Math Assessment Tools for Dyscalculia: What They Measure
You have a student who cannot hold a number in mind long enough to do anything with it, and a family who arrived at your intake with the word "dyscalculia" already in their mouth. So you go looking for the math assessment tools dyscalculia gets diagnosed with, the way you would go looking for a dyslexia screener, and you find something odd.
There is no equivalent.
On the reading side there is a construct with a name, a screening layer built for it, and a purpose-built composite bolted onto two of the biggest achievement batteries in the country.
On the math side there is a small shelf of instruments that measure how much arithmetic a student has learned, sold at four different qualification letters by five different publishers, and two federal practice guides that never use the word at all.
That gap is not a hole in your knowledge. It is the actual state of the field, and it changes what you should be asking. This is a map of what exists, what each instrument claims to measure, who is permitted to buy it, and where the published record simply stops.
Nothing here is legal advice, and nothing here is diagnostic guidance.
The word you are searching for is not the name of anything you can score
Start with the awkward fact, because everything else follows from it.
When the American Psychiatric Association assembled DSM-5, it considered "dyscalculia" as a disorder name and declined. Its own summary document is blunt about the reason: "The DSM-5 Neurodevelopmental Work Group concluded that the many definitions of dyslexia and dyscalculia meant those terms would not be useful as disorder names or in the diagnostic criteria." The diagnosis that exists is specific learning disorder with impairment in mathematics.
The APA has not abandoned the word. Its own public-facing page uses "dyscalculia" descriptively, for difficulties "with learning number-related concepts, with processing numerical information, with learning arithmetic facts or with using the symbols and functions to perform accurate or fluent math calculations." But the specifier is what gets written down.
The word everyone searches for is a description, not a score. No instrument on the market produces it, and the instrument that comes closest produces three index scores instead.
That matters for you in a practical way. A test can tell you a student's standard score on calculation. It cannot tell you the student has dyscalculia, because dyscalculia is not an operationalised outcome that any of these instruments was built to yield.
The four axes that are safe to compare on
This site has used the same four axes for every instrument comparison it has published, and they hold here for the same reason they held for executive function assessment tools: each one is a matter of published fact rather than clinical judgement.
What the publisher says the instrument measures — in the publisher's words, from the publisher's page.
Who supplies the data — direct performance, rating scale, or a mix.
The age or grade band, exactly as published.
The qualification level required to purchase it, and from which storefront.
What is deliberately not on that list is which instrument to reach for given a described student. That is a clinical decision, and the honest version of this article gives you a sequence rather than an answer. The same discipline governed the achievement test qualification levels piece, and it is doing more work here than usual, because the math shelf is short enough that a recommendation would function as an endorsement.
The instruments built for mathematics alone
Six instruments are worth knowing by name. Two are early-years, three are school-age batteries, and one is the only widely distributed product that organises itself around dyscalculia at all.
KeyMath-3 Diagnostic Assessment
Pearson's KeyMath-3 DA is the instrument most educational therapists have heard of. Pearson's own description: "an individually administered measure of conceptual mathematics knowledge, computational skills, and problem solving for students in kindergarten through age 21."
Ten subtests across three content areas — Basic Concepts, Operations, Applications. Ages 4:6 to 21:11, grades K to 12, authored by Austin J. Connolly and published in 2007. Pearson lists it at qualification level B.
Note the date. It is the oldest instrument on this list by a wide margin, and it predates the DSM-5 decision described above by six years.
Test of Early Mathematics Ability, Third Edition
TEMA-3, from PRO-ED, covers ages 3-0 through 8-11 in about forty minutes, individually administered. Ginsburg and Baroody, 2003. It measures, in the publisher's framing, numbering skills, number-comparison facility, numeral literacy, mastery of number facts, calculation skills and understanding of concepts.
PRO-ED lists the TEMA-3 at Test Level B. WPS, which also distributes it, also lists it at Level B. Hold that thought — it is the only instrument in this article where two storefronts agree.
Test of Mathematical Abilities, Third Edition
TOMA-3, Brown, Cronin and Bryant, 2013. Ages 8-0 to 18-11, sixty to ninety minutes, and administrable in a group as well as individually. Four core subtests — Mathematical Symbols and Concepts, Computation, Mathematics in Everyday Life, Word Problems — plus a supplemental Attitude Toward Math scale, which is the only formal affective measure on this shelf.
PRO-ED describes it as "an easily administered, norm-referenced, assessment tool used to identify, describe, and quantify mathematical deficits in school age children." Its technical white paper reports a normative sample of 1,456 students across 21 states, and states the user requirement in a single sentence: "The TOMA-3 can be used by any professional with some formal training in standardized test administration."
It is listed at Test Level A. A 2006 review of its predecessor sits in the ERIC record for anyone who wants the independent read.
Comprehensive Mathematical Abilities Test, Second Edition
CMAT-2, Hresko, Herron and Sherrbenou, 2022 — the newest instrument here. Ages 8-0 to 18-11. Six core subtests (Addition, Subtraction, Multiplication, Division, Problem Solving, and Charts, Tables, and Graphs) and six supplemental ones (Algebra, Geometry, Rational Numbers, Time, Money, Measurement). PRO-ED says it assesses "a broad spectrum of mathematical abilities in the areas of calculation, reasoning, and application." Scoring is online only.
It is also listed at Test Level A.
Feifer Assessment of Mathematics
The FAM, by Steven G. Feifer, is the outlier and the reason this article exists. PAR describes it as an instrument that "examines the underlying neurodevelopmental processes that support proficient math skills" and helps identify subtypes of dyscalculia.
Nineteen subtests, administered in tiers: nine at pre-K, fourteen from kindergarten to grade 2, eighteen from grade 3 upward. It yields three indices that are the closest thing in commercial publishing to a dyscalculia typology:
Verbal Index — automatic fact retrieval and the linguistic components of mathematics.
Procedural Index — counting, ordering, sequencing and mathematical procedures.
Semantic Index — visual-spatial and conceptual components, magnitude representation, patterns and number sense.
There is also a FAM Screening Form, fifteen minutes to administer and fifteen to score, which PAR says is "designed to identify children at risk for dyscalculia or a math learning disability (MLD) within an RTI delivery system model" and to help decide "whether the risk of dyscalculia or an MLD is high enough to warrant further assessment."
The qualification letters on this product are where the article turns. They are covered in their own section below.
Number Sense Screener
The Number Sense Screener, from Brookes, is the best-known standalone early-numeracy screener in the United States, authored by Nancy C. Jordan and Joseph J. Glutting with Nancy Dyson, published in 2012.
It is worth being precise about what it is. The full published title is "Number Sense Screener, K–1, Research Edition" — and it is still a research edition fourteen years later.
The user's guide reports three norm groups: fall of kindergarten, spring of kindergarten, and fall of first grade. Internal-consistency reliability across those groups runs from .82 to .87, with an averaged value of .85, calculated on a development sample of 425 children.
Those are respectable numbers for a twenty-nine-item screener. They are not the national benchmark architecture that sits behind the reading screeners a practitioner is used to, and the "Research Edition" label on the cover is the publisher saying so.
The mathematics inside the general achievement batteries
Most educational therapists will meet math scores not through any of the above, but inside a comprehensive battery administered by somebody else — which is the situation the psychoeducational evaluation article maps in full.
WIAT-4
The Wechsler Individual Achievement Test, Fourth Edition, published 2020, ages 4:0 to 50:11, level B. Its published subtest and composite structure puts two subtests in the Mathematics composite — Numerical Operations and Math Problem Solving — plus a separate Math Fluency composite built from three timed subtests: Addition, Subtraction and Multiplication.
KTEA-3
The Kaufman Test of Educational Achievement, Third Edition, 2014, ages 4:0 to 25:11, level B. Three math subtests: Math Concepts and Applications, Math Computation, and Math Fluency.
WRAT5
The Wide Range Achievement Test, Fifth Edition, 2017, ages 5 to 85 and over, level B, is the brief one — fifteen to twenty-five minutes for ages 5 to 7 and thirty-five to forty-five from age 8 up. One of its four subtests is Math Computation, which Pearson describes as measuring "an individual's ability to count, identify numbers, solve simple oral math problems, and calculate written math problems with a time limit."
It is a single timed computation subtest inside a four-subtest screener. That is worth holding in mind, because it is one of the two instruments Pearson's own dyscalculia toolkit lists under screening.
PAL-II Math
Process Assessment of the Learner, Second Edition: Diagnostics for Math covers grades K to 6 at qualification level B.
Its published flyer makes the strongest claims on this shelf: that it "measures not only basic math skills, but also the development of cognitive processes that are critical to learning math skills," that it "introduces novel quantitative and spatial working memory tasks associated with math computation skills," and that it "helps in formulating a diagnosis of math disability."
Two things there deserve a practitioner's attention.
The flyer's own headline calls PAL-II Math "the only test to investigate the cognitive processes related to math." The FAM, from a different publisher, describes itself as examining "the underlying neurodevelopmental processes that support proficient math skills." Both cannot be the only one. Marketing superlatives are not claims either publisher would defend in a technical manual, and they are a reason to read the manual rather than the flyer.
And "helps in formulating a diagnosis" is a sentence written for the psychologist the level-B tier contemplates, not a warrant for anyone who clears the purchase screen. That distinction is the whole of the next-but-one section.
Woodcock-Johnson V
Riverside Insights' Woodcock-Johnson V, sold here, is the third of the big batteries. Riverside is the publisher that does not play the letter game at all.
Its conditions of sale organise purchasers by category — educational institutions, teachers, hospitals and private practices, individuals and home schools, university students — rather than by an alphabetic tier, and require that "all first-time test purchasers to furnish evidence of their qualifications to use tests by either completing a Test Purchaser Qualification Form or by creating an account at riversideinsights.com."
This is the same finding the qualification-levels article reached from the other direction: there is no industry scale. There are publishers' policies.
One instrument, three letters, two storefronts
Here is the sharpest thing in this article, and it is a purchasing fact rather than a clinical one.
The Feifer Assessment of Mathematics is sold by two publishers. PAR lists the FAM at qualification level B. WPS lists the same instrument at Level C. The FAM Screening Form, at PAR, is level S.
Those letters are not decorative. PAR's published definitions set Level B at "a degree from an accredited 4-year college or university in psychology, counseling, speech-language pathology, or a closely related field plus satisfactory completion of coursework in test interpretation, psychometrics and measurement theory, educational statistics, or a closely related area."
Level C requires "all qualifications for level B plus an advanced professional degree." Level S is a separate track for holders of "a degree, certificate, or license to practice in a health care profession or occupation."
So on one storefront a bachelor's degree plus a measurement course qualifies you to buy the full FAM, and on the other you need a doctorate for the identical kit.
The same instrument, describing the same three dyscalculia indices, requires a bachelor's degree at one distributor and a doctorate at another. Neither publisher is wrong. There is nothing for them to be wrong about.
This is the second confirmed instance of the pattern. The first was the TOWL-4, documented in this site's dysgraphia and writing intervention tools article and extended in the achievement test qualification levels piece: level B at PAR and Pearson, Level C at WPS. Two instruments, same shape, same two publishers on opposite sides.
And the letters do not even mean the same thing across publishers. Pearson's Level B requires "a master's degree in psychology, education, speech language pathology, occupational therapy, social work, counseling, or in a field closely related to the intended use of the assessment." PAR's Level B requires a bachelor's.
The same letter, one tier apart. Pearson also publishes a separate page on purchasing as a qualified user, which is worth reading once before you assume your credential clears a level.
Which makes Test Level A the number worth staring at
PRO-ED's qualifications policy defines its lowest tier in one sentence: "There are no special qualifications to purchase these products."
The TOMA-3 and the CMAT-2 are both Test Level A.
Read that against PRO-ED's own description of the TOMA-3 as a tool "used to identify, describe, and quantify mathematical deficits in school age children." A norm-referenced instrument whose stated purpose is to quantify a deficit is purchasable with no credential check at all, while an instrument that produces three index scores about the same domain requires a doctorate from one of its two distributors.
Nothing in that arrangement tells you what you are competent to do. It tells you what a shopping cart will let you complete, which is a different question and a much smaller one.
The screening layer, and what is actually in it
If you want to know why the reading and math shelves feel so different, compare the screening layers rather than the batteries.
The NCII academic screening tools chart carries mathematics screeners — Acadience Math, aimswebPlus Math, easyCBM, FastBridge, SpringMath, i-Ready, MAP Growth, Star, mCLASS, ASPENS, and others. Every one of them is a universal screener: group-administered where possible, benchmark-referenced, designed to sort a whole grade into risk bands three times a year.
That is a school's instrument, not a practitioner's. It answers "which of these ninety children needs a closer look" and it was never built to answer "what is going on with this one." The access problem is the same one the progress monitoring tools article documented in detail: most of these are district-licensed platforms, and the licence names schools as the permitted user.
Two things about that chart are worth knowing before you read a bubble on it.
First, the ratings come from what vendors chose to send in. NCII's Technical Review Committees rate each standard as convincing, partially convincing or unconvincing evidence, or as data unavailable, working from material "submitted by developers," per the tools chart user guide. A tool that is absent from the chart has not been judged and found wanting. It has not been judged.
Second, the chart says explicitly that it is not a recommendation: "The TRCs rate submitted tools against technical standards but do not compare them to other tools on the chart. The presence of a tool on a chart does not constitute endorsement and should not be viewed as a recommendation from either the TRCs or the NCII." The review process and the chart's own landing page carry the same framing.
Now look at what Pearson itself puts in the screening column.
Pearson publishes a Dyslexia Index for both the KTEA-3 and the WIAT-4 — a purpose-built composite of two or three subtests, twenty minutes or less, with its own manual, its own record forms and published effect sizes of 1.76 to 2.11 separating dyslexia groups from matched controls. Its own flyer opens: "If dyslexia screening is on your to-do list."
There is no dyscalculia index. In Pearson's own 2025 dyscalculia technical report, by Kristina C. Breaux and Tina Eichstadt, the screening column of the toolkit table contains the KTEA-3 Brief Form and the WRAT5 — two general achievement measures. The assess column contains KeyMath-3 DA, the KTEA-3 Comprehensive Form, the WIAT-4 and PAL-II Math.
So on the reading side the answer to "what do I screen with" is a purpose-built twenty-minute composite with published effect sizes. On the math side, from the same publisher in the same year, the answer is a general achievement brief form and a four-subtest screener that contains one timed computation task.
The same publisher built a purpose-made screener for one construct and recommends a general achievement brief form for the other. That is not an oversight. It is what the evidence base supports.
What that same Pearson report says next, which is the part to quote
The 2025 report is unusually candid for a publisher document, and two passages are worth carrying into a parent conversation verbatim.
On the relationship between the word and the category: "To qualify students for services in educational settings, the terms dyscalculia and SLD-math are typically used synonymously," followed immediately by the instruction that "practitioners are strongly encouraged to define how terms are being used when communicating with families, educators, and other professionals."
And then, in the same answer: "A diagnosis of dyscalculia or SLD-math is not sufficient to guide intervention. To develop an effective treatment plan, it is necessary to understand the individual's specific learning profile and the factors contributing to the impairments, regardless of classification terminology."
A test publisher telling you that the diagnosis its instruments contribute to will not tell you what to teach is the most useful sentence on this whole shelf.
The report also answers the question families ask most often. Is math anxiety or poor math performance the same thing as dyscalculia? "No, not all individuals with math difficulties or math anxiety have dyscalculia. Math difficulties manifest in a variety of ways and for different reasons."
What the federal practice guides say, and what they never say
Two What Works Clearinghouse practice guides govern this territory, and the difference between them is the story.
The 2009 guide, Assisting Students Struggling with Mathematics: Response to Intervention (RtI) for Elementary and Middle Schools — full text here — opens with Recommendation 1: "Screen all students to identify those at risk for potential mathematics difficulties and provide interventions to students identified as at risk."
The panel rated the level of evidence behind it Moderate. It also added a caution worth keeping: "We stress that no one screening measure is perfect and that schools need to monitor the progress of students who score slightly above or slightly below any screening cutoff score."
Its 2021 replacement, Assisting Students Struggling with Mathematics: Intervention in the Elementary Grades — full PDF, ERIC record — removed the subject.
In the panel's own words: "We narrowed the focus for this update for three reasons. First, MTSS is more widely used than when the original guide was published, and a broad overview to facilitate implementation of MTSS was no longer necessary. Therefore, topics such as progress monitoring, screening, and motivation, though important, are not included in this guide."
So the most recent federal guidance on mathematics screening dates from 2009, was rated Moderate, and was written about school-wide universal screening rather than individual identification.
The same 2021 panel also stated the asymmetry directly: "Historically, mathematics intervention research has received far less energy and funding than reading intervention research." That sentence, from a federal expert panel rather than an advocacy organisation, is the cleanest available answer to a family asking why so much less exists on this side.
And a full-text search of both guides returns the same result. The word "dyscalculia" does not appear in the body of either one. Its only occurrence anywhere in the 2009 guide is a chapter title in the reference list; in the 2021 guide it does not appear at all. This site's math intervention for dyscalculia article found the same absence from the intervention side, and the assessment side confirms it.
Administering is not interpreting
Everything above is a purchasing map. The ethics are a separate surface and they are narrower.
The AET Code of Ethics, Revision #9 carries three clauses that bear directly on a math battery in an educational therapist's hands.
II.I.C — educational therapists "use only those assessment instruments for which they have been adequately trained."
II.I.D — they "seek interpretation of assessment data from professionals in related fields (e.g. medical, psychological, speech/language, neuropsychological)."
I.I.B — they "provide professional services only within the boundaries of their competence based on their education, training, supervised and professional experience."
Read alongside the AET Fact Sheet, which says educational therapists provide intervention for "dyscalculia and/or other math-related difficulties," there is the same two-document tension this site has now documented four separate times — in the scope map, in the psychoeducational evaluation article, in the executive function scope of practice piece, and again here. The fact sheet claims the territory. The Code sends the interpretation elsewhere.
Can you use any of these instruments to identify a disability? No. Not the FAM, not the KeyMath-3, not the TOMA-3 you could buy this afternoon without showing anyone a credential. A purchasing permission is not a scope of practice, and II.I.D is explicit about where interpretation goes.
What you can do is describe performance, and that is not a small thing. Which operations are intact. Where fluency and accuracy diverge. Whether word problems fail on the arithmetic or on the reading. That is instructional information, it is yours to gather, and it is what a treatment plan is actually built from — the mechanics of which are in the treatment plan article.
A sequence, not a decision
Four questions, in order, and none of them is "which test."
What decision is this assessment for? Eligibility belongs to a school team or a licensed evaluator. Instructional planning belongs to you. The instrument follows from the decision, not the other way round.
What has already been administered? A recent comprehensive battery almost certainly contains math subtests already. Re-administering inside a practice effect window buys you worse data than reading the report you were handed.
Are you qualified for it, separately from whether you can buy it? The letter is the storefront's rule. II.I.C is yours, and the two are not the same test.
Who interprets it? If the answer involves a diagnosis, the answer is not you. Route it, under II.I.D, and say so plainly at intake — a conversation the first session article sets up.
When several organisations share a vocabulary
The word "assessment" is doing at least five different jobs across the documents cited above, and this site has now traced each one.
Publishers mean a norm-referenced instrument with a qualification letter attached — the finding of the achievement test qualification levels article, where WPS runs A, B, C and N with C in the middle and Riverside uses no letters at all.
Schools mean three-times-a-year universal screening against a benchmark, which is not the same thing as a percentile.
The clearinghouses mean a rated intervention or screening tool, and their ratings attach to outcomes rather than to labels — the point the evidence-based article makes at length.
Programme vendors mean whatever placement task ships in the box, which is the reading intervention finding transposed into math.
Cognitive training vendors mean a trained-task score, and the gap between that and anything else is the whole subject of the working memory training article published on this site yesterday.
If you hold more than one credential, the vocabulary problem compounds with the continuing education problem and the supervision problem, both of which this site has mapped. And if you are still working out where this sits relative to adjacent roles, the educational therapist, learning specialist and executive function coach comparison is the place to start.
Where the published record runs out
Six gaps, named rather than papered over.
WPS's qualification level definitions could not be read. Its product pages state the letter for each instrument — Level C for the FAM, Level B for the TEMA-3, both verified directly — but the Qualification Guidelines document that defines what those letters require is not retrievable by automated fetch from here, and its FAQ page exposes only the definition of Level N.
So this article states which letter WPS assigns and declines to state what WPS means by it. If that distinction is load-bearing for a purchase you are about to make, call WPS.
Remote administration terms are not covered here at all. Every instrument above has its own telepractice position and they differ sharply by publisher; that is the subject of a separate article.
No prices appear in this article, by policy. Kit prices move, and a figure published today is a correction waiting to happen.
The Number Sense Screener's current publication status was not independently confirmed. The technical report is public and was read directly for the reliability figures quoted above; the Brookes product page returned an access error and the "Research Edition" designation is taken from the title page of the user's guide itself.
Test-retest and predictive validity figures are not compared across these instruments. Doing that properly requires the technical manuals, most of which ship with the kit and are not public. Anyone telling you one of these instruments is more accurate than another without citing those manuals is guessing.
No state's eligibility criteria are described anywhere above. What counts as a specific learning disability in mathematics for special education purposes is set by state rule, it varies, and it is a question for the school team and, where it matters, for an attorney.
Frequently asked questions
Is there a test that diagnoses dyscalculia?
No. Dyscalculia is not a diagnosis in DSM-5; the diagnosis is specific learning disorder with impairment in mathematics, and the APA's own summary says the work group rejected "dyscalculia" as a disorder name because its definitions vary too widely. The Feifer Assessment of Mathematics is the only widely distributed instrument organised around dyscalculia subtypes, and it produces three index scores rather than a diagnosis.
Can an educational therapist administer the KeyMath-3 or the TOMA-3?
Purchasing and competence are two different questions. Pearson lists the KeyMath-3 DA at level B, which requires a master's degree in a relevant field or equivalent professional standing; PRO-ED lists the TOMA-3 at Test Level A, which its policy says carries no special purchasing qualifications at all.
AET's Code of Ethics II.I.C separately requires that educational therapists use only instruments for which they have been adequately trained, and that requirement does not move when a storefront's does.
Why does the same test have different qualification levels at different publishers?
Because the letters are each publisher's own purchasing policy rather than an industry standard. The Feifer Assessment of Mathematics is level B at PAR and Level C at WPS. Even where the letter matches, the definition may not: PAR's Level B requires a bachelor's degree plus measurement coursework, while Pearson's Level B requires a master's degree.
What is the difference between a math screener and a math assessment?
A screener sorts a population into risk bands quickly and is usually group-administered against a benchmark; the mathematics tools on NCII's academic screening chart are all of this kind and most are licensed to districts. A diagnostic assessment is individually administered, norm-referenced, and yields subtest-level performance. The FAM Screening Form sits between them: PAR describes it as identifying children at risk within an RTI model and deciding whether further assessment is warranted.
Does the federal government recommend screening for math difficulties?
The most recent federal guidance is the 2009 What Works Clearinghouse practice guide, whose first recommendation is to screen all students, rated Moderate for level of evidence. Its 2021 replacement explicitly removed screening from its scope, stating that topics such as progress monitoring, screening and motivation "though important, are not included in this guide."
Why is there so much more available for dyslexia than for dyscalculia?
The 2021 What Works Clearinghouse panel answered this in its own introduction: "Historically, mathematics intervention research has received far less energy and funding than reading intervention research." The panel adds that the discrepancy has begun to change over the past ten to fifteen years, but the assessment shelf still reflects the older pattern — Pearson publishes a purpose-built Dyslexia Index for two achievement batteries and no equivalent for mathematics.
Where this leaves your caseload
There is no instrument that will hand you the word the family arrived with. What the shelf will give you is a description of what a student can and cannot do with number, at a level of detail good enough to plan from, provided you are honest about which questions belong to you and which get routed.
That is a smaller claim than the marketing around any of these products, and it is the one that survives contact with the sources.
If you work with students on the mathematics side and want the next of these maps when it publishes, join the Illuminate educator list. It is where new practitioner resources go first.
Comments