How Girls Present With Autism Differently, and What’s Untested

The ratio and what moved it

Four or five boys for every girl. That was the male-to-female ratio reported for autism prevalence across most of the condition’s history, and it remains the figure most often quoted. Large population-based studies published before 2014 put it lower, between two and five boys to every girl, the older estimate sitting at the top of the newer range rather than outside it.

Meng-Chuan Lai and four colleagues, all affiliated with the Autism Research Centre at the University of Cambridge, set out to account for that movement. Their review, published in the Journal of the American Academy of Child and Adolescent Psychiatry in 2015, screened 1,906 papers on sex, gender or females and autism and examined 329 in detail.

The lower ratios came from a particular kind of study: nationwide, general-population cohorts rather than the clinic and school caseloads that supply most autism research, screened with instruments better able to detect subtle presentations. Earlier work had found the ratio widening at the higher-functioning end; some newer studies found no association with intellectual disability at all. The male bias is real, the authors conclude, and less pronounced than was previously believed.

Research samples ran further from the population than clinical caseloads did: roughly eight males for every female in studies measuring brain volume, about fifteen to one in functional imaging, so the neurobiology of autism has been assembled almost entirely from boys and men.

Autism is defined by behaviour, and the behaviours in that definition were drawn from a caseload overwhelmingly male since the first published case series. The standard diagnostic instruments, the Autism Diagnostic Interview and the Autism Diagnostic Observation Schedule, were built from items whose selection was likely shaped by that predominance. A meta-analysis found that females on average show fewer restricted and repetitive behaviours than males, with social and communication difficulty comparable, and the review treats that result as genuinely ambiguous: the difference may be real, or the instruments may be insensitive to how the behaviour looks in girls.

Studies that looked past the algorithm scores found that clinically diagnosed females meet the same criteria by a different route. They hold reciprocal conversation better and adjust their behaviour to the situation more readily, while their underlying difficulty understanding social situations matches that of males. Their friendship trouble takes another form again: initiation goes well and maintenance does not, and they are more often overlooked by peers than rejected by them.

What this means for you and your daughter

The review keeps apart two words usually run together. Sex, on the World Health Organization definition it adopts, refers to biological and physiological characteristics; gender to the socially constructed roles a society considers appropriate. The authors write “sex/gender” almost throughout, because gendered socialisation begins at birth and no human study can cleanly separate the two.

The explanation most popular among parents is camouflaging, meaning masking difficulty or compensating for it by imitating social behaviour and following learned scripts. The evidence available in 2014 was thin in a specific way. Among high-functioning autistic adults with comparable childhood symptoms and mentalising ability, women showed less evident autistic behaviour in interpersonal settings while reporting greater deviation from other women on autism-related traits than men reported from other men. Those are adults, and no quantitative measure of camouflaging yet existed; building one appears on the review’s list of work still to be done.

The familiar list of characteristics said to distinguish autistic girls, the restricted interests in people and animals rather than objects, the passivity read by a teacher as shyness, appears beneath a table heading that calls it anecdotal, drawn from clinical and autobiographical report. The authors are explicit that these descriptions have not been tested psychometrically.

On the diagnostic side the findings are harder to dispute. Age of diagnosis is on average later in females, and at similar levels of autistic features males are more readily diagnosed, while females require more concurrent behavioural or cognitive problems before a diagnosis is made. Interpretation bias at the point of referral and diagnostic substitution, where anxiety, depression or a personality disorder is recorded in place of autism, are both offered as candidate explanations.

Against that, the authors hold open a possibility parents rarely encounter: some autistic girls may need clinical services later than boys do, or need them less, and pathologising a child who is managing without functional impairment carries its own cost. A diagnosis rests on more than a trait level; distress and a need for services are part of what a clinician must establish.

The literature here stops at September 2014, and the review is a narrative synthesis rather than a meta-analysis, so it weighs arguments instead of pooling effects. Its point about who autism was described from has held: the male predominance is genuine at two to five boys for every girl, and part of the older figure came from under-recognition of girls.

Clinicians and autistic women have been writing that description of the female presentation down for years. If you recognise your daughter in it, you have recognised her in an account the field has not yet learned to test.