Outdated Test Norms: What Actually Makes a Score Too Old
Updated: 4 days ago
Somebody in a meeting says the testing is too old to use. Everybody nods. Then somebody asks what the cutoff is, and the room produces a number nobody can source.
Outdated test norms are a real problem with a real literature behind them, and the rule most practitioners carry in their heads is not in any of the documents they would cite for it. There is no ten-year rule.
What exists instead is a set of clauses that are vaguer in one direction and considerably more demanding in another, plus a quiet publishing practice that makes the question you are probably asking the wrong one.
This is the shape of it. The ethics code that governs the psychologists who write the reports you read names no interval at all. The measurement standards everybody defers to tie revision to research and to changes in the construct, not to the calendar.
One state licensing board has published an actual number, and it is one or two years, not ten. And the test publishers themselves say, in print, that time alone does not make a test obsolete, while simultaneously selling you new norms for tests whose items have not changed since 2004.
So the question is not how old the test is. It is which number you are reading, and whether the number you need is published at all.
The rule you have heard, and the documents that do not contain it
The ten-year rule circulates as though it were a citation. It is usually attributed to one of three places: the American Psychological Association's ethics code, the joint testing standards published by AERA, APA and NCME, or the publisher's own manual.
It is in none of them.
That is worth sitting with, because the rule is doing real work in real meetings. It gets evaluations dismissed, it gets re-testing ordered, and it gets practitioners talked out of instruments that are perfectly defensible. A rule that strong deserves a source, and this one does not have one.
What follows is what the documents actually say, in each case quoted from the document rather than from the folklore about it.
What the APA Ethics Code actually says
The relevant provision is Standard 9.08 of the Ethical Principles of Psychologists and Code of Conduct, and its heading is the first useful thing about it.
The heading splits the question in two
The heading is "Obsolete Tests and Outdated Test Results." Two different problems, named separately, in one standard.
(a) Psychologists do not base their assessment or intervention decisions or recommendations on data or test results that are outdated for the current purpose. (b) Psychologists do not base such decisions or recommendations on tests and measures that are obsolete and not useful for the current purpose.
Subsection (a) is about a result. A child was tested three years ago and the scores are stale because the child has changed, or because the question has changed. Subsection (b) is about an instrument. The test itself is no longer fit for the job.
Those are genuinely different failures and they have different remedies. A stale result is fixed by testing again. An obsolete instrument is fixed by choosing a different instrument, which may mean re-testing with something else entirely.
What it does not say
Read both subsections again and notice what is missing. No number of years. No interval. No reference to the publication date of the test or to the age of its normative sample.
The operative phrase in both is "for the current purpose." The standard makes the judgement situational on purpose, and it puts the burden on the person making the decision rather than on a calendar.
What the measurement standards say, and where to read them free
The Standards for Educational and Psychological Testing is the document practitioners mean when they say "the standards." It is jointly published by the American Educational Research Association, the American Psychological Association and the National Council on Measurement in Education, and the current edition is 2014.
The revision clause ties revision to knowledge, not to time
The relevant passage, quoted by PAR from page 83, reads:
revisions or amendments are necessary when new research data, significant changes in the domain, or new conditions of test use and interpretation suggest that the test is no longer optimal
Three triggers, and not one of them is elapsed time. New research data. A changed domain. New conditions of use. A test whose construct has been stable and whose research base has not moved is not, by this clause, due for revision merely because a decade went past.
It is free, and most practitioners do not know that
NCME notes that in March of 2021 it, "along with its coauthoring organizations AERA and APA, announced the open access release of the 2014 Standards for Educational and Psychological Testing."
If you have ever argued about what the standards require without having read them, that is now a solvable problem. NCME also records that the edition before this one was published in 1999, which puts fifteen years between the last two editions.
A new edition is in progress
A survey document from NCME's Standards and Test Use Committee records that in autumn 2023 the sponsoring organisations nominated nearly two hundred people as potential Joint Committee co-chairs and members for a revision of the 2014 edition. The document does not give a completion date.
One respondent to that survey made a point worth repeating here, because it is the same complaint practitioners make about test manuals: there ought to be incremental updates, because waiting ten to fifteen years for a total overhaul is too long.
So check the edition year before you quote the standards at anybody. The document is being revised and the paragraph numbers may not survive.
One state board has published a number, and it is not ten years
The North Carolina Psychology Board devoted part of a newsletter to the use of tests and norms, and it is the only place in this research where a regulator put a figure in print.
The figure is not ten years.
Revised tests or new norms are considered current if they are within one or two years of the release of the revised test.
Read plainly, that is a far shorter window than the folk rule, and it is pegged to the release of a revision rather than to the age of the instrument. The board also states a rule about mixing editions that is absolute in a way the rest of this material is not:
You should never administer one edition of a test and then score and interpret the test using the norms of another edition.
And where a practitioner decides against the most recent version, the board requires the decision to be defensible: it "must be based on valid research data and sound clinical judgment" and must be clearly clinically supported in compliance with Standard 9.08.
This is one board, in one state, writing for licensed psychologists. It is not a national rule and it does not bind an educational therapist. It is quoted here because it is the clearest published statement of the principle available, and because the reports you read are frequently written by people this board regulates.
What the publishers say about their own tests
The publishers have an obvious commercial interest in telling you your test is out of date. What is interesting is that at least one of them says the opposite in writing.
PAR: time alone does not do it
PAR's own guidance on when a test should be updated or revised includes this sentence:
tests do not become obsolete simply because of the passage of time
PAR's reasoning is that different constructs drift at different rates. Cognitive measures need updating more often than personality measures, because measured intellectual performance moves over cohorts in a way that personality content does not. The obligation PAR describes is to monitor for change, not to replace on a schedule.
The International Test Commission's version
PAR also quotes the International Test Commission's guidelines, from page 9, on what drives a revision:
Test revisions may be driven by knowledge that the assessed behaviors are subject to substantial change over time, by significant demographic changes
Again: change in the behaviour, change in the population. Not the calendar.
Oral reading fluency is the clearest live case of this. The most recent compiled fluency norms were published in 2017 and their newest input is data from the 2013 to 2014 school year, which we work through in oral reading fluency norms: what the table does not say.
The number on the manual is not the number that matters
Here is the part that changes what you do on Monday.
Practitioners judge currency by the year on the manual, because it is the only year they are given. That year is the copyright or publication year. It is not the year the normative data were collected, and on the instruments an educational therapist actually encounters, the two numbers are four to five years apart.
CTOPP-2: copyright 2013, data collected 2008 and 2009
PRO-ED's own product page for the Comprehensive Test of Phonological Processing, Second Edition, carries a copyright of 2013 and says this about the norms:
All new normative data were collected in 2008 and 2009.
The test was normed on 1,900 individuals aged 6 through 24, it covers ages 4-0 through 24-11, and it sells at Test Level B. The copyright year is 2013. The data are from 2008 and 2009.
In 2026, that is a manual stamped thirteen years ago sitting on top of a sample gathered seventeen and eighteen years ago.
TOWRE-2: the identical four-year gap
The Test of Word Reading Efficiency, Second Edition shows the same pattern. Copyright 2012. Over 1,700 individuals. Ages 6-0 through 24-11. Test Level B. And:
All new normative data were collected in 2008 and 2009.
Same collection years as the CTOPP-2, a year earlier on the copyright. If you have been treating the TOWRE-2 as newer than the CTOPP-2 because of the edition dates, it is not. They were normed in the same two calendar years.
TOMA-3: a population year rather than a collection year
The Test of Mathematical Abilities, Third Edition does something slightly different and worth noticing. Copyright 2013. Normed on 1,456 students aged 8-0 through 18-11 living in 21 states. Test Level A, which PRO-ED's own policy defines as requiring no special qualifications to purchase. And the norms are described as follows: "All new normative data were collected from a demographically representative sample of the 2011 U.S. school-age population."
That tells you the reference population year, 2011, but not the years of collection. It is a different claim from the CTOPP-2 and TOWRE-2 wording, and it is less specific in exactly the dimension you are trying to evaluate.
Two Pearson pages that do not publish the number at all
Pearson's product pages are organised differently. The WIAT-4 page gives a publication year of 2020, an age range of 4:0 through 50:11 and qualification level B. The TOWL-4 page gives 2009, ages 9:0 through 17:11 and level B.
Neither page publishes the normative sample size or the years the data were collected.
That is not a criticism of the instruments. It is a statement about what a practitioner can establish from the public record without buying the technical manual. For the PRO-ED instruments above, the collection years are on a free product page. For these two, they are not.
If the age of the norms is load-bearing for a decision you are making, you need the technical manual, and on these two instruments there is no free substitute for it.
The normative update: the same test with different norms
Now the practice that makes the edition question unanswerable.
Publishers re-norm tests without revising them. The items stay the same, the structure stays the same, the edition number stays the same, and the norms are replaced. The product is sold separately, usually as a manual supplement and new record forms, and it is labelled NU.
Three examples, all current, across two publishers.
DAS-II NU: same items, 2007 norms replaced with 2022
Pearson's page for the Differential Ability Scales, Second Edition, Normative Update, School-Age, states the position directly:
The existing items, subtests, and test structure did not change so customers familiar with the existing DAS-II product line will be comfortable administering the normative update without additional training.
The new normative data were gathered in 2022 from children aged 7:0 through 17:11, replacing norms from 2007. Pearson's own comparison document lists administration time, age range and subtest count as unchanged, and marks the school-age norms and the scoring platform as the differences.
KABC-II NU: a 2004 test with 2018 norms
The Kaufman Assessment Battery for Children, Second Edition, Normative Update carries normative data from March 2018 on a test originally published in 2004. The normative sample is 700 children aged 3 through 18 years 11 months, stratified by gender, race and ethnicity, parent education and region. Existing customers are told they do not need to buy a new kit. The same battery is also listed by WPS.
Fourteen years between the test and its current norms, with the items untouched.
GORT-5 NU: PRO-ED's 2026 entry
The Gray Oral Reading Tests, Fifth Edition, Normative Update carries a 2026 copyright and describes what it did: "Additional data were collected to update the original normative data from 2008–2010 to align with those of the U.S. population."
It also reinstated separate basal and ceiling rules for Fluency and Comprehension and reports extensive studies of floors, ceilings and item gradients. It covers ages 6-0 through 23-11 at Test Level B.
That is a working note for anyone holding a GORT-5. The original norms are from 2008 to 2010, which the fifth edition's own copyright year does not tell you, and there is now a normative update on the market.
Why this breaks the question "which edition do you have?"
Put the three together and the problem is plain. Two practitioners can both hold the DAS-II. Both are using the same items in the same order with the same structure. One is scoring against 2007 norms and the other against 2022 norms, and the edition number is identical on both kits.
So "which edition?" does not identify the norms. The question that identifies the norms is which norms table you scored from, and that question is almost never asked.
It also puts a wrinkle in the North Carolina board's otherwise clean prohibition on administering one edition and scoring with another edition's norms. A normative update is not another edition. It is the same edition with a different norms table, and the board's sentence does not reach it.
What re-norming actually does to a score
Re-norming matters because the same raw performance can earn a different standard score against a different sample. The direction and size of that shift is an empirical question, and the usual shorthand for it is only half right.
The Flynn effect, and why it gets cited
The standard argument for re-norming cognitive measures is that measured performance rose across cohorts through the twentieth century, so an old normative sample flatters a current test-taker: scored against a sample that performed less well, a child looks better than they would against their own cohort. PAR's guidance leans on exactly this in explaining why cognitive instruments need updating more often than personality ones.
The reversal, and why it matters here
The drift is not guaranteed to run one way. A systematic literature review by Dutton, van der Linden and Lynn, published in the journal Intelligence in 2016, found nine studies reporting negative Flynn effects in seven countries — Norway, Denmark, Britain, the Netherlands, Finland, France and Estonia. In the majority of those studies the decline ranged between 0.38 and 4.30 IQ points per decade.
Two cautions before anybody does arithmetic with that. These are national cohort studies in European countries, largely on conscript and school data, and they are not a statement about the normative sample of any instrument you own. And a reported decline in one country over one period does not license a correction to an individual score.
What the reversal does establish is narrower and still useful: you cannot assume you know which way an old norms table is biased. The safe formulation is that an old sample introduces an unknown, not that it inflates by a predictable amount.
The honest position is that an old normative sample makes a score less interpretable. It does not make it wrong in a known direction by a known quantity, and an evaluator who tells you it does is going further than the record allows.
What changed in 2025 and 2026
Two things moved recently enough that they may not have reached everybody.
WJ V, and the first post-pandemic norms in the series
Riverside Insights released the Woodcock-Johnson V on 1 February 2025. The release describes it as the first in the series to include norms developed after the COVID-19 pandemic.
Riverside's product page states that the WJ V was normed from 2022 to 2024, aligned to 2020 United States Census demographics, on a sample of more than 5,000. It describes a co-normed system of 26 achievement tests across 22 clusters and 20 cognitive tests across 12 clusters.
Whatever you make of the marketing, the norm years are published and specific, which is more than several of the instruments above manage.
GORT-5 NU
Covered above, and it belongs on this list because a 2026 copyright is as current as this market gets. If your referral sources are still reporting GORT-5 scores, the norms behind those scores may now be two tables old.
Where AET's Code of Ethics lands on this
An educational therapist is not bound by the APA ethics code or by a state psychology board. The binding document is AET's Code of Ethics, Revision #9, 2025, and the clause everybody would reach for is Section Two, I.B.
The clause names three limitations, and norms are not one of them
Section Two, I.B requires that educational therapists:
select and use appropriate assessment instruments, recognizing their limitations with respect to reliability, validity, and bias
Reliability. Validity. Bias. Three named limitations, and the currency of the normative sample is not among them.
That is not a loophole and it is not an accusation. Norm currency arguably lives inside validity, and a reasonable reader would get there. But the clause does not say so, and a practitioner looking for a rule about old norms in the document that governs them will not find one written down.
The same section continues with I.C, which requires that educational therapists "use only those assessment instruments for which they have been adequately trained," and I.D, which directs them to "seek interpretation of assessment data from professionals in related fields (e.g. medical, psychological, speech/language, neuropsychological)." Section One, III.B obliges them to "systematically advance their knowledge and skills" through continuing education.
Taken together, those clauses put the norm-currency question where this post has been pointing all along: it is part of reading a report competently and knowing when to ask the person who administered it.
Added 9 October 2026. The age of the norms is one question about a score; its precision is another, and the second is printed right beside it on almost every report. The confidence band and the standard error behind it covers what that band is, and why the same score can carry a band of two different widths.
Reading a report that used old norms
Most educational therapists meet this question as readers rather than as administrators. Here is the sequence, and it is a sequence of questions to the report rather than a verdict on it.
Four questions, in order
Which instrument and which edition. The name of the test and its edition, as printed in the report, not as remembered.
Which norms table. The one question nobody asks. If the instrument has a normative update, the report may or may not have used it, and the edition name will not tell you. This is a question for the evaluator.
When were those norms collected. Not the copyright year. For the PRO-ED instruments above the answer is on a free web page. For several Pearson instruments it is in the technical manual only.
What is the score being used for. Standard 9.08's phrase is "for the current purpose," and purpose is doing real work. A score used to pick a starting point for instruction and a score used to establish a discrepancy are not being asked to carry the same weight.
What you may not conclude
You may not conclude from the age of the norms that a reported score is wrong, that it is inflated, or that it should be adjusted. You may not re-score a protocol against a different norms table than the one the evaluator used. And you cannot determine from a date whether a student has a disability, which is not a question an educational therapist answers in any case.
What you may do is note the age of the norms in your own documentation, state which number you are relying on, and route the interpretive question to the professional who holds it. That is AET Section Two, I.D working as designed.
If you administer tests yourself
Two practical rules fall out of the material above, and both are cheap.
First, record the norm collection years alongside the instrument name in your own files, not the copyright year. It takes one extra line and it is the number you will want in two years when somebody challenges the score.
Second, check whether the instrument has a normative update before every intake cycle, not when you remember. Three of the instruments in this post acquired one between 2018 and 2026, and nothing on a kit you already own will tell you.
The failure mode is not using an old test. It is not knowing which norms you used, and discovering that two years later when somebody asks.
What is out of scope here
This post is about what professional bodies and test publishers publish. It is not about what any law requires.
Re-evaluation intervals, eligibility timelines, and what a district must or must not accept are statutory and regulatory questions, and nothing above should be read as an answer to any of them.
If a dispute turns on whether an evaluation is too old to be accepted for a particular purpose, that is a question for the district and for an attorney who practises in that area, not for this post and not for a test manual.
Nothing here is legal advice.
Where the published record runs out
In the house tradition of saying what could not be established, here is what this research could not close.
The exact wording and paragraph numbering of the 2014 Standards on norm currency could not be verified against the document itself. The APA's own standards page and the open-access copy were not retrievable by automated fetch during this research.
The passage from page 83 quoted above is quoted as PAR quotes it, and PAR is a test publisher rather than a sponsoring organisation. If that sentence is load-bearing for a decision, read it in the open-access document rather than taking it from here or from PAR.
The normative sample sizes and collection years for the WIAT-4 and the TOWL-4 are not on their public product pages, and this research did not obtain the technical manuals. Those two rows are blank on purpose and should not be filled in from memory by anybody.
PRO-ED's site returned a maintenance page on some product paths during this research. The pages cited above resolved; others did not, so a check that comes back empty may be the site rather than the product.
And the age of the GORT-5 NU's own new normative data is not stated on its product page. The page says additional data were collected to update the 2008 to 2010 norms. It does not say when.
Frequently asked questions
Is there a ten-year rule for test norms?
No. Neither the APA Ethical Principles of Psychologists and Code of Conduct nor the 2014 Standards for Educational and Psychological Testing states an interval in years. Standard 9.08 turns on whether data or an instrument is outdated or obsolete "for the current purpose," and the Standards tie revision to new research data, changes in the domain and new conditions of use.
What is the shortest published interval anybody has put in print?
The North Carolina Psychology Board's winter 2022 newsletter states that revised tests or new norms "are considered current if they are within one or two years of the release of the revised test." That is one state board writing for licensed psychologists, not a national rule.
Does the copyright year on a test manual tell me when the norms were collected?
No, and the gap is routinely four years or more. The CTOPP-2 carries a 2013 copyright and PRO-ED states its normative data were collected in 2008 and 2009. The TOWRE-2 carries a 2012 copyright and the same collection years.
What is a normative update, and how is it different from a new edition?
A normative update replaces a test's norms while leaving its items and structure unchanged. Pearson states for the DAS-II NU that "the existing items, subtests, and test structure did not change." It is sold separately and the edition number does not change, so two practitioners holding the same edition may be scoring against different norms.
Can I tell from an old normative sample that a score is inflated?
No. The usual argument is that measured performance rose across cohorts, which would flatter a test-taker scored against an older sample. But Dutton, van der Linden and Lynn's 2016 review in Intelligence found nine studies reporting declines in seven countries, so the direction of drift is not something you can assume. An old sample makes a score less interpretable, not wrong by a known amount.
Does AET's Code of Ethics require me to use current norms?
Not in those words. Section Two, I.B of Revision #9 requires selecting and using appropriate instruments "recognizing their limitations with respect to reliability, validity, and bias." Norm currency is not named. Section Two, I.D directs educational therapists to seek interpretation of assessment data from professionals in related fields, which is where the question usually belongs.
Related reading
Several of the posts on this site share a vocabulary with this one, and each answers a question this post deliberately leaves open.
On which storefront will sell you which instrument and at what level, see achievement test qualification levels, which found that the letter is a purchasing rule for one publisher rather than an industry scale.
On the score type that cannot survive subtraction, see grade equivalent scores, where the publishers printing them disagree about what they are for. On what you may do with a report somebody else wrote, see psychoeducational evaluation for educational therapists, which found AET's two governing documents disagreeing about interpretation.
On three batteries publishing a score called Written Expression and no two building it the same way, see written expression assessment tools. On the absence of a dyscalculia analogue to the reading screening ecosystem, see math assessment tools for dyscalculia. On three different tasks wearing one label, see reading comprehension assessment tools. On the instruments in the executive function space, see executive function assessment tools, compared.
On testing through a screen and what each publisher permits, see remote assessment administration. On the systems that rate programs rather than people, see reading program evidence ratings and progress monitoring tools compared, where a 50th percentile in one system is not a 50th percentile in another. On what the evidence base for the service model itself does and does not contain, see is educational therapy evidence-based.
On the standards document that four different things get called, see the IDA knowledge and practice standards. On what the training routes actually cost in hours, see structured literacy training programs compared. On the hours that count toward a credential, see educational therapy supervision requirements, and on the units those hours are measured in, see educational therapy continuing education.
And for the scope questions underneath all of this, see the role map at educational therapist vs. learning specialist vs. executive function coach and the practitioner scope map at what is educational therapy.
Get the practitioner list
We publish one of these a day for educational therapists, learning specialists and executive function coaches, sourced to the issuing bodies and the publishers rather than to other people's blogs. If that is useful to your practice, join the educator list and the next one will reach you.
Comments