What the 2025 EFPA Test Review Model Changes

What does the model evaluate, and what does its name establish?

The 2025 EFPA Test Review Model changes how experts organize an evaluation; its name does not certify an intelligence score. EFPA describes Version 5.0 as a structure for rigorous descriptions and evaluations of psychological and educational assessments, intended for expert reviewers and adaptable to national arrangements. The model is not itself an IQ test, a score conversion, a validation sample, or a clinical assessment. As “EFPA Test Review Model – Version 2025” explains, this structure supports test choice through expert review. A framework reference does not show that this particular test was reviewed or which interpretations its evidence supports. The model sets out areas for description and evaluation, including test materials, norms, reliability or precision, validity, support for fair use, and digitally generated reports. They do not create missing comparison data or make a score precise simply by appearing in a framework. Here, validity means evidence and theory supporting a particular interpretation of scores for a proposed use; it is not a permanent label attached to a test name. EFPA’s Version 2025 model also states directly that ratings made under it do not imply EFPA endorsement, approval, or recommendation of a test. So “reviewed under the EFPA model” and “EFPA-approved” are not interchangeable claims. To interpret an IQ result responsibly, look for an actual test-specific review and identify the exact instrument and version, intended population, score interpretation, and use it addresses. The model’s value is the questions it directs reviewers to ask; the citation alone establishes no norm, precision, validity, or fitness for a consequential decision.

Sources: EFPA Test Review Model – Version 2025 – EFPA; EFPA Test Review Model Version 2025, Version 5.0

What is genuinely different from the 2013 model?

Version 5.0 is a major revision, but it did not introduce the basic technical questions used to evaluate an intelligence test. The 2013 model already covered norms, reliability and validity, computer-based and web-based materials, and computer-generated reports. The clearest change is emphasis and organization: the 2025 model has a distinct section on supporting fair use, addresses digital and print testing more deliberately, and describes generated reports separately from its evaluation of their quality. This expanded framework is not evidence that a particular test became fairer, more accurate or suitable for a decision.

The 2013 article “Assessing the Quality of Tests: Revision of the EFPA Review Model” makes the continuity clear. Its account of Version 4.2.6 says practical criteria focused mainly on paper-and-pencil tests, although computer-generated reports already had their own evaluation criterion. It also describes provisions for computer-based and web-based tests. Digital assessment review, therefore, was not invented in 2025. The earlier revision also identified gaps: more systematic attention to methods beyond classical test theory, cross-cultural adaptation, and a scope broader than workplace testing.

The 2025 EFPA Test Review Model describes later pressures. Its “Note on this revision” says testing practice had diversified across Europe with technological and psychological developments and wider professional use in some countries. It points to online delivery in some assessment areas, digitized interpretation and remote feedback. It also notes varying professional requirements, inclusive language, and relevant data and digital regulations. EFPA says the revision aims to rebalance attention between digital and print tests and give reviewers more places to reflect these developments. These are the model’s stated reasons, not evidence that every country or test changed alike.

The structural signal is clear in the contents. Version 5.0 includes “Supporting Fair Use of Tests” as its own evaluation section. It asks reviewers to describe digitally generated reports and separately evaluate their quality. In 2013, reports also appeared in both the description and evaluation parts, so this is a refinement, not an entirely new subject. The clearer addition is fair use as an explicit review domain alongside norms, reliability or precision, validity and report quality. That makes intended use and context harder to overlook.

This matters when someone sees a test described as “EFPA reviewed.” The 2025 framework directs attention to conditions around use, but its title alone does not settle whether an intelligence score is appropriately interpreted for a person or purpose. A fair-use section also does not establish a test-specific fairness finding. Reviewers still need evidence relevant to the population, administration and score claim; a framework can organize those questions, not supply missing evidence.

The useful comparison is not “the old model ignored digital tests; the new model fixed them.” The 2013 framework already scrutinized core measurement properties and included digital materials and reports, while Version 5.0 makes current digital practice more central and gives fair-use review a clearer place. Compare the questions asked, evidence available and stated limits. A 2025 review is not automatically stronger because it is newer, and a 2013 review did not necessarily ignore digital testing. What matters is whether evidence and reasoning support the specific score interpretation and use now being claimed.

Sources: EFPA Test Review Model Version 2025, Version 5.0; Assessing the quality of tests: revision of the EFPA review model; Assessing the quality of tests: revision of the EFPA review model – PubMed; Consultation: Model for the Review, Description and Evaluation of Psychological and Educational Tests – EFPA

Why does the raw score need a reference population?

A raw score records performance on a particular test under particular conditions. It may be a count of correct answers or another total derived from responses. That is useful as a description of performance on that form, but it does not show how typical or unusual that performance is among adults, students, or another population. For that comparison, the test needs a scoring procedure and reference data that fit the interpretation.

A norm-referenced score interprets performance relative to a specified reference population. An IQ score is ordinarily presented this way: its meaning depends on comparison with an appropriate group, not just the number of items answered correctly. A percentile is a position in that comparison, not the percentage of questions answered correctly. A criterion-referenced interpretation instead compares performance with a stated standard, such as a proficiency threshold. “The GRoNC: Guidelines for Reporting on Norm-Referenced and Criterion-Referenced Scores” explains why these interpretations must be distinguished.

The GRoNC also separates three populations that can be blurred together: the target population, to whom a test is intended to be administered; the normative population, from which norm data are collected; and the reference population, against which a person's score is interpreted. They may overlap, but are not automatically identical. Its intelligence-test examples show how age boundaries can matter: normative data may cover a wider span than the intended population, while a person is compared with others of the same age. Confusing these populations can encourage use outside the test's intended range. These examples explain the terms; they do not establish norms for another test.

Reviewing a norm claim therefore requires more than asking whether the sample was large. A reviewer needs enough information to judge who was included, how the sample relates to the intended population, what testing conditions applied, and how responses became scores. The GRoNC asks whether the reference population fits the test's aim, sampling is suitable for estimating norms, the procedure is described, and score precision and interpretation are explained. Size alone cannot resolve a mismatch. The paper reports that the guidance was developed through literature review and two Delphi rounds with 11 theoretical experts and 14 test developers. That process supports development of the reporting guideline, not the validity of any individual test.

Mapping a browser quiz's correct-answer count onto a familiar IQ number cannot create an adult population comparison. A conversion rule can assign a label, but without appropriate norm evidence the label does not establish where performance stands among adults. A band produced without a validated adult reference population is not a normed adult IQ estimate or a confidence interval. The “Standards for Educational and Psychological Testing” likewise call for clear procedures and rationale for deriving standardized scores, so users can judge quality and precision. The standards organize evaluation; they do not certify a particular score or conversion.

Keep the inference proportional to the evidence. If a result gives only a raw total, it describes performance on that form under its scoring rules. If it supplies a norm-referenced score, ask which reference population it uses, how the conversion was derived, and whether the comparison fits the person and decision. For a criterion category, ask what standard defines it and what evidence supports its use. A bare number answers none of these questions. The method and intended interpretation must be clear before the score can support a comparison; it does not establish identity, worth, diagnosis, or fixed potential.

Sources: The GRoNC: Guidelines for Reporting on Norm-Referenced and Criterion-Referenced Scores; EFPA Test Review Model Version 2025, Version 5.0; Standards for Educational and Psychological Testing

What evidence should a review make visible about an IQ claim?

A useful review connects each score claim to evidence for the stated population, score meaning and proposed use. Norms, precision and validity answer different questions: who is being compared, how uncertain the score is, and whether the interpretation is supported for its purpose. Strong evidence on one point cannot fill a gap in another. The 2025 EFPA Test Review Model separates norms, reliability/precision, validity and fair-use support. Its form organizes evidence; it does not supply evidence a publisher has not provided. Validity means the degree to which evidence and theory support an interpretation of test scores for a proposed use. The Standards for Educational and Psychological Testing make the unit clear: professionals evaluate interpretations for specified uses, not a test in the abstract. EFPA’s 2025 model likewise says the evidence needed depends on the intended use and that broader intended applications require more supporting evidence. A score describing performance on particular reasoning tasks makes a narrower claim than one estimating general intelligence or guiding placement. Evidence for one interpretation cannot simply be carried over to another. Precision addresses a different question: how consistently a procedure measures and how much measurement uncertainty may surround a score. EFPA’s review prompts reviewers to examine the reported reliability or precision evidence, the samples and procedures behind it, and whether standard errors of measurement are supplied. Consistency alone cannot establish meaning or suitability for a decision; evidence supporting an interpretation does not show how precisely an individual score was measured. Norm evidence supplies the comparison needed for a norm-referenced interpretation. A raw score is the observed result on the administered items; by itself it says how many responses met the scoring rule, not how the performance compares with a defined population. A norm-referenced score interprets performance relative to a specified reference group. EFPA’s review therefore asks about the norm group, how it was formed and whether it fits the target population and interpretation. The review should identify the version, population and score scale covered. A raw total or familiar-looking number alone cannot support a broad “adult IQ” interpretation. Consider an illustration, not a study: someone completes a short browser puzzle set and answers several items correctly. The total describes performance on those items under those conditions. Interpreting it relative to adults requires relevant reference data and a defensible conversion; diagnosis or placement requires separate supporting evidence. A checklist cannot manufacture missing data. The Standards for Educational and Psychological Testing are a separate professional reference, not an EFPA approval process. They explain why evidence must match the intended use; EFPA structures reviewers’ inspection of documentation and technical evidence. Neither a standards citation nor “reviewed” is complete evidence on its own. Ask what evidence supports the exact interpretation, for which population, with what precision, and for what use. If those particulars are not available, keep the conclusion correspondingly narrow: a score may describe observed performance, while its broader meaning remains unsupported.

Sources: Standards for Educational and Psychological Testing; EFPA Test Review Model Version 2025, Version 5.0; Assessing the quality of tests: revision of the EFPA review model

What does an overall review rating summarize—and leave open?

An overall review rating is an expert judgment about evidence considered under a defined framework. It is not a universal pass mark, nor proof that every score from a test suits every person or decision. EFPA’s 2025 Test Review Model says reviewers should judge overall validity in light of the test’s purpose and the relative importance of different evidence sources; the rating “should not be regarded as an average or as the lowest common denominator.” The summary needs its reasoning and limits beside it.

The rating is the end of a chain of judgments, not a mathematical shortcut. Reviewers examine the test’s intended construct, administration and scoring, then consider documentation and technical evidence. EFPA separates areas such as norms, reliability or precision, and validity, and asks reviewers to form an overall judgment informed by them. Its guidance says the importance of validity evidence depends on the test’s purpose. Evidence relevant to one use may not answer another: describing practice performance raises a different question from informing an educational assessment. This identifies what to examine.

Averaging component ratings can mislead because it treats each number as equally important and interchangeable. The criteria are not points that compensate for one another. Strong documentation cannot supply absent evidence for an intended interpretation, and evidence about one population does not automatically establish a conclusion for another. A gap central to the proposed use may matter more than a lower rating on less relevant evidence. EFPA calls for expert judgment rather than a mechanical average. The overall rating should show how relevant evidence was weighed, not hide the reasoning behind a label.

A rating category alone does not show what documents reviewers examined, which population the evidence concerns, what interpretation they evaluated, or how much uncertainty remains. EFPA’s model provides space for reviewers’ validity comments and calls for validity evidence to be discussed. Those comments connect the summary to evidence and the proposed interpretation. A bare statement that a test received a favorable “overall” rating leaves its scope and grounds unclear.

A framework citation and an instrument-specific review are different things. EFPA says Version 5.0 guides expert reviewers while allowing local implementation and judgment. It explicitly states that ratings under the model do not imply EFPA endorsement, approval or recommendation of a test. A useful report should identify the test and version, evidence available, interpretation or use addressed, and limitations. Without those details, readers cannot tell whether it speaks to the score claim at issue.

EFPA recognizes that technically important supporting material may be commercially confidential or restricted by copyright, contracts or intellectual-property rights. The absence of every underlying document from a public account does not alone establish that reviewers lacked evidence. Still, the report should say what was evaluated and what conclusion follows from the information considered. Readers can ask whether a test-specific review is available, which version and use it covers, and what limitations reviewers identified.

The 2013 paper “Assessing the quality of tests: revision of the EFPA review model” describes recommended procedures for that earlier framework. It is not proof that every current national program uses the same procedure. Ask what the particular report says about its reviewers, materials and scope rather than assuming uniformity from the model’s name.

Read an overall rating as a synthesis for a stated purpose, then follow its explanation to the evidence and limits. It can help assess whether a test’s technical support addresses a particular use. It cannot automatically be carried to another population, interpretation or consequential decision. If a report gives only a rating, ask for its reasoning and scope before treating it as an answer about an intelligence result.

Sources: EFPA Test Review Model Version 2025, Version 5.0; Assessing the quality of tests: revision of the EFPA review model

What does digital delivery change in the evidence a reviewer should inspect?

Digital delivery changes the conditions under which performance is produced and results are interpreted. It does not make an intelligence score more or less trustworthy. A reviewer should match evidence to the version and mode used: instructions, response method, device, timing, supervision, and adaptive routing. These are not automatic defects; the question is whether documentation and evidence address them. The 2025 EFPA Test Review Model makes this scrutiny easier to locate: it asks reviewers to describe digitally generated reports and separately evaluate their quality, within a broader review of materials, norms, precision, validity, and fair use. This is a sharper and more integrated treatment, not the first recognition of computer-based assessment. The 2013 paper “Assessing the Quality of Tests: Revision of the EFPA Review Model” already listed computer-generated reports in both the instrument description and evaluation, and said that computer- and web-based materials were subject to review. Version 5.0 expands the frame around delivery: its revision note points to online administration, digitized interpretation, and feedback at a distance as developments the model should address. Its report section also asks about the breadth and specificity of a report, whether its detail is justified by the instrument’s scores, and whether distinct score bands fit the scale’s measurement error. This added structure focuses attention on the path from test to interpretation; it does not establish that digital and paper forms are interchangeable. That distinction is consistent with the separate “Standards for Educational and Psychological Testing,” which advise users to seek a sound rationale and, when possible, empirical evidence before changing format, administration mode, instructions, or language. The Standards note that some minor changes may reasonably be expected to have little effect, while other changes can affect score meaning; where a modification becomes widespread, validity evidence should be gathered and norms may need to reflect the changed conditions. This does not mean screens necessarily alter performance. Reviewers should check that the reviewed setup matches the one used, and that evidence supports carrying the interpretation across any difference. Online delivery does not disqualify a test, but similar-looking content does not prove equivalence. The report itself needs the same discipline. In Version 5.0, reviewers assess each report under review, including its scope, reliability, validity, inclusivity, and recommendations. The model asks whether the report’s specificity matches the detail available from the instrument and whether the number of score bands can be justified in light of measurement error. For a personalized dashboard, ask whether each statement follows from the scores and evidence and whether its detail is warranted. Tailored wording can help within those limits, but a polished layout or individualized text cannot supply missing evidence. Since the 2013 model already reviewed generated reports, the change is a more explicit set of prompts for examining the route from score to language. For a reader, the request is specific: ask whether an actual review examined the same digital form and report pathway, and read what it says about the relevant version, population, score precision, and intended interpretation. The Standards also caution against relying solely on computer-generated interpretations. Automated feedback is not inherently unusable, but supporting evidence and appropriate judgment matter, especially for consequential decisions. Fine-grained claims beyond what the scale supports remain overreach, whatever the interface looks like. The 2025 model gives reviewers a clearer structure for asking these questions. It does not certify a particular platform, establish mode equivalence, or turn report detail into accuracy.

Sources: EFPA Test Review Model Version 2025, Version 5.0; Assessing the quality of tests: revision of the EFPA review model; Standards for Educational and Psychological Testing

How does explicit fair-use review change the reader's question?

Explicit fair-use review changes the question from whether a test produces identical outcomes for every group to whether, in its intended setting, scores have adequately supported meaning and accuracy for the people taking it. Reviewers also consider whether users have guidance about limits, adaptations, and appropriate use. The EFPA Test Review Model, Version 5.0, treats fairness as contextual: relevant groups depend on the population and purpose. Equal group averages are not its definition of fairness. For an intelligence-test reader, that means a group average alone cannot establish whether an individual score is accurate, comparable, or suitable for a decision.

Version 5.0 asks reviewers to consider whether people at the same level of the measured construct would receive scores with comparable accuracy and meaning, and whether technical information supports fair use in context. This connects the construct a test claims to measure with its administration, population, and intended use. A mean describes an average outcome; it does not by itself show whether score meaning or accuracy differs among people with comparable measured standing. Conversely, a difference in group averages alone does not establish that a test is biased. The interpretation and use require relevant evidence.

For an intelligence test, relevant questions might include whether language demands belong to the ability being measured or instead obstruct access; whether a person with a disability can demonstrate the intended ability; and whether the reference population fits the interpretation. These are questions for evidence, not assumptions about a group. The Standards for Educational and Psychological Testing describes accessibility as an unobstructed opportunity to demonstrate standing on the measured construct. Removing an irrelevant barrier may make performance more informative about that construct. But accessibility alone does not establish that a score supports a broader IQ interpretation, and an adapted format is not automatically interchangeable with another format. Such conclusions require evidence for the test and use.

Fair-use review also asks whether the information given to practitioners helps them choose responsibly. Technical documentation may exist while evidence remains inadequate for a particular population, setting, or decision. EFPA’s model gives reviewers a place to consider those conditions and the guidance available to users. The Standards likewise treat fairness and accessibility as relevant to valid interpretation for individuals and subgroups in the intended population. The review can make visible which populations and possible sources of error were considered, and what cautions or adaptations users should understand. They do not replace professional judgment or mechanically settle every case.

The practical question is specific: for this person, administration, and proposed use, what evidence supports the score’s meaning, and what limits should shape the decision? Language, access needs, cultural or educational context, and norm-group fit matter when evidence shows they bear on the interpretation. A generic “fair” label cannot answer this question, and group averages cannot settle it without information about construct, measurement, population, and purpose. A separate fairness section cannot guarantee fair outcomes; it organizes scrutiny but cannot supply missing evidence or erase contextual differences. Treat a fairness label as a reason to inspect the actual review and rationale, not as a finding about every intelligence test. Look for reviewer comments relevant to the people, conditions, and decision at hand, and treat unresolved gaps as limits on what the score can responsibly mean.

Sources: EFPA Test Review Model Version 2025, Version 5.0; Standards for Educational and Psychological Testing

What can you responsibly conclude and do next?

A citation to the 2025 EFPA Test Review Model tells you that a reviewer framework is being referenced; by itself, it does not show that a particular intelligence test was reviewed or that a score supports a particular conclusion. EFPA says the model is designed for expert reviewers and can be adapted to national arrangements. It also states that ratings under the model do not imply EFPA endorsement, approval, or recommendation. Look for the actual review, then check that it concerns the test version and delivery mode in question, identifies its intended population, and states what interpretations and uses the available evidence supports.

That distinction matters when the decision is consequential. The Standards for Educational and Psychological Testing frame validity around evidence and theory supporting a specified interpretation for a proposed use. Ask a qualified professional what evidence supports the interpretation for that decision, and consider the limits and alternatives described in the review.

For open, low-stakes practice, Test IQ Free offers an original adaptive 50-question test across pattern, number, and spatial reasoning, with worked answers. Its free result includes a provisional estimated IQ range, but the conversion has no validated adult reference population. The displayed band is therefore not a validated norm or a confidence interval, and should not be treated as an established comparison with adults. The optional $9 report explains the same provisional conversion and adds worked examples and practice structure; payment does not add accuracy or validity. That makes it suitable for practice feedback, not an established adult IQ comparison or a consequential decision.

Sources: EFPA Test Review Model – Version 2025 – EFPA; EFPA Test Review Model Version 2025, Version 5.0; Standards for Educational and Psychological Testing

Questions readers ask