Dyslexia Screening Tools: What a Screener Can Tell You
A parent forwards you a result and asks what it means. A school says the child "passed the dyslexia screener." A colleague recommends a dyslexia screening tool that takes four minutes, and another recommends one that takes forty-five. All four of those sentences use the same word, and they are not describing the same object.
That is the first thing to know about dyslexia screening tools: the word "screener" does not name a kind of test. Nobody owns it, nobody adjudicates it, and no body publishes a list of which products may use it. It is a claim about purpose, attached to instruments that differ from each other more than they resemble each other.
Some of them measure the child directly. One of the best known never touches the child at all. One is not normed. One is free and runs in a browser. One is a composite pulled out of a full achievement battery by a test publisher. They all end in the same sentence, and that sentence is the only thing they genuinely share.
This post is about what each kind actually is, what the published standards say a screener has to do before anyone should trust it, and what you may and may not conclude from a result somebody hands you. It is written for practitioners doing intake, not for families deciding whether to seek one.
A screener does not tell you whether a child has dyslexia. It tells you whether somebody should look again.
What screening means when somebody actually defines it
Two organisations have put a definition in writing, and they agree.
The National Center on Improving Literacy, which is funded by the US Department of Education, defines it in its dyslexia screening report:
"The term screening refers to the brief evaluation of a defined population of individuals to identify the risk for performing below a specified threshold or benchmark on a specified outcome."
Three things are doing work in that sentence. Brief. A defined population, which means everybody in a grade, not a referred child. And risk of falling below a threshold on an outcome, which is a prediction about future reading, not a statement about a present condition.
The International Dyslexia Association's Universal Screening: K-2 Reading fact sheet says the same thing from the other side:
"Screening measures, by definition, are typically brief assessments of a particular skill or ability that is highly predictive of a later outcome."
Note what neither definition contains. Neither says a screener measures dyslexia. Both describe a measure chosen because it predicts something else well, which is a completely different property from measuring the thing you care about.
NCIL is explicit about the boundary on the same page:
"The purpose of universal screening for dyslexia risk is very different than the purpose of diagnosing dyslexia... It is not appropriate to use screening results to formally diagnose whether an individual actually has dyslexia."
Hold that sentence. Every object below is sold under a word whose own definition says this.
Seven objects, one word
Here is what is currently on the market and in use, all of it described by somebody as a dyslexia screener or a dyslexia screening tool. The differences are not shades. They are differences in what gets measured, by whom, and whether the child is present.
1. The teacher rating scale that never tests the child
Pearson's Shaywitz DyslexiaScreen is probably the best-known dyslexia-branded screener in American schools. It is a rating scale. The teacher answers a short set of items about a student they already know, and the student does nothing.
Pearson's own FAQ explains the design choice directly: "classroom teachers are those professionals closest to the student academically." There are six forms, Forms 0 through 3 for kindergarten through grade three with ten to twelve items each, plus an adolescent-adult self-rating form and a corrections form of ten items each. Administration takes, in Pearson's words, "just a few minutes per rating."
It is a Level B product. The FAQ lists the professionals who may use it: "special education teachers, speech-language pathologists, psychologists, occupational therapists, reading specialists, educational diagnosticians."
And it says plainly that "a screener does not diagnose," and that an "At Risk for Dyslexia" classification "should be followed by a secondary measure or measures to establish whether or not the student actually has dyslexia."
That is a defensible instrument, well documented, with its limits in print. It is also not a reading test. If a referral arrives describing a child who "failed the dyslexia screener," it is worth knowing whether anybody has heard the child read.
2. The composite lifted out of a full achievement battery
Pearson also publishes a Dyslexia Index for the WIAT-4, with an equivalent for the KTEA-3. This is a different animal entirely: an individually administered set of subtests drawn from a full achievement battery, scored as a composite.
The composition changes by grade. For grades PK through 3 it is Phonemic Proficiency and Word Reading. For grades 4 through 12 and above it is Word Reading, Orthographic Fluency and Pseudoword Decoding. Administration runs about five minutes for the younger band and under twenty for the older. The range is ages four to fifty. It is Level B.
Pearson positions it as something that "may be used as an evidence-based dyslexia screener to identify students who may benefit from a more comprehensive evaluation or assessment battery, or for research."
And then it writes the limit into the product page itself:
"A single score such as the Dyslexia Index is not sufficient to diagnose dyslexia. Rather, a diagnosis of dyslexia is based on a convergence of evidence gathered from multiple sources."
This matters for a reason that is easy to miss. A Dyslexia Index is not an extra test. It is a scoring option on an achievement battery you may already be using. If the referral paperwork names a Dyslexia Index, the child sat parts of the WIAT-4 or the KTEA-3, and the full battery's other subtests may exist somewhere.
That is a question worth asking at intake, and it connects directly to what you may do with a report somebody else wrote, which is covered in our post on reading a psychoeducational evaluation.
3. The universal screening battery that is not about dyslexia at all
This is the largest category by use and the one least often called a dyslexia screener by its own publisher. DIBELS, Acadience Reading, aimswebPlus, easyCBM, FastBridge and the rest are brief curriculum-based measures administered to everybody, two or three times a year, producing a composite that sorts students into risk bands.
They are general reading screeners. A district uses one of them and then describes the process as dyslexia screening, which is a reasonable thing to do operationally and a misleading thing to say about the instrument. The composite predicts reading difficulty. It does not isolate a profile.
IDA's fact sheet sets out which skills belong in such a battery and when. In kindergarten: phonological awareness, rapid automatic naming, letter-sound association and phonological memory. In first grade: phoneme awareness, letter naming fluency, letter-sound association, phonological memory, oral vocabulary and word recognition fluency. In second grade: word identification with real and nonsense words, oral reading fluency and reading comprehension.
NCIL's report sets out the cadence: screen all students in kindergarten through grade three twice a year, administered "as early as possible... within two weeks to one month of the start of school."
If you work with the output of these systems, the comparison in our post on progress monitoring tools covers what the free educational use agreements actually permit, and why a benchmark in one system is not a benchmark in another.
4. The free measure that is not normed
David Kilpatrick's Phonological Awareness Screening Test, the PAST, is a criterion-referenced measure of phoneme manipulation, descended from the Rosner and Simon Auditory Analysis Test. It is distributed free, with forms and reliability and validity data published alongside.
Two of its properties routinely surprise people. It is not a normed test, and its author says so in the instructions. And every item is timed: automatic credit requires a correct response in two seconds or less, because the construct of interest is automaticity, not whether the child can get there eventually.
That second property is the whole point of the instrument and the thing that gets lost when it is used casually. An untimed administration of the PAST is a different measure that happens to use the same items.
A measure can be useful and unnormed at the same time, as long as nobody reports it as a standard score. The trouble starts when an unnormed criterion measure is summarised in a report in the grammar of a normed one. Our post on grade equivalent scores covers the adjacent version of that failure.
5. The game on the tablet
Imagine Learning's Imagine+ Screener, formerly EarlyBird, is the one tool on the federal screening chart whose own subtitle carries the word: it appears there as "Dyslexia and Early Literacy Screener," rated for kindergarten, grade one and grade two.
Imagine Learning's own overview describes it as "a literacy screener for PreK-Grade 3 that measures foundational literacy skills," administered "one-on-one or in a small group with adult oversight," and says it "helps educators identify students who may be at risk for reading difficulties, including characteristics of dyslexia."
Then there is the timing, and it is worth sitting with. Comprehensive screening takes "about 20 minutes in PreK, 45 minutes in Kindergarten, and 30 minutes in Grades 1-3."
Forty-five minutes per kindergartener is a serious and defensible assessment design. It is also four or five times what both published definitions of screening mean by "brief." The instrument is not worse for that. But a practitioner reading "we screened the kindergarten" should know that the word covers both this and a four-minute teacher rating.
Nessy's Dyslexia Quest sits in the same category and is explicit about the design intent. It is a set of game-based tests of rapid automatic naming, working memory, phonemic awareness and phonological awareness, embedded in a game, built so that "children don't realise they are being tested." It runs for ages five to eleven with adult supervision, and most children finish inside twenty minutes.
Its limits statement is the bluntest of any product quoted here, and it names a consequence the others leave out: "screeners do not give a diagnosis, and can not be used for examination accommodations." What it offers instead is "an indication of dyslexia tendencies."
That second clause is worth repeating to anybody who hopes a screening result will unlock testing accommodations. It will not, and the vendor says so.
6. The free web test a family runs at the kitchen table
Lexercise's free dyslexia test combines what it describes as the San Diego Quick Assessment of grade-level word reading with a proprietary "Dyslexia Z-Screener." It takes five to ten minutes, is administered by an adult to the child rather than self-completed, and is scored live by whoever is administering it.
To its credit, the page states the limit in its own words: "Screeners are not diagnostic or descriptive tools."
The reason to know this category exists is that its results arrive in your inbox. A family who has run one has a number, a word and an expectation, and very often no understanding of what was and was not measured.
Naming the instrument and explaining what it sampled is a useful first ten minutes of an intake, and it is also an ethics question about what you are implicitly endorsing. Our post on informed consent across five codes covers the surrounding machinery.
7. The clinician's judgement, called a screener
The last category has no product page. It is the informal screen: a practitioner spends twenty minutes with a child, listens to them read a passage and a nonsense word list, asks about family history, and forms a view. People call this screening, and it is the oldest version of the practice.
There is nothing wrong with it as a clinical act, and it is often more informative than a ten-item rating scale. It has no classification accuracy data of any kind, which means it cannot be defended in the terms the next section uses, and it should never be described in a report as though it were an instrument.
Nobody is guarding the word
Across those seven, there is no shared construct, no shared administration method, no shared duration, no shared scoring model and no shared evidence requirement. Four of them do not require the child to read anything. One of them is not a test.
There is no certification body for the term, no registry of approved dyslexia screeners, and no professional association that rules on whether a product may use the word. The nearest thing to a public arbiter is a chart that deliberately refuses the role, which is the subject of the next section.
The word "screener" is a statement of intent, not a specification. Two products can both be accurate descriptions of the word and share nothing else.
The one published bar, and what it is
The National Center on Intensive Intervention maintains an Academic Screening Tools Chart, last updated August 2026. It is the closest thing the field has to a public technical review of screening instruments, and its rating rubric is published in full.
NCII's own description of the chart explains where the criteria come from: "The Center's Technical Review Committee (TRC) on Screening independently established a set of criteria for evaluating the scientific rigor of screening tools." That independence is the point. The criteria were not negotiated with the vendors whose products are rated against them.
For the top rating on classification accuracy, a tool must satisfy three threshold questions and clear three numbers:
"All of Q1 - Q3 (below) rated as YES and the lower bound of the confidence interval around the Area Under the Curve (AUC) estimate is at least 0.80 and Sensitivity is at least 0.80 and Specificity is at least 0.80."
The three questions are worth reading on their own, because they are where most products fail before the numbers are even reached:
"Was an appropriate external measure of academic performance used as an outcome?"
"Was risk adequately defined within an RTI approach to screening (i.e., 10th - 20th percentile)?"
"Were the classification analyses and cut-points adequately performed?"
NCIL's dyslexia report endorses the same figures independently: sensitivity and specificity each at least .80, and area under the curve at least .80.
So there is a bar, it is public, it is numeric, and two federally funded centres agree on it. The difficulty is what sits underneath it.
What the chart is, and what it refuses to be
The chart states its own limits in the clearest language of any document in this post:
"The presence of a particular tool on the chart does not constitute endorsement and should not be viewed as a recommendation. All tools that meet the criteria for review are posted on the chart, regardless of results. The chart represents all tools that were reviewed, not those that were 'approved.'"
Read that twice if you have ever seen a vendor cite chart presence as a credential. Appearing on the chart means a tool was submitted and reviewed. It is compatible with an empty bubble in every column.
The chart also defines its central statistic plainly:
"AUC values closer to 1 indicate the screening measure reliably distinguishes among students with satisfactory and unsatisfactory reading performance, whereas values at 0.50 indicate the predictor is no better than chance."
And what the chart does not contain
This is the finding that changes how the bar should be read. The chart is an academic screening chart. It rates general reading and mathematics screeners. Exactly one tool on it carries the word dyslexia in its name, and that is the Imagine+ EarlyBird entry described above.
The Shaywitz DyslexiaScreen is not on it. Neither is any free web screener, nor the PAST, nor a Dyslexia Index.
That is not an accusation against any of those instruments. Submission to the chart is voluntary, several of them are not universal screeners in the sense the chart reviews, and absence of a rating is not a poor rating.
But it means the one public bar in the field is almost entirely unapplied to the products that carry the field's name, and a practitioner cannot close that gap by looking something up. The same structural problem runs through reading program evidence ratings, where five public systems rate programs and none of them rates what a practitioner is actually choosing between.
Sensitivity, specificity, and the choice nobody tells you you are making
NCIL's seven factors for choosing a literacy screener defines the pair in one sentence: "Sensitivity is correctly identifying those who will develop a disability, and specificity is correctly identifying those who will not."
A false positive flags a child who turns out fine. A false negative misses a child who does not. You cannot minimise both at once; moving the cut score trades one for the other.
NCIL's report takes a position on which way to err:
"The consequences of a false negative error... is perhaps more serious because of the need for students... to receive explicit reading instruction as early as possible."
The seven-factors guidance says the same thing operationally, and adds the constraint everyone discovers eventually: a system should prioritise catching every at-risk student, accept the extra false positives that follow, and then adjust the threshold to the resources actually available to follow up.
That last clause is the honest part. A cut score is partly a psychometric decision and partly a statement about how many children a programme can afford to look at twice. When somebody tells you a screener "over-identifies," they may be describing a validity problem, or they may be describing a staffing problem.
And the Shaywitz FAQ is unusually candid about the consequence at the level of the individual child: "a few students who are flagged as 'at risk' will end up not having dyslexia upon further assessment. They may need something else, or nothing beyond excellent Tier 1/general education instruction."
What this means for a child in front of you
A flagged child is a child with an elevated prior probability, nothing more. A child who was not flagged is not a child who has been cleared. Both of those sentences follow from the arithmetic, and both of them are regularly over-read in the direction the reader was already leaning.
What "brief" turns out to mean
Lay the durations side by side and the range is hard to defend as one category.
Shaywitz DyslexiaScreen: "just a few minutes per rating"
WIAT-4 Dyslexia Index: about five minutes for PK through 3, under twenty for 4 through 12 and above
Lexercise free test: five to ten minutes
Imagine+ Screener, comprehensive: about twenty minutes in PreK, forty-five in kindergarten, thirty in grades one to three
A four-minute rating scale and a forty-five-minute battery are both "brief" relative to a full evaluation, and that is the only sense in which the word covers both. When a school or a parent reports that screening was done, the duration is one of the fastest ways to work out which object you are dealing with.
What every publisher says happens next
This is where the seven converge, and it is the most useful single observation in the post. Every one of these instruments, in its own publisher's or author's words, ends the same way.
Pearson on the Shaywitz: a positive classification "should be followed by a secondary measure or measures."
Pearson on the Dyslexia Index: "a diagnosis of dyslexia is based on a convergence of evidence gathered from multiple sources."
NCIL: "It is not appropriate to use screening results to formally diagnose whether an individual actually has dyslexia."
IDA: a "formal clinical evaluation is necessary to determine a diagnosis of dyslexia if the student continues to struggle with literacy skills, despite high-quality instruction."
Lexercise: "Screeners are not diagnostic or descriptive tools."
Nessy: "screeners do not give a diagnosis, and can not be used for examination accommodations."
The British Dyslexia Association: screening tests "can only give an indication of possible dyslexic difficulties."
Seven objects that share no construct, no method and no duration, and they all terminate in the identical instruction. That is the actual definition of the category, and it is behavioural rather than psychometric. A screener is an instrument whose only authorised output is a decision about whether to look again.
Can a screener identify dyslexia? No.
It is worth answering the question flatly, because it is the question that arrives most often and the one that practitioners get asked to answer on the spot.
No screener identifies dyslexia. Not the rating scale, not the composite, not the battery, not the game, not the free web test. Every document quoted above says so, including the documents written by the companies selling the instruments.
What a screener can do is raise or lower the probability that a comprehensive evaluation would find something. That is a genuine and valuable contribution, and it is a different contribution from the one the word implies to a family.
For educational therapists specifically, there is a second boundary underneath the first. The AET Code of Ethics directs the educational therapist to seek interpretation from related-field professionals, and the AET Fact Sheet lists administering cognitive, intelligence or psychological tests among the things an educational therapist does not do, with the parenthetical "(unless otherwise qualified to do so)" intact.
Administering a brief literacy screener is not the same act as diagnosing, and the codes treat them differently. We cover how those two AET documents fit together, and where they disagree, in our scope map of educational therapy.
Nothing here is legal advice. State screening mandates exist, they vary, they change, and they are outside the scope of this post entirely. If your question is what a statute or district policy obliges you or a school to do, that question belongs with your state education agency and, for independent practice, with an attorney.
What a qualification level tells you and what it does not
Two of the instruments above are Level B. It is tempting to read that as a difficulty rating or a rigour rating. It is neither.
A qualification level is a purchasing rule set by one publisher for its own storefront. It is not an industry scale, the letters do not mean the same thing across publishers, and one publisher's Level A can mean there are no special qualifications to purchase at all. We worked through the whole mess, publisher by publisher, in achievement test qualification levels.
For screening specifically the implication is narrow and useful. The fact that you can buy a dyslexia screener tells you nothing about whether you are qualified to interpret what it produces, and the two Level B products discussed here both say in print that their output requires something further.
The same word means something different in Britain
Practitioners read across borders, and the British material is both excellent and built on a different model, which is a reliable source of confusion.
The British Dyslexia Association draws the line like this:
"The only way that dyslexia can be formally identified is by a Diagnostic Assessment for Dyslexia carried out by a qualified assessor."
Screening in that system "can only give an indication of possible dyslexic difficulties," may be administered by a trained SENCo or online, and is followed by a full diagnostic assessment when the probability returned is moderate or high. The diagnostic assessment "will determine the precise nature of the difficulties, and how they can best be supported," and is carried out by specialist teachers or psychologists holding a specific assessment qualification.
The underlying logic is the same as the American one. The vocabulary is not interchangeable, and a UK screening report describing a "moderate probability of dyslexic difficulties" is a different statement from an American report describing a child as "at risk."
How to read a screener's claim in four questions
When somebody hands you a result, or a vendor hands you a brochure, these four questions separate the categories quickly and none of them requires psychometric training.
1. Who or what was measured? The child's reading, the child's phonological processing, or an adult's opinion of the child. All three are legitimate inputs. Only two of them involve the child doing anything.
2. Against what was risk defined? A screener predicts something. Ask what. NCII's second threshold question is exactly this, and it asks specifically whether risk was defined at the tenth to twentieth percentile on an external measure. A tool that cannot answer it has not been validated as a predictor at all, whatever its reliability looks like.
3. What classification accuracy is published, and on whose sample? Sensitivity, specificity and AUC, with confidence intervals, on a sample that resembles the children you see. NCIL's third factor is about exactly this: check whether the norming group included students like yours.
4. What does the publisher say happens next? If the answer is anything other than further assessment, the product has over-claimed relative to every standard quoted in this post.
What this changes about an intake
Very little of this changes what you do in the room. It changes what you write down and what you say out loud.
When a screening result arrives, name the instrument in your notes rather than the word. "Teacher rating scale, ten items, completed by the classroom teacher in March" is a record. "Passed the dyslexia screener" is not, and in six months neither you nor anyone else will be able to reconstruct what it meant.
When a family asks what the result means, give them the publisher's own sentence. It is more accurate than any paraphrase and it has the advantage of being quotable back to the school.
And when you are deciding whether to recommend further assessment, remember that the screener has already done its only job. It has told you whether to look again. The decision about what the looking consists of is yours and the evaluating professional's, and it depends on the child, not on the instrument that raised the flag.
The decision sequence runs: identify which of the seven objects produced the result; establish what it measured and what it did not; decide whether what it did not measure matters for this child; and route the remaining question to whoever is qualified to answer it. That sequence is the deliverable. The decision itself is not something a post can make for you.
Where this sits in the rest of the library
Several of the threads above are worked out at length elsewhere. The letters on a test publisher's storefront are a purchasing rule and not a scale, which is the subject of achievement test qualification levels.
Two comprehension scores on the same child can disagree for reasons that have nothing to do with the child, in reading comprehension assessment tools. The ten-year rule about test norms appears in none of the documents practitioners cite for it, in outdated test norms.
The phrase "Orton-Gillingham based" is a claim no organisation adjudicates, in what that label actually means — the same structural finding as this post, applied to intervention rather than assessment. "Aligned to the IDA Knowledge and Practice Standards" turns out to mean four different things, in the IDA standards post. And the question of whether the service model itself has an evidence base is handled in is educational therapy evidence-based.
On the intervention side, what follows a confirmed reading difficulty is covered in reading intervention programs compared and, for the training routes behind them, structured literacy training programs compared. For the writing side of the same intake, see written expression assessment tools and dysgraphia and writing intervention tools.
For mathematics, math assessment tools for dyscalculia makes the parallel argument about a word that does not name anything you can score. And if you are still working out where your own role ends relative to the neighbouring ones, start with educational therapist versus learning specialist versus executive function coach.
Where the published record runs out
Four things in this area could not be established from publicly readable sources this week, and it is better to say so than to fill the gaps.
Classification accuracy figures for most dyslexia-branded products are not public. Pearson's FAQ reports that manual tables show "Good to Excellent levels of sensitivity, specificity, and AUC" for the Shaywitz, but the numbers themselves live in the manual, not on the product page. The same is true of the Dyslexia Index. You cannot compare two products on the one dimension that matters most without buying both manuals.
Proprietary screeners do not publish their algorithms. The Lexercise "Dyslexia Z-Screener" is described as proprietary. What it computes, on what sample it was validated and against what criterion are not stated on the public page.
The NCII chart's per-tool ratings change. The chart was last updated August 2026, and ratings are added and revised as vendors submit data. Anything you read here about which tools are and are not on it should be checked against the live chart before it goes in a report.
State screening mandates are deliberately absent from this post. They are the single most common reason a practitioner encounters a dyslexia screener, and they are legal territory. Nothing above should be read as describing what any jurisdiction requires.
Frequently asked questions
Can a dyslexia screener diagnose dyslexia?
No. Every source quoted in this post says so, including the test publishers. NCIL states that "it is not appropriate to use screening results to formally diagnose whether an individual actually has dyslexia," and Pearson writes on its own product page that "a single score such as the Dyslexia Index is not sufficient to diagnose dyslexia."
What is the difference between a dyslexia screener and a dyslexia evaluation?
A screener is brief, given to a whole defined population, and predicts risk of a future outcome. An evaluation is lengthy, individually administered, and describes present functioning across multiple domains. IDA frames the distinction as brief and predictive versus extensive and descriptive.
How accurate does a dyslexia screening tool have to be?
The one published bar comes from the National Center on Intensive Intervention and is echoed by the National Center on Improving Literacy: sensitivity at least .80, specificity at least .80, and the lower bound of the confidence interval around the AUC at least .80, alongside three threshold questions about the outcome measure, the definition of risk and the cut-point analysis.
Is a teacher rating scale a real dyslexia screener?
Yes, in the sense that it is a validated instrument sold and documented as one, and its publisher states its limits clearly. It is not a reading test, and the child does not read anything during it. Both of those facts belong in your notes when a rating-scale result arrives.
Why did a screener flag a child who turns out not to have dyslexia?
Because cut scores trade false positives against false negatives, and the published guidance deliberately favours catching every at-risk student even at the cost of extra flags. Pearson's own FAQ notes that some flagged students "may need something else, or nothing beyond excellent Tier 1/general education instruction."
Should an educational therapist administer a dyslexia screener?
That depends on the instrument's qualification requirements, your own training, and the ethical boundary between administering a measure and interpreting it. The AET Code of Ethics directs educational therapists to seek interpretation from related-field professionals, and the AET Fact Sheet reserves cognitive, intelligence and psychological testing unless otherwise qualified.
A brief literacy screener is a different act from a psychoeducational evaluation, and the distinction is worth making explicitly in your own scope statement.
Join the Illuminate educator network
Illuminate Educational Therapy Group matches students with educational therapists, learning specialists and executive function coaches. If you work with students who arrive carrying a screening result somebody else produced, and you would like referrals from a group that understands the difference between a flag and a finding, join the educator list.
Comments