Show the full module text
Module 1 — How to Read a Screener (and What a Score Can't Tell You)
Before the eighteen questionnaires: what a screening score actually measures, the four validity tiers every instrument in this course is labelled with, and why a positive result is a reason to look closer rather than an answer.
Autistic Self-Discovery · Part One — Start Here · 11 min read, about 18 min with the workbook
The big idea
Short on capacity today? The big idea: a screener is a set of questions that compares your answers to a research threshold and tells you whether it is worth looking further. It cannot diagnose you, it cannot measure how autistic you are, and it cannot fail you. Some of the questionnaires in this course are peer-reviewed and validated; some are honest tools we built ourselves. This module tells you which is which. Stop here if that is enough for today.
If you have ever taken an online quiz and then felt worse — either because the number seemed to say something enormous, or because it seemed to say nothing at all — start here. Eighteen questionnaires follow this module. Ten minutes now will change what every one of them means to you.
Step 1 — The lesson
A. What a screener actually is
A screening questionnaire is a net with a fixed mesh size. Researchers took a group of people who had already been carefully assessed, looked at how they answered a long list of questions, and kept the questions that best separated one group from the other. Then they set a threshold: above this line, look further; below it, probably not.
That is the entire mechanism. It is genuinely useful and it is much smaller than people imagine.
Diagram — A · A net with one mesh size. A screener catches what its questions were built to catch. Things it was never designed to notice pass straight through — and they were still there. This is why a low score is information, not an all-clear.
Two consequences follow, and they matter more than any number you will see in this course. A high score does not mean you have the thing. A low score does not mean you do not. The score tells you how much your answers resembled the answers of a particular research group, and nothing more than that.
B. Four tools on one pegboard
Here is the part almost no free-quiz site will tell you: the questionnaires you find online are not all the same kind of object. Some have been through years of independent testing in multiple countries. Some were written thoughtfully by a clinician last year and have never been tested at all. Both can be useful. They are not the same claim.
Every screener in this course carries a tier label, and every module says plainly which tier its instrument sits in.
Diagram — B · Four tools on one pegboard. Four tiers, four different kinds of claim. A Tier 1 instrument has been validated and published; a Tier 4 tool is one we built in-house because nothing suitable existed. Both belong on the board. Only one is a calibrated gauge, and you deserve to know which you are holding.
Tier 1 — validated and published. Built by external researchers, tested against clinical assessment, published in a peer-reviewed journal, and usually re-tested by other teams in other languages. The RAADS-14, AQ-50, GQ-ASC, CAT-Q, RBQ-2A, Monotropism Questionnaire, TAS-20, OCI-R, CPQ and FMPS in this course are Tier 1.
Tier 2 — published, but the construct is contested. The questionnaire exists in the literature, but what it measures is not a settled or formally recognised category. The EDA-QA, which measures demand avoidance, is the one Tier 2 instrument here.
Tier 3 — a clinical checklist in long use, never formally normed. Clinicians have organised conversations around it for years. Nobody has run the validation study.
Tier 4 — a reflection tool we built. When a real experience had no decent questionnaire attached to it — autistic burnout, intimacy, sensory profile — we wrote one. These are honest and they are useful. They are not normed, and we will never pretend otherwise. Where the underlying idea is itself still emerging or debated, the module says that too.
WHY WE SAY THIS OUT LOUD
Under-claiming is the safe direction.
It would be easy to present all eighteen questionnaires with the same confident styling and let you assume they carry the same weight. Plenty of sites do. We would rather you finish this course able to tell the difference — because that skill outlasts any single score, and because a tool you trust for the right reasons is worth more than one you trust by accident.
If you are ever unsure where a questionnaire sits, assume the lower tier. That is the rule we follow ourselves.
C. A mirror, not a verdict
You will meet that phrase on every screener page on this site, and it is not a legal disclaimer bolted on at the end. It is a description of what the object in front of you does.
Diagram — C · A mirror, not a verdict. A mirror shows you something you already had, from an angle you do not usually get. A verdict decides something about you and closes the question. These questionnaires are the first thing and never the second.
The most common way a score does damage is not the score itself — it is the sentence people build around it. "I scored 138, so I am definitely autistic." "I scored 41, so I was making it all up." Both sentences take a rough comparison and turn it into an identity. Neither is supported by the number.
A more accurate sentence is duller and far more useful: "My answers looked more like the group than not, in these particular areas, on this particular day." That is something you can take to a real conversation.
D. One signpost, six paths
People arrive at this course with a question, not a diagnosis. The eighteen questionnaires are not a ladder to climb in order — they answer different questions, and you only need the ones that match yours.
Diagram — D · One signpost, six paths. Each part of this course answers a different question. You do not need all eighteen. Start with the arm that points at the thing you actually came here wondering about.
If the question is am I autistic at all, start with Part Two — the broad screeners. If it is why am I so tired all the time, go to Part Five. If it is why do I feel like I am acting in every room, Part Three. If you do not know what your question is yet, Part Two is still the right door.
INSIDE THE INSTRUMENT
Why a positive screen usually is not a case
This section appears in every module of this course, in the same order, so that a clinician — or a curious reader — can find the same six facts about any instrument in the same place. In this module it covers the arithmetic that governs all of them.
The base-rate problem
Screening accuracy is usually reported as sensitivity (of people who have the thing, what proportion does the screen catch) and specificity (of people who do not, what proportion does it correctly clear). Both can look excellent while the number a person actually cares about — if I screened positive, what is the chance I have it — is poor.
That number is the positive predictive value, and it depends almost entirely on how common the condition is in the group being screened. In a population where 3 in 100 people have a condition, a screen with 90% sensitivity and 70% specificity will return roughly 32 positives, of which about 3 are real. Nine out of ten positive results are false. Nothing is wrong with the screen; that is what screening at a low base rate does.
This is not hypothetical. A systematic review of every adult ADHD self-report measure with published accuracy data found negative predictive values above 96% across the board, while positive predictive values in clinical samples "reached 61% at best, though most fell below 20%."1 An eight-practice primary-care study of the ASRS found a positive predictive value of 0.52 within the sample that fell to 0.12 once adjusted to real population prevalence.2
What that means for the reader in front of you
Two things, and they pull in opposite directions. First: a negative screen is the more informative result. High negative predictive value is the one property these instruments reliably have. Second: the people taking a screener on this site are not an unselected population. They arrived because something already did not fit, which raises the base rate considerably and improves the positive predictive value — while also introducing exactly the confirmation bias that makes a positive feel like proof.
What a screening score legitimately supports
Language for something previously unnamed; a structured way to open a conversation; a record of pattern across domains; and a decision about whether a full assessment is worth the time and cost. It does not establish onset, cross-setting presence, functional impairment, or the exclusion of another explanation — and those are most of what a diagnosis actually requires.
The published position
The UK national guideline is explicit that a diagnosis "should not be made solely on the basis of rating scale or observational data," and describes rating scales as valuable adjuncts only.3 That is the standard this course holds itself to.
Diagram — E · Tuned to over-trigger. Screeners are deliberately built to err toward catching too much, because a missed case costs more than a second look. A smoke alarm that never went off for burnt toast would be a worse smoke alarm. When it sounds, you check. You do not evacuate.
F. What helps
Four habits that will make the next eighteen modules more useful and less destabilising.
1. Answer as the person you are on an ordinary day.
Not your worst week, not your most composed hour. Several of these questionnaires ask about the last six months for exactly this reason. If you find yourself calculating what the "right" answer is, that is worth noticing on its own — and it is more or less what the CAT-Q in Part Three is about.
2. Write the number down and then put it down.
Record the score and the date somewhere you will find it again. Then stop looking at it. Scores are most useful as a set, read weeks later next to each other, and least useful in the twenty minutes after you get one.
3. Read the tier before you read the score.
Every module tells you which of the four tiers its instrument sits in, and it is printed above the questions rather than buried under them. A high score on a Tier 4 reflection tool and a high score on a Tier 1 validated instrument are not the same event.
4. Take the result to a person, not a search engine.
The productive next step after a screener is a conversation — with a clinician, a GP, or someone who knows you well. New Path offers therapy to clients in California and coaching worldwide, all by telehealth, and a first conversation costs nothing. Saving this for later counts too.
Step 2 — Your workbook
Your answers save to this device only — we cannot see a word of what you write. This module has no screener of its own. Instead it asks you to set up how you are going to use the eighteen that follow.
What actually brought you here?
One or two sentences. Not the clinical version — the version you would say out loud.
Fields: The question I am really asking
Which path fits your question?
From the signpost in section D. You can change your mind later.
Fields: Tick: Am I autistic at all? (Part Two); Tick: Am I masking? (Part Three); Tick: Can I name what I feel? (Part Four); Tick: Why am I so tired? (Part Five); Tick: Is this sensory, or something travelling alongside? (Part Six)
Your score log
Come back and fill this in as you go. Reading them together, later, is worth far more than reading any one of them now.
Fields: Instrument, score, date
Before you start: the sentence you will not say
Write out the conclusion you are most tempted to jump to, so you can recognise it when it arrives.
Fields: The sentence I am watching out for
What would a low score actually mean?
Section A: what slips through the net was still there.
Fields: In my own words
Who is the person you would take a result to?
A name, not a category.
Fields: The person; What I would want to say to them
Appendix — Research companion
Peer-reviewed research
1. Harrison AG, Edwards MJ (2023). The ability of self-report methods to accurately diagnose attention deficit hyperactivity disorder: a systematic review. Journal of Attention Disorders, 27(12), 1343-1359. DOI 10.1177/10870547231177470. View the paper Systematic review of 20 studies and manuals reporting diagnostic accuracy for adult self-report screeners. Negative predictive values exceeded 96% across all measures, but positive predictive values in clinical samples reached 61% at best and most fell below 20%. Limitation: heterogeneous samples and reference standards, so the range is indicative rather than a pooled estimate.
2. Hines JL, King TS, Curry WJ (2012). The Adult ADHD Self-Report Scale for screening for adult attention deficit-hyperactivity disorder (ADHD). Journal of the American Board of Family Medicine, 25(6), 847-853. DOI 10.3122/jabfm.2012.06.120065. View the paper 200 adults across eight primary care practices. Sensitivity 1.0, specificity 0.71, negative predictive value 1.0, positive predictive value 0.52 within sample - falling to 0.12 once adjusted to 4.4% population prevalence. Limitation: a single-instrument study, used here to illustrate the base-rate arithmetic that applies to screeners generally.
Clinical frameworks and position statements
3. National Institute for Health and Care Excellence (NICE) (2019). Attention deficit hyperactivity disorder: diagnosis and management (NICE guideline NG87). NICE, London (published March 2018, last updated September 2019). View the source UK national clinical guideline. Recommendation 1.3.2 states that a diagnosis should not be made solely on the basis of rating scale or observational data, and describes rating scales as valuable adjuncts only. Limitation: UK-specific and written for ADHD; cited here for the general principle that a screening score is not a diagnosis.
Further reading — general background
Fayyad J, Sampson NA, Hwang I, et al. (2017). The descriptive epidemiology of DSM-IV adult ADHD in the World Health Organization World Mental Health Surveys. ADHD Attention Deficit and Hyperactivity Disorders, 9(1), 47-65. DOI 10.1007/s12402-016-0208-3. View the paper Twenty nationally or regionally representative surveys, n=26,744, interview-administered. Current adult ADHD prevalence averaged 2.8%. Limitation: DSM-IV criteria and lay-administered interviews; cited here to establish the low population base rate that drives poor positive predictive value.
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 distinguishing deep from shallow compensation, and arguing that because diagnostic criteria are behaviourally defined, successful compensation can conceal underlying differences from observation-based assessment. Limitation: a conceptual review, not an empirical test of any instrument.
Peer-reviewed = checked by independent experts before publication. Clinical model = an established professional framework, not a single study.
Up next
Module 2 - RAADS-14 — The Short Form That Starts the Conversation
All modules in Autistic Self-Discovery
This is easier with someone alongside you. This course was built by clinicians who are part of the New Path family of therapy centers. If a score has left you with more questions than you started with — which is normal, and often the point — there is somewhere to take them. 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