Show the full module text
Module 5 — CAT-Q — Measuring the Cost of Fitting In
What camouflaging actually is, why it keeps people from being recognised for decades, what the 25 questions can and cannot tell you, and how to read three subscale scores that matter more than the total.
Autistic Self-Discovery · Part Three — Masking and the Shape of Your Mind · 14 min read, about 26 min with the workbook
The big idea
If you have ever come home from an ordinary evening and been unable to speak — or watched yourself run a conversation like a performance while some quieter part of you stood back and marked it out of ten — this module is about the thing you were doing, and about a questionnaire built to measure how much of it you do.
Step 1 — The lesson
A. The mask has a weight on it
Camouflaging is the effort of appearing more socially typical than you feel. Rehearsing what you will say before you say it. Copying someone else's expression because you are not sure what yours is doing. Forcing eye contact that costs you the thread of the sentence. Building a version of yourself for work, another for family, and losing track of which one is load-bearing.
The word people reach for is "mask," and it is a good word as far as it goes. What it misses is the weight. A mask you put on is one thing. A mask you hold up, all day, with your arms, is another.
Diagram — A · The mask has a weight on it. The face people meet is the light part. The cost is in the holding — the constant, unbroken effort of keeping it up, which is invisible to everyone in the room including, very often, you.
The original research that this questionnaire grew out of interviewed 92 autistic adults and found the same three-part story again and again.1 There were motivations — to fit in, to keep a job, to be safe, and, importantly, to genuinely connect with people. There were mechanisms — hiding what would show, and building explicit strategies to do socially what does not come automatically. And there were consequences: exhaustion, described by one participant as falling to pieces, and a slow erosion of the sense of being a real person rather than a performance.
Notice that connection is in the motivations list. Camouflaging is not only fear. A great deal of it is someone wanting very much to be known, and using the only tools they have.
B. Only the stage is lit
Here is why so many camouflaging adults reach forty without anyone having asked the question.
Diagnostic criteria are behavioural. They describe things another person can watch you do or fail to do. If you have spent twenty-five years getting extremely good at doing those things — at a cost nobody can see — then an assessment that watches your behaviour will watch a competent performance.
Diagram — B · Only the stage is lit. An assessment sees the performance. It does not see the hours afterwards in a dark room, the migraine, the cancelled plans, or the rehearsal that made the performance possible. A good performance is not evidence that it was easy.
Researchers have a name for this. Livingston and Happé distinguish deep compensation, which is flexible and generalises, from shallow compensation, which is rule-based and brittle and collapses under load or novelty.2 Both look like coping from the outside. Only one of them is. And because the criteria are written in terms of observable behaviour, a person compensating well can genuinely fail to meet them while every underlying difference is still there.
This is the single strongest argument for asking about camouflaging at all. Not because a camouflaging score diagnoses anything — it does not — but because without it, an assessment is reading the stage and calling it the whole theatre.
C. Three gears, one machine
The CAT-Q asks 25 questions, each answered on a seven-point scale from strongly disagree to strongly agree. Five of them are reverse-scored, which is a quiet way of checking you are reading rather than pattern-matching. It takes most people five to eight minutes.
The questions group into three things, and the three are worth understanding separately, because they feel different from the inside and they cost different amounts.
Diagram — C · Three gears, one machine. Three subscales feed one total between 25 and 175. Compensation is what you built; masking is what you hide; assimilation is the pull to disappear into the room. Most people are high on one or two rather than evenly high across all three — and which ones tell you more than the total does.
Compensation is the active work: the scripts, the researched topics, the deliberate study of how conversation is supposed to go. It is skilled, and often it is the part people are quietly proud of.
Masking is subtraction rather than addition: suppressing a stim, flattening a reaction, keeping a face still that wanted to do something else.
Assimilation is the one that tends to hurt most. It is the pull to not stand out, the sense of forcing yourself into rooms that cost you, the feeling of performing rather than being. Of the three, assimilation is the one most consistently linked in the research to anxiety, low wellbeing and a thinned-out sense of belonging.3
D. A dial with no fever line
Your total will land somewhere between 25 and 175. Higher means more effort spent fitting in.
Now the important part, and it is the thing most sites get wrong. The people who built the CAT-Q never published a cut-off score. There is no line above which you are officially a camouflager. The widely circulated figure of "100 or more means clinically significant camouflaging" comes from a non-peer-reviewed website, not from the validation paper, and it should not be treated as a threshold.
Diagram — D · A dial with no fever line. A thermometer has a fever line because someone established one. This one does not. Your score is a position on a continuous dial, not a pass or a fail, and the researchers who built it deliberately left the line off.
New Path's own results screen does show three bands — low, moderate and high — because a bare number with no context is unkind and unhelpful. Those bands are our presentation of your score, chosen to make it readable. They are not a validated diagnostic threshold, and nobody should treat them as one. You will see them in the workbook below, worded exactly as they appear on the screener itself.
What is worth reading closely is the shape across the three subscales. Someone high on compensation and low on assimilation is running a very different life from someone with the reverse. The total flattens that. The three numbers do not.
One more finding, because a lot of people arrive at this questionnaire because of it. Autistic women score significantly higher than autistic men — averages of about 124 against about 110 in the largest study, a moderate effect that survived controlling for autistic traits. Among non-autistic adults there was no such gap.4 A separate study using an entirely different method — comparing how someone presents in an assessment against their internal profile, rather than asking them — found the same direction with a large effect.5 Two methods, one answer. That is about as solid as this literature gets.
INSIDE THE INSTRUMENT
CAT-Q — Camouflaging Autistic Traits Questionnaire
Where it came from
Hull, Mandy, Lai, Baron-Cohen, Allison, Smith and Petrides (2019), Journal of Autism and Developmental Disorders 49(3), 819-833.6 Items were generated from the same team's 2017 qualitative study of 92 autistic adults and refined by external expert review.1
What it is made of
25 items, 7-point Likert (1 strongly disagree to 7 strongly agree), five items reverse-scored (3, 12, 19, 22, 24). Three subscales: Compensation (9 items, 9-63), Masking (8 items, 8-56), Assimilation (8 items, 8-56). Total 25-175, higher indicating more camouflaging.
How well it performs
Cronbach's alpha 0.94 total (0.91 autistic, 0.93 non-autistic); subscale alphas 0.86-0.93. Three-month test-retest r = 0.77 total, 0.70-0.78 by subscale — described by the authors themselves as preliminary, from only 30 autistic adults. Exploratory factor analysis yielded three factors and confirmatory analysis showed good fit.
Where it was validated
An online, self-selected, English-language sample of 354 autistic and 478 non-autistic adults (N = 832), split across exploratory (n = 402), confirmatory (n = 430) and validity (n = 706) analyses. Diagnoses were self-reported and unverified, and participants were predominantly diagnosed in adulthood, disproportionately female, and without intellectual disability. It has not been validated in under-15s, in non-binary people, or in people with intellectual disability or limited language — the authors state all three.
What it cannot do
No cut-off exists. The authors publish none and explicitly state that clinical utility requires further research. The circulating "100+" threshold has no peer-reviewed source.
Measurement invariance between autistic and non-autistic respondents has failed to replicate. The original paper reported strict invariance; the Dutch validation (N = 674, 356 autistic)7 and the French validation (N = 1,227, 744 autistic)8 both found the opposite — the French authors state plainly that the meaning behind items differs for autistic versus non-autistic people. Direct comparison of a client's score against non-autistic norms is therefore not psychometrically defensible, even though it is the most common use.
The factor structure does not travel. The Swedish adaptation (N = 639, ages 10-83) did not fit the original three-factor model and produced a two-factor solution in which the reverse-coded items formed their own factor — a method artefact. Discrimination in under-15s was at chance, AUC = 0.505.9
Construct specificity is genuinely contested. In the original autistic sample the total correlated more strongly with social anxiety (r = 0.44) than with autistic traits (r = 0.34), and the Masking subscale did not correlate with autistic traits at all (r = −0.03). A representative US general-population sample (N = 972) replicated the three-factor structure and found camouflaging predicted more strongly by ADHD traits (β = 0.27) than by autistic traits (β = 0.21); the authors conclude the CAT-Q measures camouflaging across neurotypes rather than specifically in autistic people.10 A 2026 review of 389 papers reaches the same conclusion.11
It requires self-awareness it cannot verify. Self-report and parent-report versions correlate only r = 0.47 with each other, and r = 0.20-0.46 with the behavioural discrepancy method — the two dominant ways of measuring camouflaging do not converge.12 Someone who masks automatically may score low for that reason alone.
The causal story is not settled. The cross-sectional association with anxiety and depression is well replicated,3 but a preregistered two-timepoint longitudinal study of 332 autistic adults found higher baseline camouflaging predicted a decrease in mental health difficulties over roughly two years, with small effects.13 The honest position is that masking and distress travel together and the direction is not established.
What a clinician does with it
In 2026 the CAT-Q's own developers published guidance on its use and misuse.14 They state that clinicians should not use CAT-Q scores in diagnostic decision-making, and should not use them as an intervention monitoring or outcome measure, noting that very little research has examined the instrument in clinical settings at all.
What it legitimately supports: structured language for something a client has often never had words for; context for a presentation that does not look autistic on observation; formulation of co-occurring distress, particularly where high assimilation sits alongside anxiety, exhaustion or burnout; and post-diagnostic work in which the cost of masking is named and unmasking becomes a legitimate goal.15
Validity tier: 1 — validated and published. A high-quality instrument with a genuinely contested construct. Both halves of that sentence are true.
Diagram — E · Two rulers that do not agree. Measurement invariance is the assumption that a question means the same thing to two groups. The Dutch and French validations both found it does not hold here. That is not a reason to discard the CAT-Q — it is a reason not to read your score as a comparison against non-autistic people.
STRENGTHS LENS
The skill was real. That is the part nobody says.
A high compensation score describes something genuinely difficult that you taught yourself, usually without help and often without knowing you were doing it. Reading a room from first principles, building scripts that work, running a social interaction on manual — these are not failures. They are an enormous amount of learning.
The problem was never the skill. It is that it has no off switch, that nobody has ever relieved you at the post, and that the version of you people know is the version that costs the most to produce. Unmasking is not unlearning the skill. It is getting to choose when to use it.
F. What helps
Camouflaging is not a habit you drop on a Tuesday. It kept you employed, safe, and sometimes alive. What follows is about choice and cost, not about stopping.
1. Find one room where you do less.
Not all rooms — one. A friend, a partner, a group, an hour alone in the car. Unmasking research consistently points to the value of somewhere the performance is not required, rather than a wholesale change. Notice what you do differently there. That is your baseline, and most people have not seen theirs in years.
2. Budget the cost instead of denying it.
If a two-hour work event reliably takes you a whole evening to recover from, that event costs six hours, not two. Plan the recovery into the calendar rather than being surprised by it. People who do this report far less of the shame that comes from wondering why they are so tired after something that looked easy.
3. Let the stim back in, privately first.
Suppressing self-regulating movement costs attention that then is not available for anything else. Rocking, fidgeting, pacing, humming and pressure all do a job. Reintroduce them somewhere unobserved before you worry about anywhere else.
4. Say the cost out loud to one person.
The isolating part of camouflaging is not the effort — it is that the people closest to you are often responding to the performance and have no idea there is one. One sentence to one person changes that: "that was fine, and it took everything I had."
5. Treat exhaustion as data, not weakness.
Sustained camouflaging is one of the stressors most consistently named in the autistic burnout literature, and unmasking is named as a recovery factor.15 If you are flattened rather than merely tired, Part Five of this course is built for exactly that, and the ABSI-24 in Module 13 asks about right now rather than in general.
Step 2 — Take the screener
Twenty-five questions, five to eight minutes, free and confidential. You will get a total between 25 and 175 plus three subscale scores. Read section D first if you can — the shape across the three subscales is worth far more than the total, and there is no pass mark to hit.
Before you start
The CAT-Q (Hull et al., 2019) is a validated research questionnaire measuring how much you camouflage in social situations. It reliably captures self-reported camouflaging, but it is a screen, not a diagnosis — only a clinician can diagnose, and the authors published no cut-off score, so no number here means you are or are not autistic. Treat your result as a mirror for noticing patterns and a starting point for that conversation.
Step 3 — Your workbook
Your answers save to this device only — we cannot see a word of what you write. This module lets you score the CAT-Q if you took it, then map where your camouflaging actually happens, what it costs you, and the one room where you already do less of it.
Your CAT-Q results
Took the CAT-Q? Put your numbers in below to see where each one lands. Entirely optional — skip it if you would rather just read.
Score bands: 25–80 = Low Camouflaging; 81–120 = Moderate Camouflaging; 121–175 = High Camouflaging
Fields: CAT-Q · Camouflaging Autistic Traits Questionnaire; Total score (25–175); My total (enter 25–175); Compensation (9–63); Compensation (enter 9–63); Masking (8–56); Masking (enter 8–56); Assimilation (8–56); Assimilation (enter 8–56)
Which of the three is highest for you?
The shape matters more than the total. Compensation is what you built; masking is what you hide; assimilation is the pull to disappear into the room.
Fields: My highest subscale: Compensation / Masking / Assimilation / They are close together; Does that match how it feels from the inside?
The room where you do the most
Not the hardest room — the one where the performance runs longest.
Fields: Where; What specifically am I doing there that I would not do alone?
What does it cost?
Section F, item 2. If a two-hour event takes an evening to recover from, the event costs six hours.
Fields: A recent event, and how long it took; What the recovery actually looked like
One room where you already do less
There is usually one, even if it is the car. If there is not one yet, name the likeliest candidate.
Fields: The room or the person; What is different about me there
The one sentence
Section F, item 4. Write the version you would actually say.
Fields: Who I would say it to; What I would say
One thing to let back in
Somewhere unobserved, first.
Fields: What I am going to stop suppressing, and where
Appendix — Research companion
Peer-reviewed research
10. Ai W, Cunningham WA, Lai M-C (2024). The dimensional structure of the Camouflaging Autistic Traits Questionnaire (CAT-Q) and predictors of camouflaging in a representative general population sample. Comprehensive Psychiatry, 128, 152434. DOI 10.1016/j.comppsych.2023.152434. View the paper N = 972 US representative general-population sample. Three-factor structure replicated. Camouflaging was predicted by ADHD traits (beta = 0.27) more strongly than autistic traits (beta = 0.21), plus social anxiety; the authors conclude the CAT-Q measures camouflaging across neurotypes rather than specifically in autistic people. Limitation: all measures self-report and cross-sectional; it does not capture how effective camouflaging actually is.
8. Bureau R, Riebel M, Weiner L, Coutelle R, Dachez J, Clement C (2024). French validation of the Camouflaging Autistic Traits Questionnaire (CAT-Q). Journal of Autism and Developmental Disorders, 54(9), 3549-3558. DOI 10.1007/s10803-023-06048-w. View the paper N = 1,227 (744 autistic, 483 non-autistic). Three-factor structure, internal consistency and test-retest all good, but measurement invariance testing indicated the meaning behind items differs for autistic versus non-autistic people. Limitation: small test-retest subsample (n = 22); the invariance failure means cross-group score comparisons are not psychometrically safe.
12. Hannon B, Mandy W, Hull L (2023). A comparison of methods for measuring camouflaging in autism. Autism Research, 16(1), 12-29. DOI 10.1002/aur.2850. View the paper Systematic review identifying 16 camouflaging measurement tools plus development of a parent-report CAT-Q. Self-report and parent-report correlated only r = 0.47 with each other, and r = 0.20 and r = 0.46 respectively with the behavioural discrepancy approach. Limitation: the parent-report validation sample was young people rather than adults.
4. Hull L, Lai M-C, Baron-Cohen S, Allison C, Smith P, Petrides KV, Mandy W (2020). Gender differences in self-reported camouflaging in autistic and non-autistic adults. Autism, 24(2), 352-363. DOI 10.1177/1362361319864804. View the paper 306 autistic and 472 non-autistic adults. Autistic women scored higher than autistic men (124.35 versus 109.64; d = 0.65), with no gender difference among non-autistic adults; the effect survived controlling for autistic traits. Limitation: non-binary subsamples were underpowered (16 autistic, 27 non-autistic) and the authors state the CAT-Q has not been psychometrically validated in non-binary populations.
3. Hull L, Levy L, Lai M-C, Petrides KV, Baron-Cohen S, Allison C, Smith P, Mandy W (2021). Is social camouflaging associated with anxiety and depression in autistic adults?. Molecular Autism, 12, 13. DOI 10.1186/s13229-021-00421-1. View the paper 305 autistic adults aged 18-75. Higher CAT-Q scores predicted greater generalised anxiety, social anxiety and depression after controlling for age and autistic traits; effects small but consistent; the relationship was linear with no threshold effect. Limitation: cross-sectional, so the authors state they cannot determine which developed first.
6. Hull L, Mandy W, Lai M-C, Baron-Cohen S, Allison C, Smith P, Petrides KV (2019). Development and validation of the Camouflaging Autistic Traits Questionnaire (CAT-Q). Journal of Autism and Developmental Disorders, 49(3), 819-833. DOI 10.1007/s10803-018-3792-6. View the paper Original validation. Online sample of 354 autistic and 478 non-autistic adults; total alpha 0.94; three-month test-retest r = 0.77 from only 30 autistic adults, described by the authors as preliminary. Limitation: participants were largely diagnosed in adulthood with self-reported, unverified diagnoses; the authors state the CAT-Q is not a diagnostic tool and publish no cut-off score.
1. Hull L, Petrides KV, Allison C, Smith P, Baron-Cohen S, Lai M-C, Mandy W (2017). "Putting on my best normal": social camouflaging in adults with autism spectrum conditions. Journal of Autism and Developmental Disorders, 47(8), 2519-2534. DOI 10.1007/s10803-017-3166-5. View the paper Qualitative thematic analysis of 92 autistic adults; the study the CAT-Q items were derived from. Produced the three-stage model of motivations, mechanisms and consequences. Limitation: self-selected online sample, 55% British, no behavioural verification.
5. Lai M-C, Lombardo MV, Ruigrok ANV, Chakrabarti B, Auyeung B, Szatmari P, Happe F, Baron-Cohen S, MRC AIMS Consortium (2017). Quantifying and exploring camouflaging in men and women with autism. Autism, 21(6), 690-702. DOI 10.1177/1362361316671012. View the paper 30 autistic men and 30 autistic women, age- and IQ-matched, aged 18-49, without intellectual disability. Camouflaging measured as the discrepancy between external presentation and internal disposition rather than by self-report; women camouflaged more than men, d = 0.98. Limitation: small sample, excludes intellectual disability, and this method correlates only weakly with self-report measures such as the CAT-Q.
2. Livingston LA, Happe F (2017). Conceptualising compensation in neurodevelopmental disorders: reflections from autism spectrum disorder. Neuroscience & Biobehavioral Reviews, 80, 729-742. DOI 10.1016/j.neubiorev.2017.06.005. View the paper Theoretical review introducing the deep versus shallow compensation distinction and the argument that behaviourally defined criteria can be met despite persisting underlying differences. Limitation: a conceptual review, not an empirical test, and not a validation of any instrument.
9. Lundin Remnelius K, Bolte 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. DOI 10.1007/s10803-023-05909-8. View the paper Swedish adaptation, N = 639 (100 autistic, 539 general population), ages 10-83. The original three-factor structure did not fit; an exploratory analysis produced a two-factor solution in which reverse-coded items formed their own factor. Discrimination was at chance for under-15s (AUC = 0.505). Limitation: only 100 autistic participants, and cross-cultural measurement invariance was not formally tested.
7. van der Putten WJ, Agelink van Rentergem JA, Radhoe TA, Torenvliet C, Groenman AP, Mol AJJ, Geurts HM (2023). How to measure camouflaging? A conceptual replication of the validation of the Camouflaging Autistic Traits Questionnaire in Dutch adults. Research in Autism Spectrum Disorders, 100, 102072. DOI 10.1016/j.rasd.2022.102072. View the paper N = 674 (356 autistic), aged 30-92, with ADOS-2 in 90 autistic adults. Three-factor structure had acceptable fit, but found no measurement invariance between autistic and non-autistic individuals, contradicting the original validation. Limitation: an older cohort (30+), so findings may not generalise to younger autistic adults.
Clinical frameworks and position statements
14. Hannon BR, Hull L, Lai M-C, Magiati I, Mandy W (2026). The use and misuse of the Camouflaging Autistic Traits Questionnaire in autism research and clinical practice: issues, considerations, and suggestions. Autism in Adulthood, published online 26 March 2026. DOI 10.1177/25739581261435868. View the source Guidance paper co-authored by the CAT-Q's own developers. States that clinicians should refrain from using CAT-Q scores in diagnostic decision-making or as an intervention monitoring or outcome tool, and that very little research has examined the CAT-Q in clinical settings. Limitation: published online first, so volume, issue and page range are not yet assigned.
Lived experience
15. Raymaker DM, Teo AR, Steckler NA, Lentz B, Scharer M, Delos Santos A, Kapp SK, Hunter M, Joyce A, Nicolaidis C (2020). "Having all of your internal resources exhausted beyond measure and being left with no clean-up crew": defining autistic burnout. Autism in Adulthood, 2(2), 132-143. DOI 10.1089/aut.2019.0079. View the source Community-based participatory research with autistic co-researchers; 19 interviews plus 19 social media sources. Identifies masking as a prominent life stressor producing cumulative exhaustion, warns against teaching autistic people to mask, and identifies unmasking as a recovery factor. Limitation: qualitative and self-selected; does not quantify the camouflaging-burnout relationship or use the CAT-Q.
Emerging or contested
11. Arnold WM, Bitsika V, Sharpley CF (2026). Camouflaging and autism: conceptualisation and methodological issues. Autism, 30(5), 1131-1146. DOI 10.1177/13623613261420085. View the paper Review of 389 papers reporting that CAT-Q correlations with autistic traits are often weaker than with social anxiety, that the Masking subscale is unrelated to autistic traits across multiple studies, and that the three-factor structure fails replication in several translations. Limitation: a narrative and conceptual critique rather than new primary data.
13. van der Putten WJ, Mol AJJ, Radhoe TA, Torenvliet C, Agelink van Rentergem JA, Groenman AP, Geurts HM (2025). Camouflaging in autism: a cause or a consequence of mental health difficulties?. Autism, 29(10), 2604-2617. DOI 10.1177/13623613251347104. View the paper Preregistered two-timepoint longitudinal study, 332 autistic adults aged 30-84, mean 2.0-year follow-up. Higher baseline camouflaging predicted a decrease in mental health difficulties over time, complicating the common cross-sectional claim that masking causes poor mental health. Limitation: small effect sizes, only two timepoints, cohort aged 30+.
Further reading — general background
Bradley L, Shaw R, Baron-Cohen S, Cassidy S (2021). Autistic adults' experiences of camouflaging and its perceived impact on mental health. Autism in Adulthood, 3(4), 320-329. DOI 10.1089/aut.2020.0071. View the source Thematic analysis of open-ended survey responses from 277 autistic adults. Camouflaging described as exhausting and linked to anxiety, depression, loss of identity and delayed diagnosis, while also enabling employment and offering protection from bullying; the amount of time spent camouflaging, not camouflaging itself, was most damaging. Limitation: qualitative and self-selected, including both diagnosed and self-identified participants.
Cook J, Hull L, Crane L, Mandy W (2021). Camouflaging in autism: a systematic review. Clinical Psychology Review, 89, 102080. DOI 10.1016/j.cpr.2021.102080. View the paper Systematic review of 29 empirical studies covering both measurement traditions, sex and gender differences, and the association between higher self-reported camouflaging and worse mental health outcomes. Limitation: the authors note the reviewed samples cannot be generalised to the whole autistic community.
Peer-reviewed = checked by independent experts before publication. Clinical model = an established professional framework, not a single study.
Up next
Module 6 - RBQ-2A — Repetitive Behaviour, Explained Kindly
All modules in Autistic Self-Discovery
Putting the mask down is easier with company. This course was built by clinicians who are part of the New Path family of therapy centers. Unmasking is not something most people manage alone, and a high assimilation score alongside exhaustion is worth taking to someone. Therapy for clients in California, coaching worldwide, all by telehealth. A conversation costs nothing and there is no pressure. Saving this for later counts too. Talk with the New Path team