What Camouflaging Actually Is, and What Your CAT-Q Score Can and Cannot Tell You

A woman in her late thirties brings me a printout. She has taken an online camouflaging questionnaire, scored 138, and read on a forum that anything over 100 means autistic. She has spent three weeks reorganizing her understanding of her entire life around that number, and she wants to know what I think.

What I think is that the number is real and the reading of it is not, and that the conversation we are about to have will be more useful than either of those.

I see this often enough to have opinions about it. The instrument she used is good. The interpretation she found is a distortion that has been repeated until it acquired the texture of fact. And underneath both of those is a person who has correctly noticed that she is working hard at something other people appear to do for free, which is the observation worth keeping.

Camouflaging describes the strategies a person uses to adapt to or conceal autistic characteristics during social interaction.

The thing that makes this construct unusual in the autism literature is where it came from. Laura Hull and colleagues at University College London built it out of autistic adults' own qualitative accounts of what they were doing and what it cost. The categories were derived from the people performing it, who were in a considerably better position to describe it.

That origin matters, because a great deal of autism research measures what autistic people look like to non-autistic observers. This one measures what autistic people report doing.

The factor analysis produced three distinct strategies, and they are different activities with different costs.

Compensation is the active production of social behavior. Scripts written in advance for predictable situations. Phrases lifted from other people and kept for reuse. Conversations rehearsed before they happen, sometimes for hours, sometimes for days. Facial expressions studied and reproduced deliberately. A mental library of what a person is supposed to say when someone mentions a death or a new baby.

Compensation is the most invisible of the three, because it works. It produces behavior that reads as socially fluent, which means the person doing it gets no credit for the labor and no allowance for the cost. It is also the reason a great many autistic adults have been told, at assessment, that they present as too socially skilled to be autistic.

Masking is suppression. Holding eye contact that is physically uncomfortable to hold. Sitting on a stim that would regulate the nervous system if it were permitted. Controlling the face so it does not broadcast what is happening inside. Flattening a voice that would otherwise carry too much or too little. This is subtraction, where compensation is addition, and it runs continuously in a way compensation does not.

Assimilation is the felt experience of doing the other two. The sense of performing a role rather than participating in an event. The feeling of being in the room without being in the room. Exhaustion after social contact that does not match the apparent difficulty of the contact. A background awareness that the version of you present is a version you are maintaining.

Most people who camouflage lean heavily on one of these and lightly on the others. Knowing which one is carrying the load matters more clinically than knowing the total, because the interventions differ. Someone whose cost is concentrated in compensation needs the demand for spontaneous social production reduced. Someone whose cost is concentrated in masking needs environments where the stim is permitted. Someone whose cost is concentrated in assimilation is usually describing something closer to dissociation, and that is a different conversation.

The Camouflaging Autistic Traits Questionnaire is 25 items answered on a seven-point scale, producing a total between 25 and 175, with compensation, masking, and assimilation scored separately.

Hull and colleagues published the validation in 2019 in the Journal of Autism and Developmental Disorders, after administering it online to 354 autistic and 478 non-autistic adults. Exploratory factor analysis produced the three factors, confirmatory factor analysis showed good model fit, and measurement invariance analyses demonstrated that the factor structure held across gender and across diagnostic group. Internal consistency came in at 0.94 for the total scale, with test-retest reliability acceptable at 0.77. Convergent validity was established against measures of autistic traits, wellbeing, anxiety, and depression.

By the standards of self-report instruments in this field, that is solid work. A cross-cultural validation followed in 2024, and a short form has since been validated as well.

What it measures is camouflaging. That is the entirety of what it measures, and the authors were clear about it.

The printout goes wrong at this point, and I want to be precise about it, because the error has spread far enough that people now encounter it before they encounter the instrument.

The figure of 100 circulates as a validated cutoff. It appears on testing sites, in forum posts, in a considerable amount of secondary writing, and in at least one place that describes it as "the validated cutoff from Hull et al. 2019." The original validation paper does not establish a diagnostic cutoff. It reports factor structure, reliability, and convergent validity. It does not present the instrument as a screener for autism and does not offer a threshold above which a person should be considered autistic.

What the paper does report is group means, with autistic adults in the sample averaging around 124 and non-autistic adults around 97.

Those are the centers of two overlapping distributions, and the overlap is the part that matters. A substantial number of non-autistic participants scored above the autistic average, and a substantial number of autistic participants scored below the non-autistic one. Drawing a line at 100 and treating everything above it as a positive result only sorts scores into higher and lower, and sorting scores is all the instrument was built to do.

This is the same failure that produces many of the online quiz problems I write about. An instrument built to describe a dimension gets repurposed as a gate, a number gets attached to the gate, the number gets repeated, and eventually people are making identity decisions on the strength of a threshold that no researcher published.

There are several reasons, all ordinary, and none of them is evidence of anything wrong with the person.

Social anxiety produces large quantities of rehearsal, self-monitoring, post-event review, and advance planning for conversations that may not happen. The behaviors the CAT-Q asks about are the behaviors anxiety generates.

Growing up somewhere that reading the room was a safety requirement produces the same thing. A child who learned to scan a parent's face on entering a house has been running compensation and masking since before they could name it, and the machinery does not switch off because the house is gone.

Being visibly different in an environment that punishes difference produces it too, whether the difference is race, class, disability, queerness, an accent, religion, body size, or anything else that draws comment. Anyone who has been told to make themselves smaller has been doing the work this questionnaire asks about.

So does any occupation where managing your own presentation is part of the job.

Camouflaging is what a person does when the social environment costs more to be in than to perform in. Autistic people do a great deal of it because the environment is calibrated against them more or less continuously. They have no monopoly on the reason.

One more finding from the same group.

Hull and colleagues published separate work on gender differences in self-reported camouflaging, and autistic women reported camouflaging more than autistic men. That is one of several threads that help explain a pattern clinicians see often, where women and gender-diverse adults arrive for assessment in their thirties and forties after decades of accumulating diagnoses that did not fit.

The mechanism is not complicated. An assessment that relies on observing autistic presentation in the room will miss anyone who has spent thirty years becoming excellent at not presenting that way. The skill that got them through school and work is the same skill that gets them declined at assessment.

A high compensation score is a direct account of that skill. It is, in a literal sense, the receipt.

The CAT-Q answers a useful question, which is how much of your social functioning is being manufactured and by what mechanism.

That matters because camouflaging correlates with anxiety, with depression, and with poorer wellbeing, and because the bill arrives later and larger than people anticipate. A great deal of what gets called autistic burnout is the accumulated cost of years of unbilled camouflaging coming due at once.

It also points at what would help. A general instruction to mask less is useless. Knowing that your cost is concentrated in compensation, and that what you need is fewer situations demanding spontaneous social production, is something you can act on.

And if you have been assessed and turned away on the grounds that you seemed too socially capable, the compensation score is evidence about the very thing that produced that impression.

Read the number as a description of effort. Look at which subscale is highest, since that is where the cost is concentrated. Bring it to an assessment as one piece of evidence among several, which is what the authors intended. And if you are really asking whether you are autistic, that answer comes from a diagnostic assessment with a clinician who understands adult presentation rather than from a questionnaire about how hard you are working.

The separate question, which is what happens to your relationships when you stop, is covered here.

You can take the CAT-Q here.

References

Hull, L., Mandy, W., Lai, M.-C., Baron-Cohen, S., Allison, C., Smith, P., & Petrides, K. V. (2019). Development and validation of the Camouflaging Autistic Traits Questionnaire (CAT-Q). Journal of Autism and Developmental Disorders, 49(3), 819-833.

Hull, L., Lai, M.-C., Baron-Cohen, S., Allison, C., Smith, P., Petrides, K. V., & Mandy, W. (2020). Gender differences in self-reported camouflaging in autistic and non-autistic adults. Autism, 24(2), 352-363.

Lundin Remnélius, K., & Bölte, S. (2024). Camouflaging in autism: Age effects and cross-cultural validation of the Camouflaging Autistic Traits Questionnaire (CAT-Q). Journal of Autism and Developmental Disorders, 54(5), 1749-1764.

Content on this site is for educational and informational purposes only and does not constitute medical advice, diagnosis, or treatment. Nothing here creates a therapist-client relationship. If you are in crisis, call or text 988 (Suicide & Crisis Lifeline) or go to your nearest emergency room.

Clinical descriptions in this piece are illustrative composites drawn from patterns across many years of practice. They are not accounts of any specific client, session, or family.