Show the full module text
Module 8 — EDA-QA — Demand Avoidance and PDA Traits
Why a task can be effortless when it is your idea and impossible when it is asked, what the twenty-six statements of the EDA-QA actually measure, and an honest account of a construct that sits in neither diagnostic manual.
Autistic Self-Discovery · Part Three — Masking and the Shape of Your Mind · 13 min read, about 26 min with the workbook
The big idea
If you have ever sat for eleven days in front of an email you could answer in ninety seconds, or felt something in you go flat and hard the moment somebody said "could you just" — this module is about the difference between a task being hard and a task being asked. It is also about a questionnaire that measures something clinicians are still arguing about, which is exactly why we are going to be careful with it.
The circularity is the crux. Studies identify PDA using a questionnaire built from a clinical description, then use those groups to show that PDA exists. Nothing outside the description — no biological marker, no independent clinical standard, no treatment response — has yet been brought in to break the loop. That is a real problem with the science. It is not a statement about whether your life is hard.
Step 1 — The lesson
A. Not the task. The asking.
There is a kind of stuck that has nothing to do with difficulty. The dishes you were on your way to do, until somebody mentioned the dishes. The form you had been meaning to fill in, which became unthinkable the day it acquired a deadline. The novel you loved at fourteen and could not open at fifteen, because it turned up on a reading list.
Most avoidance scales with how hard or unpleasant a thing is. This does not. It scales with how much the thing arrives as a demand. A task you invent at two in the morning can be effortless. The same task, asked for kindly by somebody who loves you, can be flatly impossible for a fortnight. Nothing about the work changed. Only where the instruction came from.
That is why it is misread so reliably. Nobody watching can see the join between will not and cannot, so they reach for the nearest available words: lazy, rude, spoilt, defiant, difficult. A great many people who recognize themselves here had heard all five before they were twelve.
It also points inwards, which is the part almost nobody predicts. Your own plan for the morning becomes a demand the moment it is fixed. So does hunger. So does the list you wrote yourself on Sunday, full of things you genuinely want to do. The demand does not have to come from a person. It only has to arrive as a demand.
The description is not new. Elizabeth Newson, a developmental psychologist working in Nottingham, spent years seeing children who did not fit the autistic profiles she knew. They resisted everyday requests, but with social strategy rather than withdrawal — distraction, negotiation, charm, sudden incapacity. In 2003 she and her colleagues published an account of 150 clinic cases, reported an even split between girls and boys, and argued that this deserved a category of its own. She called it pathological demand avoidance.1
Diagram — A · Not the task. The asking.. Notice the gap. The clinical description came first and the measuring instrument came eleven years later, which is the reverse of how most screeners are built. Everything since 2014 has been an attempt to test an idea that was already in wide use.
The idea grew fastest outside clinics. In 2018 the PDA Society, a UK charity, ran a two-week survey that drew 1,445 responses. Among the young people described, 70 per cent were unable to tolerate school or were educated at home, and 71 per cent of the 79 adults who answered named a lack of acceptance and understanding as the main barrier to getting support.2 Whatever the label turns out to be, the difficulty it points at is neither small nor rare.
Adults describe it in remarkably consistent terms. In a 2024 phenomenological study, participants spoke of pervasive anxiety — one described fear as her default state — of feeling like a fraud for failing standards she had set herself, and of long distrust of services that read her difficulties as attachment problems or poor parenting. Finding the concept brought relief.3 That relief is real and worth taking seriously.
And yet. Pathological demand avoidance is not a diagnosis in DSM-5 and not a diagnosis in ICD-11. It is neither a standalone entity nor a recognized autistic subtype in either manual.4 Both things are true at once: the experience is well described and heavily documented, and the category is not established. This module is going to hold both, because collapsing either one would be dishonest. Many adults who identify with the profile dislike the word pathological and read the initials instead as a persistent drive for autonomy, which describes the same phenomenon without the verdict.
B. Twenty-six statements and one number
The questionnaire began life as something a parent filled in about a child. In 2014 Elizabeth O'Nions and colleagues built the Extreme Demand Avoidance Questionnaire from parent reports on 326 children aged 5 to 17, across six groups: typically developing children, autistic children with and without disruptive behavior, children whose parents suspected PDA, children a health professional had identified, and children with disruptive behavior and no autistic diagnosis. Twenty-six items, each scored 0 to 3, total 78. For ages 5 to 11 a cut-off of 50 gave sensitivity of .80 and specificity of .85; for 12 to 17 year olds a cut-off of 45 gave .88 and .78.5
Five years later Vincent Egan, Omer Linenberg and Elizabeth O'Nions rewrote those same 26 statements in the first person so an adult could answer for themselves. That is the EDA-QA. In a sample of 347 adults its internal consistency was .94, and .92 in a second sample of 191. The items hang together as essentially one thing: a single general dimension accounted for 40.54 per cent of the variance, with a small second factor concerning fantasy and role play adding 6.68 per cent.6 There are no published subscales. You get one number.
Diagram — B · Twenty-six statements and one number. Two of the 26 items are reverse-scored, so agreeing with everything is not how a high total is built. The questionnaire is asking about a mechanism, not a mood.
What the statements cover is worth knowing before you answer them. Obsessively resisting ordinary requests. Needing to be the one in charge. Giving reasons of incapacity — illness, tiredness, an arm that suddenly will not work — when a demand lands. Mood that switches very fast. Using social strategies rather than withdrawal. Comfort in role play. Seeming socially fluent while missing the parts of hierarchy other people take for granted.
Read that list and you can already see the trouble. Several of those descriptions would raise a score in somebody who is simply frightened, or exhausted, or has spent a decade being told what to do by people who never explained themselves. The instrument is honest about what it counts. It has no way of knowing why.
One more thing, plainly. There is no validated cut-off for adults. None. The test publisher that distributes the EDA-QA states in its own documentation that no clinical cut-off has been validated and that scores should be read dimensionally rather than categorically.7 Every threshold attached to this questionnaire, including the one on our own screener, is a convention rather than a finding.
C. Anxiety, autism, or something of its own
If the score goes up, what pushed it? This is the live scientific question, and unusually for a contested area there is decent adult data on it.
In 2023 Rhianna White and colleagues ran the EDA-QA in two general-population adult samples, 267 and 549 people. Both autistic traits and anxiety came out as unique predictors of demand avoidance, and when the researchers tested which mattered more, the two were about equal — neither reliably outranked the other. The full model, including stress, sex and age, accounted for 53 per cent of the variance in demand avoidance scores.8
A separate 2023 study of 163 adults found the EDA-QA correlated .58 with a broad measure of anxious and depressive symptoms and .46 with intolerance of uncertainty. Anxiety alone accounted for 27.8 per cent of the variance in demand avoidance; intolerance of uncertainty added under two per cent on top of it.9 Anxiety is not a footnote to this construct. It is most of the engine.
Diagram — C · Anxiety, autism, or something of its own. This is the single most important limitation on the page. A raised total is consistent with several very different situations, and the questionnaire has no way of telling them apart.
It gets more tangled. In a 2020 study by Egan, Bull and Trundle, adult demand-avoidant traits tracked ADHD symptoms, emotional instability and antagonism more closely than autistic traits.10 Depending on which adult study you read, the EDA-QA looks like an autistic measure, an anxiety measure, an ADHD measure or a personality measure. That is not a bad questionnaire. It is a thing that has not been pinned down.
None of which means your experience is "just anxiety". Anxiety is a plausible account of why a demand feels like a threat. It is not an account of why the same task is effortless when you thought of it yourself. Both facts need explaining, and no single model currently explains both to everyone's satisfaction.
D. What the score does and does not mean
Our screener puts the boundary between its two bands at 45 and 46. You should know where that number came from, because it did not come from a study of adults.
Forty-five is the cut-off O'Nions and colleagues derived for 12 to 17 year olds, from parents answering about their children.5 No equivalent threshold has ever been derived in an adult sample. No study has reported the sensitivity or specificity of any EDA-QA score in adults against a clinical judgment, because there is no agreed clinical judgment to test it against.7
There is a further wrinkle that almost nobody mentions. The same 26 items are scored two different ways in circulation. Our screener, like the original EDA-Q, scores each item 0 to 3, giving a total of 0 to 78. Several published adult studies score the same items 1 to 4, giving 26 to 104 — exactly 26 points higher for identical answers.8 That matters when you try to compare your number with a published average, and it is why we are showing you the working rather than a percentile.
Here is the comparison, converted onto the 0 to 78 scale this screener uses. In a general-population sample of 548 UK adults the average total was about 17. In a mixed sample of 126 adults recruited from diagnosis-related forums and support groups, of whom 61.5 per cent had an autistic diagnosis, the average was about 35.7 Our high band starts at 46. That is roughly 29 points above the general-population average and around 11 points above the average of a group in which most people were already autistic. It is a demanding threshold.
Diagram — D · What the score does and does not mean. A score in the low band is not a clearance. Given where the threshold sits, plenty of people whose lives are shaped by demand avoidance will land below it. Read the low band as 'this questionnaire did not pick it up', not as 'this is not happening to you'.
Then there is the construct itself. In 2018 a group of senior UK autism researchers reviewed the evidence and concluded that it does not support pathological demand avoidance as an independent syndrome, proposing instead that the difficulties are better understood as social, sensory and cognitive sensitivities interacting with the environment.11 A 2021 systematic review of 13 studies found that every one of them based its criteria on Newson's original descriptions, whose method of derivation was never clearly set out; nearly all relied on parent report alone; and not one study had asked people with PDA about their own experience.12
A 2026 systematic review put twelve studies through a formal risk-of-bias appraisal: nine in children, two in adults, one mixed. Only two instruments were in use across the whole literature, the EDA-Q and a diagnostic interview, applied inconsistently. Every study was rated at high risk of bias, and the authors concluded that the reliance on subjective measures makes it difficult to validate either the tools or the condition.13 Twelve studies is a thin evidence base for a concept this widely used.
INSIDE THE INSTRUMENT
EDA-QA — Extreme Demand Avoidance Questionnaire for Adults
Where it came from
Elizabeth Newson published the defining account in 2003, arguing for a distinct category within the pervasive developmental disorders on the basis of 150 clinic cases and a discriminant analysis.1 O'Nions, Christie, Gould, Viding and Happe operationalised the description as the 26-item parent-report EDA-Q in 2014.5 Egan, Linenberg and O'Nions rewrote the same items in the first person as the EDA-QA in 2019.6
What it is made of
Twenty-six self-report statements, each scored 0 to 3 (not true, somewhat true, mostly true, very true), items 14 and 20 reverse-scored, total 0 to 78. Exploratory factor analysis in the development sample gave one dominant general dimension explaining 40.54% of variance plus a minor fantasy and role-play dimension at 6.68%; a single-factor confirmatory model reached RMSEA 0.064.6 There are no validated subscales and no published short form for adults. The eight-item EDA-8 exists, but it is a caregiver measure for children.
How well it performs
Internal consistency is consistently high: alpha .94 in the 347-adult development sample and .92 in the 191-adult replication,6 with .93 (n = 266) and .87 (n = 548) in independent samples.8 Convergent evidence is mostly with distress rather than with autism: r = .40 with a brief autistic traits screen,6 r = .58 with a combined anxiety and depression measure and r = .46 with intolerance of uncertainty.9 The only published sensitivity and specificity figures in this family are the EDA-Q's, from parents reporting on children: .80 / .85 at a cut-off of 50 for ages 5 to 11 and .88 / .78 at 45 for ages 12 to 17.5 No test-retest reliability has been published for the EDA-QA.
Where it was validated
UK online convenience samples, self-selected, predominantly female and highly educated, with self-reported rather than clinically confirmed diagnoses.6 The largest normative reference is 548 UK adults aged 18 to 67 recruited through Prolific in May 2020, mean 42.88 (SD 9.31) on the 26 to 104 scoring convention — about 17 on the 0 to 78 convention this screener uses. A mixed sample of 126 adults recruited from forums and support groups, 61.5% with an autistic diagnosis, averaged 60.52 (SD 17.22), about 35 on the 0 to 78 convention.7 No clinical, forensic or non-UK normative data exist.
What it cannot do
It has no validated cut-off in adults, and the threshold in circulation is borrowed from children. The 45 boundary is the EDA-Q's 12-to-17 parent-report cut-off.5 The publisher's own documentation states that no clinical cut-off has been validated for the EDA-QA and recommends dimensional interpretation.7 Any categorical statement built on this score is an assumption, not a result.
It cannot separate demand avoidance from anxiety. In two adult samples (n = 267 and n = 549) autistic traits and anxiety were each unique predictors and neither reliably dominated; the full model explained 53% of variance.8 In a third sample anxiety alone accounted for 27.8% of variance in EDA-QA scores.9 In a fourth, adult demand-avoidant traits tracked ADHD symptoms, emotional instability and antagonism more closely than autistic traits.10
It cannot establish that the construct it measures exists. PDA appears in neither DSM-5 nor ICD-11.4 A 2018 review concluded the evidence does not support it as an independent syndrome.11 A 2021 systematic review of 13 studies found all criteria traced back to Newson's descriptions, whose derivation was never clearly described.12 A 2026 review of 12 studies rated every one at high risk of bias and named the circular logic explicitly.13
It is not scored the same way everywhere. The EDA-Q and most screening implementations score items 0 to 3 (total 0 to 78); several published adult studies and at least one commercial platform score the same items 1 to 4 (total 26 to 104).8 A raw total is uninterpretable without knowing which convention produced it.
It does not track autistic severity, and was never meant to. When the parent-report items were refined into the eight-item EDA-8 in a sample of 334 autistic children (alpha .90), scores were independent of age, gender, ability level and parent-rated autistic severity.14 Demand avoidance is not a measure of how autistic somebody is, and a high EDA-QA score is not corroborating evidence for an autistic assessment.
What a clinician does with it
Use it as a structured interview schedule and a shared vocabulary, not as a score. The UK PDA Society's practice guidance, written by twelve contributors and endorsed by fifteen more practitioners, states directly that the questionnaire "should not be considered a diagnostic test", that it is useful mainly for structuring qualitative information-gathering even when not formally scored, and that the sensitivity and specificity difficulties seen in the child measure should be expected in adults.15 The clinically useful output is a functional analysis — which demands, from whom, in what state, with what consequence — and the environment and scripts built around it. Screen for anxiety, autistic traits and ADHD alongside it, because the score alone will not distinguish them.
Validity tier: 2 — a published instrument with sound reliability, measuring a construct that is genuinely disputed. The EDA-QA is peer-reviewed, internally consistent and replicated across independent adult samples, which puts it above an in-house tool; but it has no validated adult cut-off, no test-retest data, no clinical criterion standard, and the category it names is recognized by neither diagnostic manual, which keeps it well below the tier of an established screening measure.
Diagram — E · Two explanations, one experience. This is the honest picture as of today. Autistic traits and anxiety each independently predict demand avoidance in adults, and the argument about which is primary, or whether a third thing sits underneath both, has not been settled.
STRENGTHS LENS
You have been defending something that is worth defending.
Underneath the avoidance is a nervous system that treats autonomy as non-negotiable. That has cost you a great deal and this module is not going to pretend otherwise. It has also meant you were never easy to coerce. People who cannot be pressured do not get talked into things they will regret, and they tend to notice when a hierarchy is asking for obedience rather than competence.
The associated skills are real too. Reading a room fast enough to redirect it, negotiating instead of submitting, an imaginative life detailed enough that the original questionnaire needed a separate item for it. The problem has almost never been the drive for autonomy. It is that most environments run on unexplained instruction, and nobody ever offered you the version that would have worked.
F. What helps
The evidence for what helps is thinner than the evidence about what this is. There are no randomised trials in adults. What follows comes from practice guidance and from what large numbers of adults report works — a lower grade of evidence, and worth knowing as you read it. The consistent thread is that the demand gets changed rather than the person.
1. Change where the demand appears to come from.
There is a large difference between "take the bins out" and "the bins are full." The second states a problem and leaves you the author of the solution; the first hands you an instruction to resist. Ask the people around you to describe situations rather than issue tasks, and give yourself the same courtesy when you write your own list.
2. Build a real choice into everything.
Not a fake one. Two genuine options, or the order, or the timing, or which part you start with. Autonomy returned in small quantities does a surprising amount of work, because what triggered the freeze was not the task but the absence of any move that was yours. If no choice is available, say so rather than dressing a demand up as an offer — that reads as manipulation and costs more than honesty would have.
3. Lower the stakes before you lower the task.
People usually try to make the job smaller. The more effective move is often to make it less loaded: remove the deadline, remove the audience, make it reversible, agree that a bad first attempt is acceptable. Given how strongly demand avoidance tracks anxiety and intolerance of uncertainty in the adult data, taking the threat out of a task works on the mechanism rather than the symptom.
4. Count the demands you set yourself.
This is the one people miss. Your own plan becomes an external demand the moment it hardens into a commitment, which is why Sunday's enthusiastic list is unopenable by Tuesday. Keep a menu rather than a list, and let the order be decided on the day. Watch the biological demands too — eating, drinking, sleeping, washing — because those get avoided on the same mechanism and the cost accumulates quietly.
5. Give people the sentence, not the diagnosis.
You do not need anyone to accept a contested construct in order to change how they speak to you. Something like: "I will get this done. I cannot do it to a countdown, and I need the problem rather than the instruction." That is a workable request whatever anybody believes about PDA. The workbook below is built to help you write your own version of it.
Step 2 — Take the screener
Twenty-six statements, five to seven minutes, free and confidential. You get one number between 0 and 78 and a band description. Read section D before you read too much into the band: there is no cut-off for adults that anyone has validated, so a low result does not rule this out and a high one does not settle anything.
Before you start
The EDA-QA is a published 26-item self-report measure of extreme demand avoidance traits, adapted in 2019 from a parent-report questionnaire about children. It is a screen, not a diagnostic test. Pathological demand avoidance appears in neither DSM-5 nor ICD-11, and systematic reviews have found the evidence does not yet establish it as a condition distinct from autism and anxiety. No adult cut-off score has ever been validated, so the band boundary shown here is a convention rather than a research finding, and raised scores also occur with anxiety, ADHD, burnout and trauma. Only a qualified clinician can diagnose anything.
Step 3 — Your workbook
Your answers save to this device only — we cannot see a word of what you write. This module's workbook is less about the score than about the mechanism: where your demands come from, what else might be raising the number, and the three or four sentences that make the people around you easier to live with.
Your EDA-QA results
Took the screener? Put your number in below. Read the band description as a description, not a verdict — there is no cut-off for adults that anyone has validated, and the boundary here is borrowed from a parent-report measure for 12 to 17 year olds. Entirely optional — skip it if you would rather just read.
Score bands: 0–45 = Low Prevalence of Demand Avoidance Traits; 46–78 = High Prevalence of Demand Avoidance Traits
Fields: EDA-QA · Extreme Demand Avoidance Questionnaire for Adults; Total score (0–78); My total (enter 0–78); How the band description landed: That is my life described / Partly right / It missed the point / I have not taken the screener
Own idea, or somebody else's
Section A. The test of this profile is not whether a task is hard — it is whether the same task changes when it becomes a request.
Fields: Something I do easily when it is my idea, and cannot do when it is asked; What actually happens in my body when a demand lands; Roughly when did I first notice this pattern: Primary school / Teens / Twenties / Later than that / It has always been there
What else could be pushing the number up
Section C. Autistic traits, anxiety, ADHD, burnout and long exposure to environments that never explained themselves all raise this score. Naming yours changes what you do next.
Fields: Tick: Anxiety is a large part of this for me; Tick: I am currently burnt out or recovering; Tick: Attention and starting things are difficulties in their own right; If the demands in my life halved tomorrow, what would change
Rewriting the demand
Section F, item 1. Turn three instructions you receive regularly into statements of the problem instead.
Fields: Instruction one, rewritten as a situation; Instruction two, rewritten; Instruction three, rewritten; Who says these to me, and how I would ask them to change it
The demands I set myself
Section F, item 4. Your own plan becomes an external demand the moment it hardens. This is the group most people find unexpectedly useful.
Fields: Things I want to do that become impossible once they are on a list; Which basics get avoided on this mechanism — eating, drinking, sleeping, washing; One list I will convert into a menu this week
Lowering the stakes
Section F, item 3. Before shrinking a task, try removing the deadline, the audience or the permanence.
Fields: A task I have been stuck on; What is loading it most: A deadline / Being watched or checked on / It has to be right first time / Somebody asked me to / I genuinely do not know; What removing that one thing would look like in practice
The sentence
Section F, item 5. One request somebody could act on tomorrow, that does not require them to accept a contested diagnosis.
Fields: Who I would say it to; The exact sentence
Appendix — Research companion
Peer-reviewed research
13. Company D, Rotella JA (2026). A systematic review of Pathological Demand Avoidance (PDA): a veritable diagnosis or a case of circular logic?. Journal of Autism and Developmental Disorders, published online 9 July 2026, 1-11. DOI 10.1007/s10803-026-07432-y. View the paper Systematic review searching MEDLINE, EMBASE and PsycINFO, with dual independent screening and formal QUADAS-2 bias appraisal, to test whether PDA can be identified reliably enough to count as a distinct diagnosis. Twelve studies met criteria: nine paediatric, two in adults and one mixed cohort. Only two identification instruments appeared across the whole literature, the EDA-Q and the DISCO diagnostic interview, and both were applied inconsistently; cross-over with other conditions was common; every included study was rated at high risk of bias. The authors conclude that reliance on subjective rather than objective measures creates circular logic that prevents validation of either the instruments or the condition. Limitation: only twelve studies, English-language only, and the review itself was not prospectively registered, which the authors acknowledge.
10. Egan V, Bull E, Trundle G (2020). Individual differences, ADHD, adult pathological demand avoidance, and delinquency. Research in Developmental Disabilities, 105, 103733. DOI 10.1016/j.ridd.2020.103733. View the paper An adult study using the EDA-QA alongside ADHD, personality and delinquency measures in a sample recruited through online forums and support groups, of whom the majority reported an autistic diagnosis. Adult demand-avoidant traits related more strongly to ADHD symptoms, emotional instability and antagonism than to autistic traits, and were associated with rule-breaking behaviour. It supplies the only published reference data for the EDA-QA in a sample enriched for clinical diagnoses. Limitation: a cross-sectional online convenience sample with self-reported and clinically unverified diagnoses, no comparison group matched on diagnosis, and no ability to distinguish demand avoidance from the ADHD and personality traits it correlates with.
6. Egan V, Linenberg O, O'Nions E (2019). The measurement of adult pathological demand avoidance traits. Journal of Autism and Developmental Disorders, 49(2), 481-494. DOI 10.1007/s10803-018-3722-7. View the paper The paper that created the EDA-QA by rewriting the 26 EDA-Q items in the first person for adult self-report. Study 1 (n = 347) found Cronbach's alpha of .94, an exploratory solution with one dominant general dimension explaining 40.54% of variance and a minor fantasy dimension explaining 6.68%, a single-factor confirmatory model at RMSEA 0.064, and a correlation of r = .40 with a brief autistic traits screen; scores loaded with antagonism, negative affect, disinhibition and psychoticism. Study 2 (n = 191, alpha .92) found low agreeableness, low emotional stability and higher autistic traits predicted EDA-QA scores, which in turn predicted delinquency. Limitation: both samples were online general-population volunteers, predominantly female and highly educated, with self-reported and clinically unconfirmed diagnoses, and no cut-off score or clinical criterion measure was used at any point.
11. Green J, Absoud M, Grahame V, Malik O, Simonoff E, Le Couteur A, Baird G (2018). Pathological Demand Avoidance: symptoms but not a syndrome. The Lancet Child and Adolescent Health, 2(6), 455-464. DOI 10.1016/S2352-4642(18)30044-0. View the paper A review by a group of senior UK autism researchers and clinicians concluding that the available evidence does not support the validity of pathological demand avoidance as an independent syndrome. It accepts that the term names a real and significant cluster of co-occurring difficulties in many autistic children, and proposes that these are better understood through social, sensory and cognitive sensitivities interacting with the environment, with individualised management informed by assessment of co-occurring conditions rather than by a new diagnostic label. Limitation: a viewpoint and narrative review rather than a systematic one, presenting no new data, no meta-analysis and no adult evidence, and it drew published correspondence disputing its reading of the literature.
9. Johnson M, Saunderson H (2023). Examining the relationship between anxiety and pathological demand avoidance in adults: a mixed methods approach. Frontiers in Education, 8, 1179015. DOI 10.3389/feduc.2023.1179015. View the paper Mixed-methods study in which 163 adults completed the EDA-QA alongside measures of anxious and depressive symptoms, intolerance of uncertainty and behavioural inhibition, with 13 taking part in a qualitative phase. The EDA-QA correlated r = 0.58 with the anxiety and depression measure, r = 0.46 with intolerance of uncertainty and r = 0.35 with behavioural inhibition; anxiety alone accounted for 27.8% of variance in demand avoidance and intolerance of uncertainty added a further 1.9%, with the full model reaching 37%. Limitation: a student sample recruited partly through self-identifying PDA channels, cross-sectional, and the authors themselves flag that the EDA-QA's validity is unclear because it assumes PDA is a distinct autistic presentation.
4. Kamp-Becker I, Schu U, Stroth S (2023). Pathological Demand Avoidance - aktueller Forschungsstand und kritische Diskussion. Zeitschrift fur Kinder- und Jugendpsychiatrie und Psychotherapie, 51(4), 321-332. DOI 10.1024/1422-4917/a000927. View the paper A German critical review of the PDA research literature. It records that PDA is recognised in neither ICD-10, ICD-11 nor DSM-5, as neither a standalone diagnostic entity nor an autistic subtype, and concludes that the evidence supports a behaviour profile rather than an independent condition. It documents circular methodology, unexamined confounders and substantial risk of bias in the existing studies, and notes symptom overlap with conduct disorder, ADHD and selective mutism. Limitation: a narrative critical review rather than a systematic one, written in German, focused on children and adolescents, contributing no new data.
12. Kildahl AN, Helverschou SB, Rysstad AL, Wigaard E, Hellerud JMA, Ludvigsen LB, Howlin P (2021). Pathological demand avoidance in children and adolescents: a systematic review. Autism, 25(8), 2162-2176. DOI 10.1177/13623613211034382. View the paper Systematic review of 13 studies of PDA in children and adolescents. Every included study based its assessment criteria on Newson and colleagues' original descriptions, and the review found that the methods used to develop those criteria were never clearly described. Most studies relied exclusively on parent report and did not consider alternative explanations for the observed behaviours, and no study had obtained direct input from the people said to have PDA. The authors conclude that methodological limitations prevent firm conclusions about the consistency of the behaviour or the characteristics of the individuals concerned. Limitation: restricted to children and adolescents, with a small and heterogeneous pool of 13 studies, so it cannot speak to adult self-report measures such as the EDA-QA.
1. Newson E, Le Marechal K, David C (2003). Pathological demand avoidance syndrome: a necessary distinction within the pervasive developmental disorders. Archives of Disease in Childhood, 88(7), 595-600. DOI 10.1136/adc.88.7.595. View the paper The founding clinical account, drawn from referrals to Newson's developmental psychology unit in Nottingham. A discriminant analysis across 150 cases separated the proposed pathological demand avoidance group from autism and Asperger syndrome, an equal sex ratio was reported, and 50 children plus an 18-person young-adult follow-up formed the descriptive core. The authors argued for a category distinct from PDD-NOS. Limitation: a single-centre clinic series with no independent diagnostic standard, no blinding and no replication, and later reviews found the method by which the criteria were derived was never clearly described.
5. O'Nions E, Christie P, Gould J, Viding E, Happe F (2014). Development of the 'Extreme Demand Avoidance Questionnaire' (EDA-Q): preliminary observations on a trait measure for Pathological Demand Avoidance. Journal of Child Psychology and Psychiatry, 55(7), 758-768. DOI 10.1111/jcpp.12149. View the paper The originating instrument paper. Parents of 326 children aged 5 to 17 completed the measure across six groups: 102 typically developing, 36 autistic without disruptive behaviour, 48 autistic with disruptive behaviour, 67 parent-suspected PDA, 50 professional-identified PDA and 23 disruptive behaviour without autism. The final 26 items score 0 to 3 for a maximum of 78; Cronbach's alpha was .87 within the PDA groups and .93 across case groups; a cut-off of 50 for ages 5 to 11 gave sensitivity .80 and specificity .85, and 45 for ages 12 to 17 gave .88 and .78. Limitation: parent report only, no independent clinical verification of the reported diagnoses or PDA suspicions, small disruptive-behaviour comparison groups, and group membership defined by the very description the questionnaire was built from.
14. O'Nions E, Happe F, Viding E, Noens I (2021). Extreme demand avoidance in children with autism spectrum disorder: refinement of a caregiver-report measure. Advances in Neurodevelopmental Disorders, 5(3), 269-281. DOI 10.1007/s41252-021-00203-z. View the paper Principal components analysis and item response theory applied to caregiver reports on 334 autistic children aged 5 to 17 reduced the 26-item EDA-Q to an 8-item measure, the EDA-8, with Cronbach's alpha of .90 and convergent correlations of .66 with reactivity and .49 to .57 with non-compliance measures in a subsample of 233. EDA-8 scores were independent of age, gender, academic level and daily-living independence, and did not correlate with parent-rated autistic severity. Limitation: caregiver report on children only, drawn from an autistic sample rather than a diagnostically mixed one, with no adult version and no external clinical criterion, so it refines the measurement of the trait without testing whether the trait marks a distinct condition.
8. White R, Livingston LA, Taylor EC, Close SAD, Shah P, Callan MJ (2023). Understanding the contributions of trait autism and anxiety to extreme demand avoidance in the adult general population. Journal of Autism and Developmental Disorders, 53(7), 2680-2688. DOI 10.1007/s10803-022-05469-3. View the paper Two general-population adult studies using the EDA-QA scored 1 to 4 (range 26 to 104). Study 1 (N = 267, alpha .93, mean 43.41, SD 12.76) and Study 2 (N = 549 recruited via Prolific, alpha .87, mean 42.88, SD 9.31, aged 18 to 67) both found autistic traits and anxiety to be unique predictors of demand avoidance. Dominance analysis in Study 2 showed neither reliably outranked the other, the difference favouring one over the other appearing in only 25% of simulated runs; the full model including stress, sex and age explained R-squared = 0.53. Limitation: online general-population volunteers rather than clinical samples, data collected in May 2020 during the pandemic when distress scores ran high, cross-sectional so causal direction between anxiety and avoidance cannot be established, and the authors note possible conceptual overlap between the EDA-QA and autism measures.
Clinical frameworks and position statements
7. NovoPsych (2025). Extreme Demand Avoidance Questionnaire - Adult (EDA-QA): review paper. NovoPsych, test publisher documentation. View the source The distributing publisher's own technical documentation, compiling normative reference data for the EDA-QA from two published sources: a community sample of 548 UK adults aged 18 to 67 (mean 42.88, SD 9.31) from White et al. 2023, and a mixed sample of 126 adults aged 18 to 68 recruited from online forums and support groups, 61.5% with an autistic diagnosis (mean 60.52, SD 17.22) from Egan et al. 2020. It reports alphas of 0.87 to 0.94 across studies and a predominantly unidimensional structure with a small secondary fantasy factor, and states plainly that no clinical cut-off scores have been validated for the EDA-QA and that scores should be interpreted dimensionally rather than categorically. Limitation: publisher documentation rather than peer-reviewed research, its percentile bands are derived from two convenience samples on a 26 to 104 scoring convention that differs by 26 points from the 0 to 78 convention used by the original EDA-Q and by most screening implementations.
15. PDA Society (2022). Identifying and assessing a PDA profile: practice guidance. PDA Society, practice guidance, United Kingdom. View the source Consensus practice guidance written by twelve named contributing practitioners and endorsed by a further fifteen, setting out how a PDA profile can be identified within existing diagnostic frameworks. It takes the position that PDA is currently best understood as a profile or cluster of traits on the autism spectrum rather than a separate diagnosis, states directly that the EDA-Q should not be considered a diagnostic test, recommends it mainly as a structure for qualitative information-gathering even when not formally scored, and notes that the sensitivity and specificity difficulties seen with the child measure should be expected with the adult EDA-QA. Limitation: consensus opinion from a self-selected group of practitioners convened by an advocacy organisation, with no systematic evidence review, no formal grading of recommendations and no external methodological appraisal.
Lived experience
3. Kenny N, Doyle A (2024). I have always lived desperate and vulnerable on the edge of helplessness and collapse: a phenomenological exploration of the lived experience of adults experiencing pathological demand avoidance. Neurodiversity, 2, 1-14. DOI 10.1177/27546330241277075. View the source Hybrid thematic and interpretive phenomenological analysis of survey responses from 13 adults identifying with autism and/or PDA, with two in-depth telephone interviews of 45 to 60 minutes. Four superordinate themes emerged: bidirectional social challenges, life experiences marked by low self-esteem and feeling fraudulent, distrust rooted in services that attributed difficulties to attachment or parenting, and relief on encountering the PDA concept. Pervasive anxiety recurred throughout, one participant describing fear as her default state. Limitation: 13 self-identifying participants and only two interviewed in depth, with no diagnostic verification, so the study cannot support any claim that PDA is a distinct autistic subtype and the authors say so explicitly.
2. PDA Society (2018). Being misunderstood: experiences of the pathological demand avoidance profile. PDA Society, survey report, United Kingdom. View the source A two-week online survey run in March 2018 by the UK PDA Society, drawing 1,445 responses including 1,194 parents of children with a suspected or identified PDA profile and 79 adults. Seventy per cent of the 969 young people described were unable to tolerate school or were home educated; 71% of adults and 70% of parents named lack of acceptance and understanding as the principal barrier to support; 67% of 675 parents were dissatisfied with CAMHS and 50% said services had not helped at all. Limitation: a self-selected survey by an advocacy organisation, with PDA status self- or parent-identified and never independently verified, so it documents the strength and consistency of the community's experience rather than prevalence or diagnostic validity.
Further reading — general background
Haire L, Symonds JE, Senior J, Pace U (2023). Practicing psychologists' accounts of demand avoidance and extreme demand avoidance in children and adolescents. Frontiers in Education, 8, 1230014. DOI 10.3389/feduc.2023.1230014. View the source Reflexive thematic analysis of semi-structured interviews with twelve practising psychologists working in the UK, Australia and Ireland, conducted between July and November 2022. Participants largely rejected the necessity of extreme demand avoidance as a distinct construct, attributing the behaviour instead to autism, anxiety and mismatch with the environment, and described working relationally and transdiagnostically rather than behaviourally, mostly through the adults around the child. Included here as further reading on how the debate looks from inside routine practice. Limitation: twelve self-selected practitioners concentrated in one Irish region, with no parent, teacher or child perspectives and no adult caseload, so it describes professional opinion rather than outcomes.
Peer-reviewed = checked by independent experts before publication. Clinical model = an established professional framework, not a single study.
Up next
Module 9 - TAS-20 — Alexithymia, When You Can't Name What You Feel
All modules in Autistic Self-Discovery
The demand is the problem, not you. This course was built by clinicians who are part of the New Path Family. Demand avoidance is one of the areas where a contested label and a very real daily difficulty sit awkwardly together, and it helps to think it through with somebody who knows both the evidence and the experience. Therapy for clients in California and coaching worldwide, all by telehealth, are offered by our sister company New Path Family of Therapy Centers, Inc. A conversation costs nothing and there is no pressure. Saving this for later counts too. Talk with the New Path team