested
Methods
that girls may have more appropriate interests or interests
less typical of ASD (e.g. Hiller et al., 2014), but reviews Ethical approval for this survey-based study was provided
suggest that this is not a consistent finding (Harrop et al., through Autism Spectrum Australia (Aspect) Research
2015). Harrop et al. (2015) found similar rates of restricted Approvals Committee. Participants were provided with
repetitive behaviours in young girls and boys with a trend information about the survey and assured that it was anon-
towards higher frequency of these behaviours in boys. ymous. Participants gave informed consent and were able
Recent Australian research (Hiller et al., 2014) included to withdraw from the study at any time.
a sample of 69 high-functioning girls and boys with an
ASD diagnosis and looked not only at whether boys and
Participants
girls met various diagnostic criteria but also at the way
diagnosticians had reached their conclusions regarding A total of 198 parents of girls with ASD and 186 parents of
each criterion. The behavioural characteristics were boys with ASD gave informed consent to participate in the
extracted from reports and indicated that there were no dif- study. Following review of the data, responses that pro-
ferences between boys and girls in the broad social crite- vided no details of their child’s characteristics (i.e. behav-
ria, but that there were differences in the way boys and ioural or personal preferences) were excluded from further
girls came to meet each of the criterion. As with other analysis. This exclusion included respondents who con-
research, girls also presented with fewer and different firmed their child was over school age (n = 1 girl, n = 5
restricted interests than boys, including lower interest in boys). Notably, all the deleted responses either identified
screen time and more interests that were categorised as their child as not having a diagnosis of autism or did not
‘random’ interests. The impact of these differences means respond to that question.
that girls may be less likely to receive a diagnosis, even The sample of parents was self-selecting. Parents of
when similar levels of autistic traits are observed girls and boys on the autism spectrum were alerted to the
(Dworzynski et al., 2012). survey through the (name withheld for review) Facebook
The need for an understanding of the subtle differences page, which currently has approximately 6600 ‘likes’ from
between males and females was highlighted by Baldwin parents and professionals. No parents were approached
and Costley (2015) in their study of women on the autism directly by the researchers.
spectrum. They noted that research may not indicate a All participants were required to answer a question
unique profile of females on the autism spectrum but about whether their child had a diagnosis of ASD and, if
rather a range of subtle and subjective ways in which so, who made the diagnosis (e.g. psychiatrist, psychologist
autism is manifested in girls and women, including differ- and multidisciplinary team) and the date it was made.
ences in their perception of the social difficulties, mental Given the scope and nature of the survey, it was not pos-
health and preferences for company. Baldwin and Costley sible to independently verify diagnosis and these questions
(2015) also noted higher levels of mental health concerns were asked as a way of confirming that the children
among women than men on the spectrum that, while not described by the survey responses had autism. Four par-
diagnostic features of autism, frequently co-occur among ents of girls stated that their daughter did not have a diag-
those with ASD, adding to the difficulties at home, school nosis; one boy did not have a diagnosis. These participants
and work. also answered no further questions, leaving 171 parents of
The issues addressed by this study involved gaining an girls and 163 parents of boys as possible respondents to the
understanding of the real-life characteristics of school- remaining questions.
aged girls and boys with autism, including both diagnos- The remaining questions elicited whether the parent
tic criteria and associated features, as described by their had observed a behaviour or interest, with a response only
parents. It sought to add to the evidence base by includ- appropriate in the case where this was true. The highest
ing large sample sizes and a non-clinically referred popu- level of response to an observation was 267 agreeing that
lation, two features that contrast with much of the their child did not get social cues while the lowest response
previous research in the area. In addition, the use of par- regarded an interest in science fiction, with only 27 indi-
ent-derived descriptions (rather than clinical or diagnos- cating interest.
tic descriptors) of the strengths and difficulties of girls
and boys on the spectrum adds to the possibility of detect-
Materials
ing subtle differences that may exist between the gen-
ders. It was hypothesised that this study would highlight This study employed a questionnaire methodology to sur-
these subtle differences in social and communication vey the parents of children with ASD regarding their
Sutherland et al. 3

sons’ and daughters’ strengths and difficulties in commu- Open comments regarding differences between home
nication and social skills, as well as their special inter- and school and the types of repetitive behaviours the chil-
ests, repetitive behaviours and sensory needs. The survey dren showed underwent thematic analysis as per Braun
was developed from an earlier pilot study of 282 parents and Clarke (2006). Responses were analysed, coded and
of girls only who provided responses to open questions grouped according to the broad themes that emerged. The
about these areas. It is not known whether any of these open comments provide both qualitative information (e.g.
parents went on to complete the survey in this study. observed differences between home and school) and some
These open responses underwent thematic analysis, and quantitative data (e.g. the proportion of girls and boys
common themes were identified. These themes formed engaging in different types of repetitive behaviours).
the basis of multiple-choice questions for this survey
about communication and social strengths and difficul-
ties. Thus, the questions, while grouped in terms of the
Results
main diagnostic features of ASD (i.e. social interaction, Demographic information
communication, behaviours, interests and sensory differ-
ences), were not based on diagnostic criteria; rather, they Age. All participants described children between 5 and
were based on the real-life descriptions of skills and dif- 18 years of age (typical school-age range for students in
ficulties made by parents of girls on the spectrum. This Australia) on the autism spectrum. A total of 78% of boys
survey included some demographic information about and 76.6% of girls were of primary school age (5–12 years;
the children, including their age, whether they had a diag- see Table 1 for gender by age distribution).
nosis, who made this diagnosis and the type of school
they attended. Questions about communication and social Diagnostic status. Parents were asked whether their child
strengths and difficulties, special interests and sensory had a diagnosis of ASD and who had made the diagnosis
differences were multiple choice as described above, (e.g. psychiatrist, paediatrician, multidisciplinary team or
with an option of adding comments or other details psychologist). The most common source of a diagnosis,
provided. Questions about differences in skills across for both boys and girls, was by a paediatrician (46.60%
contexts and repetitive behaviours were open-ended and 42.10%, respectively). While diagnosis by a psychia-
questions. The survey software used was SurveyMonkey®. trist was far less common, a greater proportion of girls than
Parents of girls were invited first to participate in the sur- boys did received their diagnosis from a psychiatrist
vey with parents of boys being recruited to a separate roll (8.20% of girls compared to 1.80% of boys).
out of the survey. See Appendix 1 for survey questions,
available online School type. Similar numbers of girls and boys were
attending schools for specific purposes (sometimes known
as special schools; around 10% of girls and boys). This
Analysis figure may provide a proxy measure of intellectual func-
The data were analysed using a mix of methods, including tioning and may suggest that the groups could be similar in
statistical analysis, along with thematic analysis of open terms of cognitive skills. Similar numbers of girls and
comments. boys were in a mainstream education setting with support
(such as a teacher aide), with 86.5% of boys and 83.0% of
Statistical analysis. Analyses were performed using the girls. However, of these more girls were in classes without
Statistical Package for the Social Sciences (SPSS) v22. support than were boys (36.3% of girls compared to a
Frequencies of all variables were initially described and 25.8% of boys). Relatively high numbers of girls and boys
compared across gender groups providing an initial were home-schooled, with a higher proportion of girls in
profile of responses regarding gender characteristics of this group (7% of girls and 2.5% of boys).
communication, areas of social difficulty and interests.
Statistical analysis of these differences was not undertaken Communication, social and behaviour
individually, due to issues associated with high numbers
characteristics
of multiple comparisons. A binary logistic regression was
undertaken to simultaneously explore the possible effect The percentage of responses to each of the descriptors on
of reported actions of boys and girls and suggested predic- the survey is presented in Table 2.
tive factors for gender. All items regarding communication
and preferences were included, as no assumptions were
made about relative importance. All items were entered at Communication strengths and difficulties
the initial step. From this assessment, odds ratios (ORs) There were few marked differences between girls and boys
and 95% confidence intervals (CIs) were calculated, with regarding communication. Boys and girls were similarly
values of p < 0.05 considered significant. likely to be described as talkative or chatty, having good
4 Autism

Table 1. Age distribution, school placement and diagnosis by gender.

Boys (%) Girls (%) Total (%)

Age (years)
5–6 19.0 19.3 19.2
7–8 17.2 24.6 21.0
9–10 22.1 17.5 19.8
11–12 19.6 15.2 17.4
13–14 11.0 10.5 10.8
15–16 8.6 8.8 8.7
17–18 2.5 4.1 3.3
Respondents, n 163 171 334
School type
Mainstream without support 25.8 36.3 31.1
Mainstream with support 42.9 40.9 41.9
Support class in mainstream school 17.8 5.8 11.7
Special school (e.g. SSP, autism school) 10.4 9.4 9.9
Home school/distance education 2.5 7.0 4.8
Finished formal schooling 0.6 0.6 0.6
Respondents, n 163 171 334
Diagnosis
By a paediatrician 46.6 42.1 44.3
By a multidisciplinary assessment team 30.1 25.7 27.8
By a psychologist 21.5 24.0 22.8
By a psychiatrist 1.8 8.2 5.1
Respondents, n 163 171 334

SSP: School for Specific Purposes.

vocabulary or language skills and communicating well if special interests, with boys and girls generally showing
comfortable with the topic. Similar numbers of boys and interests along traditional gender lines (see Figure 1).
girls were described as non-verbal. More boys than girls Open comments about ‘other’ interests also showed differ-
were described as ‘communicates well if interested in the ences between boys and girls. Thirteen parents of boys
topic’ (69.9% boys and 50.3% girls). Similarly, more boys indicated a special interest in Lego® compared with only
also were described as ‘prefers to talk about his or her own two girls; seven parents of girls indicated an interest in
interests’ (68.1% boys and 50.3% girls). fantasy but no boys showed this interest; more boys were
interested in sport than girls. In contrast, small but similar
numbers of both boys (n = 5) and girls (n = 6) had an inter-
Social interaction strengths and difficulties est in television and movie characters such as Octonauts,
As with communication, there were many similarities in Frozen and Teenage Mutant Ninja Turtles.
social strengths and difficulties between the boys and girls
described in this study. About half of boys and girls were
Repetitive behaviours
described as having one or two special friends, and a similar
number were described as kind to younger children. Participants were asked whether their child displayed
Approximately 60% of both groups felt their child was bet- repetitive behaviours and, if present, to describe them in an
ter in one-to-one situations and about 10% of boys and girls open response. Similar proportions were reported by gen-
were described as popular. The overwhelming majority der with parents indicating that 67.6% of boys and 72.7%
(>76%) of boys and girls were described as having difficul- girls displayed repetitive behaviours. The open responses
ties understanding social cues and about 70% of both groups were then analysed to determine the broad type of behav-
had difficulties in group settings. Around 60% of both boys iour. The types of behaviours described included motor
and girls worried excessively about social situations. stereotypies (such as rocking, flapping, flicking), behav-
iours indicating a need for sameness (such as sticking to
routines, having meals at the same time every day), sen-
Special interests
sory-based behaviours (sniffing, licking, touching), repeti-
Both boys and girls were described as having special inter- tive verbal behaviour (such as repetitive questioning and
ests; however, clear differences were found in the types of vocalisations), self-injurious behaviours (skin picking),
 Sutherland et al.

Table 2. Distribution of survey descriptors across gender groups (boys, n = 163; girls, n = 171).

Trait % Trait %

Boys Girls Boys Girls

Communication Talkative/chatty 46.6 48.5 Social difficulties Has difficulty in group settings 71.2 65.5
skills Good vocabulary/good language skills 46.0 39.8 Does not understand friendship rules 60.1 63.2
To the point 30.7 33.3 Worries excessively about social situations 54.6 59.1
Likes talking with others 23.9 21.1 Needs to have her own way/bossy 44.8 51.5
Communicates well if interested in the topic 69.9 50.3 Has trouble maintaining friendships 53.4 49.7
Communicates well if comfortable with the people 58.3 55.6 Only interacts with older or younger children 14.7 18.7
Effective non-verbal 14.7 15.2 Interests Using technology 85.9 64.9
Communication Non-verbal or limited verbal communication 12.9 9.9 Singing/dancing/music 22.7 52.0
difficulties Difficulty understanding instructions/information 60.7 54.4 Animals 30.1 49.1
One-way conversations 35.6 36.8 Reading/books 32.5 45.0
Prefers to talk about his or her own interests 68.1 50.3 Art/craft 15.3 43.9
Limited interest in talking to other people 30.1 24.0 Collectibles 22.7 33.9
Only talks with a limited number of people 22.7 25.7 Maths and/or science 26.4 12.9
Blunt or too honest 41.7 49.7 Dinosaurs 14.7 8.2
Literal understanding 66.9 62.6 Science fiction 11.0 5.3
Poor eye contact 49.1 46.2 Transport 27.0 2.3
Social skills Has one or two special friends 42.9 48.5 Sensory Sensitivity to sound 76.1 73.1
Kind to younger children 49.1 52.6 Sensitivity to light/visual information 36.2 28.7
Kind to children who are different/vulnerable 29.4 38.0 Sensitivity to movement 19.0 16.4
Enjoys being with people 25.8 26.3 Sensitivity to smells 48.5 48.0
Gets on well with older/younger people 45.4 40.9 Sensitivity to taste (e.g. bland diet) 35.6 49.7
Popular 8.6 8.8 Sensitivity to texture/touch 58.3 48.5
Keen to interact with others 31.9 26.9 Seeks sound 11.0 12.3
Loyal 26.4 32.2 Seeks visual sensations 23.3 19.3
Better 1:1 than in groups 62.0 57.9 Seeks movement 39.3 35.1
Does not get social cues 84.0 76.0 Seeks strong tastes/smells 10.4 9.9
Seeks textures (e.g. strokes hair, fabric) 35.0 36.3
5
6 Autism

Figure 1. Comparison of boys’ and girls’ interests.

movement (running, pacing) and obsessive behaviours stronger response by boys (e.g. boys were more likely to
(such as cleaning, colour coding). There appeared to be be interested in technology, OR = 4.193).
somewhat more behaviours that could be classified as From the logistic regression, it is clear that the main
movement and obsessive behaviours in boys, but further factors differentiating girls and boys are not so much in
analysis of this was beyond the scope of this article. their communication or social behaviours or even their
sensory reactions, but in their interests. The only factors
that are identified in behaviours as significant are about
Sensory communication – ‘communicates well if interested in the
Participants were given multiple-choice options regarding topic’ (OR = 3.775, p = 0.009), where parents of boys
sensory sensitivities and sensory seeking behaviours across selected this much more often than parents of girls (69.9%
the senses (vision, hearing, smell, taste, touch, propriocep- to 50.3% agreement from males and females), and ‘prefers
tion and vestibular senses). Boys and girls were reported to to talk about his or her own interests’ (OR = 2.637,
display similarities in most areas. Parents reported more p = 0.037). On the other hand, girls indicated slightly
sensory sensitivity (71%) than seeking behaviours (29%) in greater difficulties understanding friendship rules
both boys and girls. Auditory sensitivity was the most com- (OR = 0.409, p = 0.027). In looking at sensory aspects,
mon observation for both boys and girls. again only one area demonstrated a significant gender dif-
ference (49.7% of girls compared to 35.6% of boys)
regarding sensitivity to taste (OR = 0.263, p = 0.002).
Assessing significant gender differences – In contrast, several significant areas differentiated boys
logistic regression and girls in the area of special interests. Boys were profiled as
To provide a predictive picture of the gender differences in being significantly more interested in technology (OR = 4.193,
behaviours and preferences, a logistic regression was p = 0.004), dinosaurs (OR = 4.353, p = 0.018) and transport
undertaken, with all dichotomous variables documenting (OR = 46.962, p < 0.000), while girls showed a far greater
parents’ assessments of their children’s communication interest in art (OR = 0.257, p = 0.002), books (OR = 0.400,
and social behaviours, their interests and sensory responses p = 0.034) and singing/music (OR = 0.192, p < 0.000).
entered simultaneously. Table 3 documents the findings
from the logistic regression, presenting details for all sig-
Differences across settings
nificant variables from the analysis. In this assessment,
ORs that are less than 1 indicate where girls show a greater Participants were asked whether, to their knowledge, their
level of response (e.g. girls were more likely to enjoy sing- child’s communication and social skills were different in
ing and music, OR = 0.192), and those above 1 indicate a the home and school setting. About three quarters of boys
Sutherland et al. 7

Table 3. Results of logistic regression showing significant predictor variables.

Predictor B SE Wald df Sig. OR 95% CI for OR

Lower Upper
C: communicates well if interested 1.328 0.508 6.840 1 0.009 3.775 1.395 10.215
C: own interests 0.970 0.466 4.334 1 0.037 2.637 1.058 6.572
Soc: does not understand friendships rules −0.894 0.405 4.883 1 0.027 0.409 0.185 0.904
I: art −1.360 0.437 9.675 1 0.002 0.257 0.109 0.605
I: technology 1.433 0.502 8.141 1 0.004 4.193 1.566 11.223
I: books(1) −0.917 0.432 4.505 1 0.034 0.400 0.172 0.932
I: singing/music −1.650 0.411 16.149 1 0.000 0.192 0.086 0.429
I: dinosaur 1.471 0.622 5.598 1 0.018 4.353 1.287 14.719
I: transport 3.849 0.835 21.254 1 0.000 46.962 9.142 241.250
Sens: sensitivity to taste −1.337 0.442 9.141 1 0.002 0.263 0.110 0.625

Nagelkerke R2 = 0.613; Hosmer Lemeshow test significance = 0.002; classification accuracy = 81.7%.
SE: standard error; df: degree of freedom; OR: odds ratio; CI: confidence interval; C: communication, Soc: social, I: interests, Sens: sensory.

and girls were described as displaying similar communica- hides under tables, refusing to talk or make eye contact.
tion characteristics across home and school settings and (Parent of a girl)
almost 85% of boys and girls were described as displaying
similar social skills across settings. Participants provided Tries hard at school, holds everything together. Two parents of
open comments describing the observed differences and boys described their sons as more rule abiding at school
qualitative analysis of these comments indicated common and as putting energy into friendships at school and one
themes, as follows. parent of a boy indicated that he tried to fit in but did so by
being silly and drawing attention to himself. However, nei-
More comfortable at home. Of those parents who indicated ther these comments nor any others from boys’ parents
that their child’s communication or social skills were dif- explicitly indicated an awareness of difference or an effort
ferent at home and school, the majority indicated that to manage or behave differently at school. In contrast, this
their child was more comfortable at home. This was was a relatively common theme for parents of girls. Thir-
described in several ways including that their child was teen parents of girls indicated that their daughters were
more anxious outside of home or that they were more con- aware of their differences to their peers and that they
fident at home. There was no difference in the numbers of explicitly tried to manage their difficulties at school, often
parents of boys and girls who reported their children felt becoming more quiet or cautious in their communication.
anxious or experienced anxiety. Comments included Comments on this theme included

He is more willing to explain himself at home where he is She tries hard at school holds everything together then
comfortable. (Parent of a boy) meltdowns when gets home. Tries hard not to let people see
her difficulties. (Parent of a girl)
More comfortable at home. (Parent of a girl)
She is less verbal at school and outside home but more careful of
Much more confident at home, sometimes quite anxious what she says – applies filter more effectively. (Parent of a girl)
about communicating elsewhere in case she gets it wrong.
(Parent of a girl) Attempts to mimic very hard at school and would contain
herself more than she does at home. Doesn’t want attention
Quiet and reserved at school. Parents of many students indi- drawn to her in class so goes into her own superhero world.
cated that they were far more quiet and reserved at school, (Parent of a girl)
with some parents saying that their child rarely spoke at
school. Many parents made comments indicating that their Outside she is more worried about not being normal and saying
child was more talkative at home. Comments in this theme things that make her seem different or strange. (Parent of a girl)
included
Discussion
Much less willing to participate in conversations at school or
ask for help when needed. (Parent of a girl) This study aimed to look at differences between girls
and boys on the autism spectrum using parent-derived,
At school, she is much more anxious, so all of these issues are non-clinical descriptors of skills with a large, non-
magnified. Where at home she is loud and bossy, at school she clinically referred sample of school-aged students. To
8 Autism

our knowledge, this is one of the largest studies to look behaviours such as stacking blocks (Mandy et al., 2012).
specifically at the similarities in communication and In contrast, this study found broadly similar rates of repeti-
social strengths and difficulties of boys and girls, as tive behaviours in boys and girls but potentially with sub-
well as behaviours and interests. The outcomes of this tle differences in the types of those behaviours.
study may provide useful information to those involved Similarly, and in contrast to the literature, this study
in an autism specific diagnostic assessment, as well as found similar rates in the numbers of girls and boys who
the education and treatment of children on the spectrum. were reported to have special interests. There were, how-
In this study, few differences were found in communica- ever, significant differences in the topics of interest to girls
tion or social interaction skills or in the proportions of and boys. The topics were clearly divided along gender
boys and girls displaying strong interests or repetitive lines, with very small numbers of girls engaging in inter-
behaviours. However, highly significant gender differ- ests that are seen as typically related to the autism spec-
ences were found in the types of interests. In addition, trum such as dinosaurs or an interest in transport. Rather,
girls and boys varied on some key demographic varia- girls engaged in the types of interests typically expected of
bles as well as in parent-reported behaviour differences their gender (Gould and Ashton-Smith, 2011; Johnson
between settings that suggest a desire to mask their et al., 2004; Ruble et al., 2007), such as animals, reading,
difficulties. art and music, while boys were typically described as pre-
Differences in demographic variables, such as who ferring activities typical of boys of this age group (Rost
made the diagnosis and school placement, were not antici- and Hanses, 1994). It is possible that in some circum-
pated. The majority of both girls and boys in our study stances, such interests may be overlooked by diagnosti-
were diagnosed with autism by a paediatrician, a psychol- cians, because they do not conform to the typical male
ogist or a multidisciplinary team, which are likely reflec- presentation. Indeed, it has been suggested that diagnostic
tive of the access to these types of diagnostic assessments assessments may need to include examples that are typi-
in the community. Girls in our study were more likely to cally associated with traditional female interests (Frazier
have been diagnosed by a psychiatrist (n = 14) than boys et al., 2014; Solomon et al., 2012), a suggestion supported
(n = 3). Qualitative analysis of the responses and com- by this study. Greater focus on parent questionnaires and
ments by the parents of the 14 girls did not reveal any surveys used for diagnosis on the intensity of interests
apparent pattern of differences to other girls in the cohort (Gould and Ashton-Smith, 2011), rather than the types of
that might have led to the differences in specialists sought. special interests, may also aid in understanding the presen-
It is possible that parents initially sought contact with a tation of ASD in girls. In addition, knowledge about the
psychiatrist because of suspected mental health difficulties interests shown by girls may be of use in clinical and edu-
such as anxiety, or because they had difficulty obtaining a cation settings so that these interests can be incorporated
diagnosis through typical channels, although these were into learning and therapy programmes, as is routinely rec-
not specifically mentioned by the parents in the comments ommended for students on the spectrum.
on this survey. The other striking difference between the genders on
This study showed very few differences in parent’s this study was the way parents described the differences in
descriptions of boys and girls in terms of their communica- their children in the home and school setting. While simi-
tion and social strengths and difficulties. It had been lar numbers of girls and boys presented differently in the
expected that using a tool based on girls’ parents’ descrip- home and school settings, the parents of girls indicated
tions may have elicited some subtle differences between that there was a conscious effort by the girls to blend in or
boys and girls but this was not found. The responses did, to hide their difficulties. This observation may give some
however, provide a number of items of interest, including support to the research findings that females on the spec-
that only half of parents felt that eye contact was an issue trum may ‘mask’ or camouflage their difficulties (e.g. Lai
for their child and that 1 in 10 children on the spectrum et al., 2015; Lehnhardt et al., 2016; Tierney et al., 2016)
were seen as socially popular. The most prevalent social possibly making diagnosis less likely and support services
finding for both boys and girls was their difficulty ‘getting’ less forthcoming for some girls (Baldwin and Costley,
social cues and difficulties in groups. 2015). While there are anecdotal reports of this phenome-
Under Diagnostic and Statistical Manual of Mental non, there are limited data. Hiller et al. (2015) found that
Disorders (5th ed.; DSM-5) diagnostic criteria, all those teachers were less concerned with girls than boys with
diagnosed with autism must demonstrate significant spe- ASD and that the girls tended to mimic elements of the
cial interests, repetitive behaviours and/or sensory issues social environment in order to fit in. They suggested that
(American Psychiatric Association (APA), 2013). Previous the home and school presentations of girls on the spectrum
research and reviews (e.g. Frazier et al., 2014; Shefcyk, are likely to be starkly different which, in turn, may impact
2015; Van Wijngaarden-Cremers et al., 2014) have found on referral and detection rates for girls. This study did not
lower levels of repetitive behaviour in girls and that girls ask parents specifically about whether their child masked
are less likely than boys to engage in repetitive stereotyped or deliberately altered their presentation but simply
Sutherland et al. 9

whether they showed different skills in different settings. social and communication domains but that interests tend
This may be an important topic to explore more fully in the to fall along traditional gender lines (Gould and Ashton-
school-aged population. Smith, 2011). Similarly, support is added to previous find-
ings that girls may attempt to mask their difficulties at
school. For teachers and diagnosticians, the findings may
Strengths, limitations and areas for further
provide insight into why some girls with autism appear to
research present with fewer difficulties in the school environment
The greatest strength of this study is the sample size of compared with those seen at home. The findings also sug-
almost 350 participants and, more particularly, the inclu- gest that all those interacting with girls on the spectrum
sion of a comparable ratio of boys to girls with a range of need to be aware of and support potential underlying anxi-
abilities as reflected in their types of schooling. The use of ety and other mental health difficulties that may be masked
a non-clinical survey to elicit parent observations is a rela- or camouflaged at school and in the community. Further
tively novel approach and complements the body of research regarding access to diagnostic services, educa-
research that has used clinical and diagnostic tools to tional placement and access to support and as well as fur-
assess gender differences. ther information about masking behaviour of girls is
A major limitation of this study is that no comparison warranted.
group of typically developing boys and girls was included.
In addition, the inclusion of a similar number of girls and Acknowledgements
boys is unusual in research about ASD and does not reflect The authors wish to acknowledge the team members of Positive
the known accepted ratio of boys to girls with ASD, risk- Partnerships. This study was undertaken through the Positive
ing ascertainment bias. The reasons for the high number of Partnerships programme. Positive Partnerships is funded by the
responses from parents of girls compared with parents of Australian Government Department of Education and Training
boys are unknown. It is possible that the interest in the through the Helping Children with Autism package, to support
publically available survey from parents of girls reflects school-aged students on the autism spectrum. The views
expressed within this article do not necessarily represent the
the relatively limited level of information about girls on
views of the Australian Government or the Australian
the autism spectrum and the desire of parents to obtain Government Department of Education and Training. The authors
this. The survey was developed from a pilot study that also wish to acknowledge the families who generously provided
involved only parents of girls, with thematic analysis used their time and insight about their daughters and sons.
to develop their responses to open questions into the mul-
tiple-choice questions that were applied to both boys and Funding
girls. The use of a non-standard survey based on the obser-
The author(s) received no financial support for the research,
vations of parents of girls is a significant departure from authorship and/or publication of this article.
the tools commonly used in research about this topic and is
a limitation of the study. The survey questions, while
References
grouped in terms of the main diagnostic features of ASD
(i.e. social interaction, communication, behaviours, inter- American Psychiatric Association (APA) (2013) Diagnostic
and Statistical Manual of Mental Disorders (DSM-5).
ests and sensory differences), were not based on diagnostic
Washington, DC: APA.
criteria, nor could they form part of a diagnostic assess- Baldwin S and Costley D (2015) The experiences and needs
ment as has been used in some previous work comparing of female adults with high-functioning autism spectrum
boys and girls on the spectrum (e.g. Hiller et al., 2014). disorder. Autism. Epub ahead of print 25 June. DOI:
Rather, the survey incorporated real-life situations and 10.1177/1362361315590805.
descriptors that were relevant to parents and children. In Baoi J (2014) Prevalence of autism spectrum disorder among
addition, the use of open-ended questions allowed parents children aged 8 years. Morbidity and Mortality Weekly
to provide addition information and to clarify their Report (MMWR) 63: 1–21.
responses. It is interesting to note that few studies have Braun V and Clarke V (2006) Using thematic analysis in
engaged in research or used materials that are specifically psychology. Qualitative Research in Psychology 3(2):
designed for girls; indeed, the autism community and 77–101.
Dworzynski K, Ronald A, Bolton P, et al. (2012) How different
research has suggested that there is a need for question-
are girls and boys above and below the diagnostic thresh-
naires, descriptors and diagnostic tools that address the old for autism spectrum disorders? Journal of the American
characteristics of girls specifically and explicitly (e.g. Academy of Child & Adolescent Psychiatry 51(8): 788–797.
Baldwin and Costley, 2015; Shefcyk, 2015). Fombonne E (2003) Epidemiological surveys of autism and other
This study provides information about parent perceived pervasive developmental disorders: an update. Journal of
similarities and differences in girls and boys in autism Autism and Developmental Disorders 33(4): 365–382.
characteristics and associated features. It lends support to Fombonne E (2009) Epidemiology of pervasive developmental
earlier findings that girls and boys present similarly in disorders. Pediatric Research 65(6): 591–598.
10 Autism

Frazier TW, Georgiades S, Bishop SL, et al. (2014) Behavioral and Lai M-C, Lombardo MV, Auyeung B, et al. (2015) Sex/gender
cognitive characteristics of females and males with autism differences and autism: setting the scene for future research.
in the Simons Simplex Collection. Journal of the American Journal of the American Academy of Child & Adolescent
Academy of Child & Adolescent Psychiatry 53(3): 329–340. Psychiatry 54(1): 11–24.
Gould J and Ashton-Smith J (2011) Missed diagnosis or misdi- Lai M-C, Lombardo MV, Ruigrok AN, et al. (2012) Cognition in
agnosis? Girls and women on the autism spectrum. Good males and females with autism: similarities and differences.
Autism Practice (GAP) 12(1): 34–41. PLoS ONE 7(10): e47198.
Halladay AK, Bishop S, Constantino JN, et al. (2015) Sex and Lehnhardt F-G, Falter CM, Gawronski A, et al. (2016) Sex-
gender differences in autism spectrum disorder: summa- related cognitive profile in autism spectrum disorders
rizing evidence gaps and identifying emerging areas of diagnosed late in life: implications for the female autistic
priority. Molecular Autism 6(1): 36. phenotype. Journal of Autism and Developmental Disorders
Harrop C, Gulsrud A and Kasari C (2015) Does gender moderate 46(1): 139–154.
core deficits in ASD? An investigation into restricted and Mandy W, Chilvers R, Chowdhury U, et al. (2012) Sex differ-
repetitive behaviors in girls and boys with ASD. Journal of ences in autism spectrum disorder: evidence from a large
Autism and Developmental Disorders 45: 3644–3655. sample of children and adolescents. Journal of Autism and
Hartley SL and Sikora DM (2009) Sex differences in autism Developmental Disorders 42(7): 1304–1313.
spectrum disorder: an examination of developmental func- Rost DH and Hanses P (1994) The possession and use of toys in
tioning, autistic symptoms, and coexisting behavior prob- elementary-school boys and girls: does giftedness make a
lems in toddlers. Journal of Autism and Developmental difference? Educational Psychology 14(2): 181–194.
Disorders 39(12): 1715–1722. Ruble DN, Martin CL and Berenbaum SA (2007) Gender devel-
Hiller RM, Young RL and Weber N (2014) Sex differences in opment. Handbook of Child Psychology III: 14.
autism spectrum disorder based on DSM-5 criteria: evidence Shefcyk A (2015) Count us in: addressing gender disparities in
from clinician and teacher reporting. Journal of Abnormal autism research. Autism 19(2): 131–132.
Child Psychology 42(8): 1381–1393. Solomon M, Miller M, Taylor SL, et al. (2012) Autism symp-
Hiller RM, Young RL and Weber N (2015) Sex differences in toms and internalizing psychopathology in girls and boys
pre-diagnosis concerns for children later diagnosed with with autism spectrum disorders. Journal of Autism and
autism spectrum disorder. Autism. Epub ahead of print 25 Developmental Disorders 42(1): 48–59.
February. DOI: 10.1177/1362361314568899. Tierney S, Burns J and Kilbey E (2016) Looking behind the
Holtmann M, Bölte S and Poustka F (2007) Autism spectrum mask: social coping strategies of girls on the autistic
disorders: sex differences in autistic behaviour domains and spectrum. Research in Autism Spectrum Disorders 23:
coexisting psychopathology. Developmental Medicine & 73–83.
Child Neurology 49(5): 361–366. Van Wijngaarden-Cremers PJ, van Eeten E, Groen WB, et al.
Johnson KE, Alexander JM, Spencer S, et al. (2004) Factors asso- (2014) Gender and age differences in the core triad of impair-
ciated with the early emergence of intense interests within ments in autism spectrum disorders: a systematic review
conceptual domains. Cognitive Development 19(3): 325–343. and meta-analysis. Journal of Autism and Developmental
Kirkovski M, Enticott PG and Fitzgerald PB (2013) A review Disorders 44(3): 627–635.
of the role of female gender in autism spectrum disorders. Werling DM and Geschwind DH (2013) Sex differences in
Journal of Autism and Developmental Disorders 43(11): autism spectrum disorders. Current Opinion in Neurology
2584–2603. 26(2): 146–153.