Autism in Women: Why So Many Are Missed Until Adulthood

A woman in her late thirties sits down for an intake and says some version of this: I have been in therapy on and off since college. I have been told I have anxiety, then depression, then that I might have borderline traits. Everything helps a little and nothing quite fits. A friend sent me an article and I cannot stop thinking about it.

I hear a version of that account often enough that it no longer surprises me, and the reason it no longer surprises me is the subject of this post. Autism in women is not rarer than autism in men by anything like the margin the old numbers suggested. It is missed more, for reasons that have less to do with women and more to do with how the diagnosis was built.

The reference picture is male

The clinical description of autism comes from a specific lineage. Kanner’s 1943 case series was eight boys and three girls. Asperger’s was all boys. The decades of research that turned those case descriptions into criteria drew on samples that were overwhelmingly male, partly because boys were referred more and partly because the emerging criteria then selected for boys, which produced more male samples, and so on around the loop.

The result is a reference picture — restricted interests in trains or timetables, visible social withdrawal, obvious repetitive movement, meltdowns that other people can see — that describes a real and common presentation. It is not the only one. When the criteria were written from that picture, and the screening questionnaires were validated against it, and clinicians were trained on it, the system became very good at recognizing autism that looks like the picture and fairly poor at recognizing autism that does not.

None of this means the criteria are wrong. It means they have a shape, and the shape was fitted to one group.

What presents differently

The core features of autism are the same across sex: differences in social communication, restricted and repetitive patterns of behavior and interest, and sensory differences, present from early development. What differs, on average, is how they show on the surface. The patterns below are documented in the clinical literature, not a checklist to apply to yourself; the point is to explain why a competent observer can look directly at autism and not see it.

Social imitation and scripting. Many autistic girls learn early that social interaction has rules they cannot intuit, and they solve the problem by study. They watch, copy, rehearse, and build a repertoire of scripts for common situations. The result can be a woman who is fluent in small talk, holds eye contact because she trained herself to, and is described by everyone who knows her as warm — and who finds every social interaction effortful in a way the people around her would not believe. This is masking, and it is the mechanism underneath most of what follows.

Interests that look ordinary. Intensity of interest is a core feature; the content is not. An intense interest in horses, a particular novelist, a celebrity, psychology, or a social cause does not register as unusual, because it is socially legible. The feature is still there — the depth, the time, the regulating function the interest serves — but it does not look like the picture.

Distress turned inward. Where the reference picture has externalized distress that teachers notice, the presentation in girls more often involves internalized distress: anxiety, perfectionism, shutdown rather than meltdown, and the kind of exhaustion that gets labeled as something else — or, after enough years, arrives as autistic burnout. A girl who holds everything together at school and falls apart at home is doing the same work as the boy who falls apart at school, in a way that generates no referral.

Sensory differences framed as preference. Sensitivity to textures, sound, light, or crowds gets read as pickiness, shyness, or introversion when the rest of the picture does not raise the question of autism.

The cost of the delay

A missed diagnosis is not a neutral event. The features do not disappear because they are unnamed; they get named something else.

The most common substitutes, in my experience and in the literature, are anxiety, depression, and borderline personality disorder — each of which partly fits, which is what makes the substitution durable. Years of masking produce real anxiety. Years of being out of step produce real depression. Emotional intensity plus relationship difficulty plus self-injury, in a woman, is often read as a personality disorder, and autism and BPD overlap enough on the surface that the error is understandable in a fifty-minute intake.

The cost is treatment aimed at the wrong target. Exposure work for social anxiety assumes the social situation is being read accurately and the fear is disproportionate; it does not help, and can harm, when the difficulty is in the reading. Treatment for a personality disorder assumes a developmental history that is not there. Antidepressants treat the depression and leave the thing underneath it intact. People in this position often describe therapy as having helped with everything except the actual problem, which is a precise description of what has happened.

There is also a quieter cost: decades of a self-account built on the premise that something is wrong with one’s character, when the more accurate account is that one’s nervous system works differently and nobody said so.

What a competent adult evaluation does

The reason a thorough evaluation can see what a short appointment cannot is that it does not depend on the thing masking defeats — a single observed interaction.

A proper adult autism evaluation is built on a developmental history: what you were like at three, at seven, at twelve, gathered from your own memory and, where available, from a parent, sibling, or school records. It uses structured diagnostic measures designed for adults rather than screening questionnaires designed to flag children. It looks at sensory and routine-related features explicitly rather than inferring them from social presentation. And it is a differential evaluation: the question is not only “is this autism?” but “is this autism, ADHD, social anxiety, a trauma response, a personality disorder, or some combination?” — because for women specifically, the combination is common and the alternatives have been in play for years.

What it does not do is confirm a conclusion you arrive with. Some women who come in with a strong hypothesis receive an autism diagnosis; some receive a different and better-fitting answer; some receive both. The value is in the clarity either way, because what follows — the right treatment, workplace accommodations, a different relationship with your own history — depends on getting the answer right rather than on getting the answer you expected.

Getting a clear answer

If parts of this are familiar, the next step is not a quiz and not a decision; it is an evaluation that is built to see through what has been missed before. FRTC conducts autism evaluations for adults and adolescents from our Greenwood Village office in the south Denver metro, with a particular focus on late-identified presentations, as part of our broader psychological testing service. It begins with a free 15-minute consultation to confirm the evaluation fits the question you are bringing.

Dr. Rachel Grace, Psy.D., is a licensed clinical psychologist who leads assessment services at Front Range Treatment Center. This article is general education, not a diagnosis; autism can only be identified through a clinical evaluation.