Opinion Conversations on the science of autism research.

Studies struggle to pin down gender differences in autism

by  /  14 October 2014


After decades of describing autism as a disorder of boys, scientists have only begun exploring how it affects girls and women in the past few years. For example, there is some consensus now that females may need a bigger genetic hit to develop autism and that the current diagnostic tests may miss many girls with the disorder.

Given all that, it comes as a surprise that two new studies, both by reputable research teams, report no detectable differences between boys and girls with autism.

The first, led by Catherine Lord, assessed various aspects of development in 234 boys and 54 girls diagnosed with autism. The study design, described 5 September in the Journal of Autism and Developmental Disorders, was rigorous: The researchers used multiple standard measures of language, social and communication skills and autism symptoms. They drew from a large community sample of more than 14,000 children and narrowed the numbers down to a group of preschoolers who had received best-estimate diagnoses of autism. They then compared the final group with an age-matched group of typically developing children in similar numbers.

The researchers expected — based on the literature — to find that girls with autism have better verbal skills, poorer nonverbal skills and fewer repetitive behaviors than boys with the disorder. But in fact, they found no significant difference in any of these aspects. (They did find that girls in the control group have better language skills than boys in that group, which is consistent with what’s known about early language development.)

Hide and seek:

The second study, published 13 September in the same journal by Connie Kasari and her colleagues, is both smaller and more specific. It compared play behavior in 40 girls with autism and 40 boys matched for autism severity, with an average age of 40 months.  

Over a 20-minute recorded session, the researchers analyzed how the children played: Did they make a doll ‘walk’ to a dollhouse, for example, or pretend to be Superman, or did they treat all objects alike? They analyzed two play skills in particular: joint attention, which means attending to the same object as another person; and behavioral requesting, which means eliciting help in getting objects from someone else or responding to such requests. Trouble with these skills can signal autism.

There is some evidence that both typically developing girls and girls with autism have better play skills than boys do. But once again, Kasari and her colleagues didn’t find any significant differences in play type or complexity between the boys and girls in their sample.

These new studies are in good company. Many teams have tried and failed to find gender differences among children with autism of comparable intelligence. As Lord told me at a conference a couple of weeks ago, “I’ve been looking for the differences for years, but I haven’t been able to find them so far.”

But that doesn’t mean no differences exist.

The numbers of girls with autism in the studies — 54 and 40 — are bigger than usual for this type of research, but they’re still woefully small. More to the point, they are probably far too small to drown out the vast amount of noise from the many variables.

For example, Kasari’s study did an admirable job of trying to match the children on severity, but more boys were clustered at the lower-functioning end than girls.

There are also external confounds at play. Lord’s study drew children from two research centers — one in Michigan and the other in Florida. The Florida center gave children with autism higher scores on some aspects than the Michigan center did, hinting at the bias introduced by the choice of research site.

Intuitively, and from the many small pieces of anecdotal and empirical evidence, it’s clear that autism manifests variably in girls and boys. As Lord put it, “I don’t think we can say there are no differences; they’re just overshadowed by bigger things.” To get to those dissimilarities, we obviously need to do better at recruiting girls with autism into research studies.

13 responses to “Studies struggle to pin down gender differences in autism”

  1. Michelle says:

    The difference between boys and girls with autism, is the same as the difference between boys and girls with ADHD, or boys and girls with dyslexia, or any other type of challenge… SOCIETAL programming of those without such challenges to NOTICE them more in boys than in girls. If those researching this area want a real answer, then the study of neurotypical behaviors regarding how much a gender is expected to cope for itself rather than get help, of how external behaviors are interpreted, and so on, is really the only way for professionals to get a handle on why so many girls slip through the cracks and don’t get the supports they need.

  2. Julian from Surrey says:

    Non-ASD girls mature quicker than boys, but this difference recedes in early adulthood.
    Maybe ASD girls mature quicker than ASD boys, and with the developmental delay, this difference is observed later into adulthood, thus accounting for some differences we think we see?

  3. Louise Masters says:

    In my xperience Girls tend to be diagnosed later than boys so studies need to focus on children overt the age do 8

  4. Louise Masters says:

    In my experience girls tend to be diagnosed later than boys so studies need to focus on children overt the age do 8

  5. Aline says:

    Many studies don’t find significant gender differences because they choose boys and girls that are already diagnosed for the sample and hence, miss the girls that don’t fit so well in the “male criteria” and those are the girls that present the differences in a more significant way

  6. Sharon says:

    Girls are missed because they are quiet & just try to fit in, but never quite do. Clever, yet have next to zero social skills. I dont understand why in the UK SENCO are reluctant to diagnose Aspergers & go along with ADD or ADHD instead..

  7. Angelica Ronald says:

    The results from both these studies are really interesting.

    Regarding the point made about sample size, the numbers of girls with autism in both these studies — 54 and 40 – is considered ‘woefully small’ by the author Apoorva Mandavilli. Adequacy of sample size depends on what type of effect size the researchers are aiming to find. I would consider these samples sizes impressive for mean group differences analyses, and I expect power calculations would suggest these sample sizes were very acceptable except for detecting very small effect sizes.

    For those interested in this literature, my previous student and I have just published a similar study where we also found no sex differences between boys and girls with ASD, albeit on a smaller sample:

    Nguyen, C. & Ronald, A. (2014). How do girls with low(er) functioning autism compare to boys with autism and typically developing girls in their behaviour, cognition and psychopathology? Scandinavian Journal of Child and Adolescent Psychiatry and Psychology, 2, 55-65.

  8. MosaicofMinds says:

    Interesting study. How do they control for potential biases in diagnosis that, at any point on the spectrum, could lead fewer girls than boys to be diagnosed? I’m not surprised that the girls who *are* diagnosed would be similar to the boys diagnosed, but what about those who are ultimately assigned some other diagnosis–or never evaluated at all despite atypical behavior?

    I’m also curious, how did they decide which play behaviors to choose, and are these really likely to be the most sensitive measures of gender differences? One way you could tell would be if there were gender differences in these behaviors in the typically developing population. IIRC typically developing girls and boys do differ in amount of pretend play, but do they also differ in joint attention and requesting?

  9. maia says:

    as someone diagnosed late in life with autism, my hunch is, emotional skills are expected of women, as ‘social work’ ’emotional work’, and failure to do so is considered moral failing (see different expectations of helpfulness research), whereas in men failure is read as refusal is read as masculine or ‘being a cold fish’ ‘withdrawn’. The moral interpretation and condemnation blocks people thinking along another route. This is my suspicion based on my life,anyway

  10. Anonymous says:

    The thing that jumped out at me in this study was that the girls and boys for the study were selected for equivalence given existing criteria normed against boys. So if you select only girls who fit the male criteria for autism, you’re automatically going to miss any gender differences. You selected them out to begin with.

  11. EJ says:

    The only difference between autistic males and females is that females mask and mimic, a few men do too but it’s a female presentation to do so. Professor Tony Attwood believes, as have I for a long time, that the gender ratios are equal: http://evolutian.wix.com/planetautism#!autism-research-and-media/c16fm

Leave a Reply

Your email address will not be published. Required fields are marked *