Retaking an IQ test: how much scores rise and what to trust

0 The short answer

Yes, you can retake an IQ test, but the second score will usually run higher for reasons that are not all ability, so the useful questions are how long to wait, which test to use and whether the change is larger than error. In a meta-analysis of 122 studies and 153,185 test takers, scores rose by 0.33 standard deviations on the first retest, which is about 5 IQ points in our arithmetic, and the average did not climb further from the third to the fourth attempt. The real options are a psychologist giving a Wechsler battery, a supervised society exam with its own retest rule, and a normed online battery such as ACIS, which reports a Full Scale IQ with its 95 percent confidence interval for 15, 30 or 50 dollars one time (prices read on October 29, 2026, and they can change) in about 45, 110 or 175 minutes. ACIS is online, unsupervised, not clinical and in English only, and a retake never certifies a "true" score.

1 Can You Retake an IQ Test, and Is There a Limit on How Many Times?

Nothing in psychology forbids retaking an IQ test, and the limit on how many times is set by the provider's rules and by the shrinking information a repeat carries, not by any property of your mind. The phrase people type most often is "can you take an IQ test more than once", and the plain answer is yes. Hausknecht and colleagues (2007) open their meta-analysis in the Journal of Applied Psychology by noting that, according to earlier studies, as many as 25 to 50 percent of applicants in organizational and educational settings are retested on cognitive ability measures. The question is never whether a second attempt is possible. It is what the second number means.

Who may retake, and when, is a matter of policy. The Standards for Educational and Psychological Testing (AERA, APA and NCME, 2014), which we read in their open access edition, say in a comment on the standards for test administration that retest policies should be established by the test developer or user, that the administrator should follow them and that they should be clearly communicated as part of the conditions for standardized administration. They add that retesting is intended to decrease the probability that a person will be incorrectly classified as not meeting some standard. Providers therefore differ. American Mensa's retest page, which we opened on October 29, 2026, says that people who took its Admission Test and failed to qualify may retest after eight weeks. A press release dated April 8, 2024 describes the change and says the qualifying line is unchanged at the top 2 percent on a standardized IQ test. Neither page states a fee or a maximum number of attempts, so we say nothing about either. A free online screener, a clinic and a job-screening platform each set their own rules, and the only reliable source for those rules is the provider's own terms. This page states no retake terms for any product.

The second limit is informational. The same meta-analysis that measured the first gain also measured later ones. Between the second and third administrations the average gain was 0.17 standard deviations, and between the third and fourth it was 0.02, with a 95 percent confidence interval running from minus 0.020 to 0.062. In plain terms, the average score stops climbing after about three attempts. That does not mean the score becomes accurate at that point. It means that the extra attempts mostly reflect familiarity with that particular test, and a fourth attempt adds very little that the third did not already show.

Two nearby questions belong to other pages and stay there. Whether intelligence itself changes across the lifespan is covered by the page on how IQ changes with age. Whether training can raise measured ability, and which claims survive scrutiny, is covered by the page on what training really moves. How to prepare, and what helps or hurts before a first attempt, is covered by the guide to preparing for an IQ test. This page owns one decision: whether and when to take an IQ test again, and how to read the pair of scores that results.

2 How Much Do IQ Scores Rise When You Take the Test Again?

On average, scores rise by about a quarter to a third of a standard deviation on the first retest, which is roughly 4 to 5 IQ points in our arithmetic, and the spread around that average is wide enough that some people score lower the second time. The best single source is the multilevel meta-analysis by Jana Scharfen, Judith Peters and Heinz Holling, published in Intelligence in 2018. It pooled 174 samples from 122 studies, 786 test outcomes and 153,185 test takers (172 of those samples contribute to the first retest comparison), applying strict inclusion criteria for study design, participant age, health status and the cognitive ability tests used. Its effect size is the standardized mean change, the gain in a group divided by a standard deviation of the raw scores, so it is expressed in standard deviations and not in IQ points.

The last column is ours. It multiplies each effect by 15 on the assumption that the test is scaled like an IQ test, with a standard deviation of 15, and the authors do not report it. The conversion is approximate because the effect is standardized against the spread of raw scores in each study, and the pooled tests differ in content and scaling. The page on why the IQ standard deviation is 15 explains what a standard deviation means on that scale. Two readings follow from the table. The first retest carries most of the gain: 0.33 of the 0.50 reached by the third attempt. And the gain from the second to the third attempt is real but about half as large, which is what one would expect if a skill is being learned and then saturates.

The authors also report how unevenly the average applies. The standard deviation of the true effect for the first retest was 0.32, which led them to a prediction interval for a new study of minus 0.30 to 0.96 standard deviations. That is their calculation. In IQ points by our arithmetic it spans from a drop of about 4.5 points to a gain of about 14. The average gain is a property of groups. For one person, a second score a few points lower than the first is entirely ordinary.

Other meta-analyses agree on the order of magnitude. Hausknecht and colleagues, in their 2007 paper, analyzed 50 studies of practice effects for cognitive ability tests and report that results from 107 samples and 134,436 participants revealed an adjusted overall effect size of .26. In the authors' accepted manuscript, which we read on the Cornell University repository, the sample weighted mean was .24 before correction for sampling error, the median interval between the first and second test was 20 days and about 90 percent of the tests were group administered on paper. They summarize it as score improvements of about one quarter of a standard deviation across the first and second administrations. Matthew Calamia, Kristian Markon and Daniel Tranel reviewed neuropsychological tests in The Clinical Neuropsychologist in 2012, pooling nearly 1600 individual effect sizes for changes in mean performance, and found that the use of alternate forms, the age of participants, clinical diagnoses and the length of the retest interval were associated with the size of the change in many cases.

For a Wechsler battery specifically, the best direct evidence we found is a single study. Eduardo Estevis, Michael Basso and Dennis Combs gave the WAIS-IV twice to 54 adults with a mean age of 20.9 years, a mean of 14.9 years of education and an initial Full Scale IQ averaging 111.6, with the second session 3 or 6 months later. The Full Scale IQ rose by approximately 7 points, Verbal Comprehension by 5, Working Memory by 4, Perceptual Reasoning by 5, Processing Speed by 9 and the General Ability Index by 6, and the gains were similar at both intervals. They concluded that the gains reflect practice effects and not genuine intellectual change. One sample of young adults with above average starting scores cannot be extrapolated to everyone, and a figure of 7 points sits above the 5 points implied by the larger meta-analysis. The honest summary is a range, not a number: expect the second score to run roughly 4 to 7 points higher on average, with a spread of several points either side.

3 Why Do Scores Rise on a Second Attempt, and Is It Real Ability?

Scores rise for three different reasons, only one of which could be a real change in the ability the test is meant to measure, and the research does not yet say whether the first or the second score better represents a person. Scharfen and colleagues organize the causes into three groups, following earlier work by Lievens and colleagues. The first possible cause is that the latent construct itself is enhanced by retesting, an idea related to the testing effect in learning research, where retrieval practice strengthens memory. The authors argue that studies showing that the validity of scores changes after retesting speak against this being the main cause, and that because cognitive ability is defined as a stable construct, an improvement due to retesting is viewed critically.

The second cause is the reduction of construct-irrelevant factors: test anxiety, lack of understanding of the task and lack of familiarity. A person who is less anxious, or who understands the rules fully, can show more of what they can do. Scharfen and colleagues cite evidence that these factors contribute to the retest effect. The third cause is the development of test-taking strategies and test-specific skills, together with memory for particular items. A person who has seen a task type learns what it rewards, and with the same form may also remember answers. The authors cite empirical evidence that strategies are developed when retesting takes place and that using them raises scores.

Only the first cause would be an ability gain, and even that is contested. The second is a better measurement of an unchanged ability, because it removes noise from the first attempt. The third is a score gain with no ability change at all. Hausknecht and colleagues make a related point in their manuscript: practice effects still occurred when alternate forms were used, which suggests that score gains are not simply recall of items, and mere repetition accounted for some portion of the practice effect. Estevis and colleagues reached the plainer conclusion for the WAIS-IV, that the gains reflect practice effects and not real intellectual change.

Scharfen and colleagues state the consequence directly in their discussion of impact: one cannot be sure whether a score from a repeated or an initial test is more representative of a person's latent ability. They warn that, in clinical settings, a patient could reach a higher score on a repeated battery without being recovered, simply because the test had been taken before. That is the reason this page makes no promise that retaking raises a "true" score. A practice gain is a gain on that test, in that format, for that person, and it need not transfer.

A fourth force deserves a mention, because it affects people who choose to retake: regression to the mean. People who decide to retest are often those whose first result disappointed them, and a score that sat low by chance tends to move toward the person's typical level on a second try even if nothing else changes. Hausknecht and colleagues estimated that less than 10 percent of the total effect could be attributed to regression to the mean in the two studies that reported complete data, and warned that the share would be larger for a more selected group of repeat test takers.

The Standards address the same point from the side of test publishers. Standard 1.7 says that if test performance is claimed to be essentially unaffected by practice and coaching, the propensity for performance to change with those forms of instruction should be documented. Its comment adds that interpretive materials should summarize the evidence on how much improvement with practice or coaching can be expected, and that materials for test takers should give practical guidance about the value of test preparation. The page on reliability and validity explains why a score can be consistent from one occasion to the next and still fail to measure what it is supposed to measure, which is the distinction a practice gain puts under pressure.

4 What Makes the Gain Bigger or Smaller: Form, Interval, Number of Attempts and Ability?

The gain depends most on whether you see the same form again and how many times you have already taken it, a little on the length of the gap, and surprisingly little on which ability is tested or how old you are at the first retest. The table collects what the two meta-analyses report for each factor. Where a number comes from the Hausknecht manuscript and not from its published abstract, the table says so.

The most practical row is the first. A second attempt on the same form combines three things: familiarity with the format, strategies learned and memory for particular items. An alternate form removes the last one, and the meta-analysis shows that this matters most early on, at the first and second repetitions, and less after that. Even so, alternate forms do not remove the gain. Scharfen and colleagues conclude that the recommendation to use parallel instead of identical forms was supported by their results as long as only a few retests are administered, but that alternate versions still produced retest effects, so any earlier experience with a similar cognitive test can account for higher scores. A related warning comes from the Standards in their discussion of reliability. The correlation between two scores on the same form can be inflated by the test taker's recall of initial responses, which is one reason test-retest coefficients on a single form are an imperfect guide to how stable a score is.

The interval row is less dramatic than people expect. A 50 week gap lowered the average effect by only 0.04 standard deviations in the Scharfen meta-analysis. The authors' own extrapolation, that the first retest effect would vanish after about 426 weeks or 8.19 years, is a linear projection and not a measurement. Hausknecht and colleagues found an interval effect only for identical forms, and concluded that organizations might minimize memory-based practice effects with a minimum retest interval of at least one year.

The ability row corrects a common belief that speeded tasks are immune or that reasoning tasks are. At the first retest, none of the differences among ability types was statistically significant, so no kind of test was shown to be immune. The only significant ability difference the authors report is later, between the third and fourth administration, where processing speed tasks stopped gaining. The CHC names behind these labels are set out in the guide to the CHC model.

Gains from structured coaching were much larger in the Hausknecht data than gains from repetition alone, but coaching teaches strategies that serve the test and not necessarily the ability. Our own reading is that a reader who intends to retake should treat the first score as the cleaner measurement and the second as the one carrying a practice allowance.

5 How Long Should You Wait Before Retaking an IQ Test?

No source we opened sets a universal waiting period, but in the studies we read a gap of weeks, and even six months, did little to remove a practice effect, and a year or more is what it takes for same form gains to approach alternate form gains, so wait as long as the decision allows and never read a quick repeat as fresh information. Start with what we did not find. Neither the Standards, the American Psychological Association's ethics code nor the Pearson documents we opened state a minimum interval for repeating an intelligence test. Intervals appear in the research, in publishers' own retest studies and in the rules of individual programs, and these answer different questions.

The research intervals are the first source. Hausknecht and colleagues found that the median gap between the first and second test in their data was 20 days, with a range from 15 minutes to 5.18 years, so most of the evidence on practice effects concerns short gaps. Estevis and colleagues compared 3 and 6 months for the WAIS-IV and found similar gains at both, which means six months was not long enough to remove the effect in that sample. Scharfen and colleagues found only a small average decline of the effect with longer gaps, and Hausknecht and colleagues estimated that for identical forms it would take roughly 685 days, just under two years, for the gain to fall to the alternate form level.

The publishers' intervals are the second source, and they are short on purpose. Canivez and colleagues report in a 2026 paper in Assessment that the WAIS-5 technical manual reports short-term test-retest stability for subtests, index scores and ancillary scores, with a mean interval of 29 days and a range of 12 to 124, for several standardization subgroups, and that it includes stability coefficients, corrected coefficients and standardized differences across the retest interval. They add that, at the time of their paper, long-term stability studies were not yet available for the WAIS-5. A 29 day interval is a design choice for measuring stability. It is not a recommended waiting period for a person who wants a clean second score.

The program rules are the third source. American Mensa lets people who failed to qualify retest after eight weeks. That is a policy, and the pages we read do not say what evidence it rests on, so it should not be read as a scientific minimum. The right interval also depends on the purpose. The American Psychological Association's Ethical Principles of Psychologists and Code of Conduct, in its 2017 text, says in standard 9.08(a) that psychologists do not base assessment or intervention decisions or recommendations on data or test results that are outdated for the current purpose. The Standards make the same point in their chapter on psychological testing and assessment: professionals are responsible for guarding against reliance on outdated test scores, and in such cases retesting is appropriate. A score from many years ago can be too old for a present decision, while a score from last week can be too practiced to compare with a new one.

The adult IQ testing process page covers who gives a professional test, what it costs and how long the wait is. The sensible default for a curious adult is simple. If the first attempt went normally, wait several months before a second attempt, use a different form or test if you can, and treat a change of only a few points as within noise.

6 Which Options Exist for Taking an IQ Test Again, and How Do They Compare?

A retake can happen in four kinds of setting that differ in who supervises, what the norm group is and what a repeat means, and the strongest comparison comes from the setting that keeps conditions and norms constant. The table lists the real options. Prices and timings are those we read on October 29, 2026 where we opened the page; where a cell says not stated, the pages we read did not say.

Read the table by the question you are asking. If you need a score a third party will accept, such as a clinician, a school or an employer, the first row or a program that requires its own test is the route, and the guide to official IQ tests separates the routes that count from those that do not. If you want a normed number for yourself, the page on every way to take an IQ test, the comparison of the best online tests and the page on IQ tests for adults cover the options, and a free IQ test is a fair way to see the format before paying for anything. The free row deserves a plain note. A free screener may have few items per score and a norm group the page does not define, so check both before reading a repeat, because a repeat of a short quiz partly measures how well you remember it.

7 What Can a Retest Tell You, and What Can It Not?

A retest can show whether a first result was a one-off, whether two instruments agree and whether a change exceeds error, but it cannot certify a true score, because the second administration carries a practice effect the first did not. The Standards describe three situations in which communications to test takers should state a retest policy: when major disruptions occur during administration, when the test taker feels the present performance does not appropriately reflect true capabilities, and when the test taker improves on the underlying knowledge, skills, abilities or other characteristics. Those are the legitimate reasons. A noisy room or an interruption is a plain reason to doubt a first result, and a second attempt in better conditions is a reasonable response.

A retest on a different instrument tells you something else. It is convergent evidence: two instruments built by different teams, normed on different groups, pointing to the same region. It is not a retest in the strict sense, because the norms, the subtests and the scale may differ. The page on why a WAIS-5 score can drop compared with WAIS-IV shows how the same person can land on a different number simply because the edition changed, and the page on how IQ scores are normed explains why a score is always relative to a reference group.

What a retest cannot do is turn the higher of two scores into the real one. Reporting the best of several attempts mixes practice with chance. Our reasoning is simple: when you keep the highest number from several tries, the selection pushes the reported value up on top of the practice effect, and the result describes a peak, not a typical performance. If you do report more than one attempt, say how many there were and which came first. The Standards also note that testing programs sometimes treat large changes in scores upon retesting as one of several indicators that may trigger careful scrutiny of a result, which is another reason to report attempts openly.

A retest of a single index, such as one speeded subtest, is a noisy comparison, because a short set of tasks has a wider error band than a full battery. The page on index score differences explains how common large gaps between indices are in the general population, and why a low index is not automatically a weakness.

8 How Do You Compare Two IQ Scores With Their Error?

A change between two scores means something only when it is larger than the combined measurement error of both scores plus the practice gain you would expect anyway, and a simple worked example shows how often that is not the case. The Standards put the principle in Standard 2.4: when a score interpretation emphasizes differences between two observed scores of an individual, reliability and precision data, including standard errors, should be provided for those differences. The comment adds that the precision of change scores can be much lower than the precision of the separate scores involved.

We can illustrate with a real report layout. Pearson's sample WAIS-5 score report, dated September 16, 2024, which we opened on October 29, 2026, lists for its invented examinee a Full Scale IQ of 111 with a 95 percent confidence interval of 106 to 116 and a standard error of measurement of 3.00. The rest is our arithmetic and a teaching approximation, not a clinical method. If both of a person's scores carry an error of 3.00 and the errors are independent, the standard error of the difference is the square root of 3.00 squared plus 3.00 squared, about 4.24. Multiplying by 1.96 gives about 8.3 points: a difference smaller than that is within what measurement error alone produces. A second step subtracts the practice gain. The first retest gain in the Scharfen meta-analysis, 0.327 standard deviations, is about 5 points when multiplied by 15.

Three cautions apply. First, the practice gain is an average, and the authors' prediction interval for a single study ran from minus 0.30 to 0.96 standard deviations, so the right adjustment for one person is uncertain and the threshold is softer than the table suggests. Second, Estevis and colleagues computed reliable change two ways for the WAIS-IV, with a simple difference and with a regression method, and found that the regression method gave more accurate estimates, so published change tables from a test's own manual should replace this approximation when they exist. Third, an interval that you can read directly is more useful than a calculation. If two reports each give a score of plus or minus 5, the combined error is about 7 points, our arithmetic from the square root of 5 squared plus 5 squared. Pearson's report also prints critical values for comparing two composites from the same sitting, such as 8.81 for the General Ability Index against the Full Scale IQ, but those are not retest thresholds.

The page on score versus percentile explains why a change in points does not translate into a constant change in percentiles, and why you compare standard scores and not percentiles. An index, which rests on fewer subtests than the Full Scale IQ, carries a wider band, as the Pearson report shows with standard errors of measurement from 3.00 to 5.61 across its five primary indices.

9 Which Abilities Shift Most on a Retake, and What Do the ACIS Subtests Look Like?

The first retest gain was statistically similar across processing speed, memory and reasoning tasks, so no ability is immune, and what differs between the abilities is which index deserves the most caution in a quick repeat. The ACIS subtest pages describe the task formats, and we use only what they state. Coding gives 120 seconds to match symbols to numbers with on-screen buttons. Symbol Search gives 80 puzzles with a 2 minute limit and a penalty for wrong answers, so speed trades against accuracy. Both feed the processing speed index, the Gs domain in CHC terms. Digit Span plays digits at one per second and asks for recall forward, backward and in sequence order; it feeds working memory, the Gwm domain, and needs working sound. Matrix Reasoning shows 32 visual matrices with a missing piece and allows 30 seconds per item; it feeds fluid reasoning, Gf. Each page describes the task and not its content, and so does this one.

The evidence says two things about these abilities. At the first retest, the meta-analysis found no significant difference among processing speed, memory and reasoning tasks, with effects from 0.28 to 0.37 for each. In the one WAIS-IV study we cite, the largest gain was on Processing Speed, 9 points, and the smallest was on Working Memory, 4 points, in a sample of 54 young adults, which is suggestive and not a rule. A task with a simple rule and a clock, such as a symbol matching task, may reward a learned routine quickly. That is our hypothesis, consistent with the strategy mechanism described earlier, and we did not find a source that tests it for these subtests.

For a reader, the practical consequence is to compare the Full Scale IQ, which rests on all the subtests, before reading any single index, and to treat a change in one index as weaker evidence than a change in several. The page on what IQ measures describes how the abilities relate to the general factor, and the processing speed test page shows what a short normed measure of that index looks like.

10 Should You Retake an IQ Test to Reach a Cutoff Such as Mensa or the Top 5 Percent?

Retaking to cross a cutoff can work as simple arithmetic, because a practice gain of several points can lift a score over a line, but a score lifted by practice is the least informative score to report, and the first repeat gives the largest lift. Scharfen and colleagues make the same point about cutoffs in other settings: a score gain from retesting could change an outcome because a cutoff value is exceeded that had not been reached before. American Mensa's testing page, read on October 29, 2026, defines a qualifying score as at or above 98 percent of the general population, and its retest rule lets those who failed to qualify try again after eight weeks. A rise of about 5 points on a repeat is enough to carry a score across a line for anyone who started within 5 points of it. That is the arithmetic, and it is one reason programs write explicit retest rules.

For a personal question, such as whether you belong to the top 5 percent of adults, the reasoning runs the other way. If the first attempt was clean, the first score is the better estimate and a second attempt is more likely to flatter than inform. If you are close to a line, the interval matters more than the point estimate, and the percentile tools turn a standard score into a rank: the IQ percentile calculator and the page on the top 5 percent of adults cover the translation, and the guide to gifted tests for adults covers what a result near such a line does and does not establish. ACIS is not meant for admission to high IQ societies or for school accommodations, so it cannot be the retake that counts for either.

11 How Should You Set Up a Second Attempt So the Comparison Is Fair?

A fair second attempt holds conditions constant, uses a different form or test when one exists, and records everything that differed, because a comparison is only as clean as its least controlled condition. The Standards warn test users to be aware of the implications of less controlled administration conditions, and an unsupervised online session is exactly that. The practical list is short. Use the same device and the same type of room. For subtests that play audio, check the sound beforehand. ACIS lists Digit Span, Alphanumeric Sequencing and Arithmetic as audio subtests that need working sound. Take breaks the way you did the first time; ACIS allows breaks, saves progress between subtests and gives 30 days to finish. Write down the time of day, how rested you felt, what interrupted you and which form you used.

Do not look up the items or answers between attempts, and do not rehearse the specific tasks, because that converts a measurement into a memory test. The Standards list among test takers' responsibilities following the directions and representing themselves honestly. If anxiety was the problem on the first attempt, the research suggests a calmer second attempt can produce a more accurate score, which is one of the three mechanisms above, not a gain in ability. If time pressure was the issue, the page on untimed IQ tests looks at the question of format, and the page on what makes an IQ test scientific lists what to ask of any test before reading its second score.

12 Where ACIS sits, and what a buyer should do

ACIS is a normed online battery that reports a Full Scale IQ with a 95 percent confidence interval, and it is useful for a second look only when you treat the interval, the form and the gap as part of the answer. The Quick form (6 subtests in 3 domains, about 45 minutes) costs 15 dollars one time, the Optimized form (13 subtests in 5 domains, about 110 minutes) costs 30 dollars and the Full Scale form (all 20 subtests in six domains, about 175 minutes) costs 50 dollars, with prices read on October 29, 2026 that can change. The Custom Test costs 5 dollars per subtest, and any domain you complete returns its full index score. There is a free trial with no card, an example report and a technical manual for documentation, and this page quotes no figure from the manual.

The limits are plain. ACIS is online and unsupervised, it is not a clinical or diagnostic instrument, it is not for hiring, school accommodations or admission to high IQ societies, and it is in English only. The technical manual says that verbal and knowledge tasks are affected by English proficiency and educational exposure. This page also states no retake terms for ACIS or any other provider. As the Standards for Educational and Psychological Testing (AERA, APA and NCME, 2014) say, retest policies belong to the developer or user and should be clearly communicated, so read the terms on the page where you buy. The APA ethics code adds that psychologists do not base decisions on results that are outdated for the current purpose, and the same logic applies to your own use: a score is evidence for a purpose, and a practiced score is weaker evidence for ability.

The concrete next step depends on where you are. If you have not taken a normed test, take one once, under good conditions, and read the interval before the point estimate. If you have taken one and doubt it, write down why, wait months if you can, and compare the two with the error method above. If you need a score for an institution, ask what it accepts before you pay for anything.

13 Sources Behind This Page

Every figure on this page comes from one of the documents below, which we opened or verified on October 29, 2026, and every calculation of ours is labeled as our arithmetic. Details beyond the published abstract of Hausknecht and colleagues come from the authors' accepted manuscript, so they should be checked against the journal version before being quoted elsewhere.

• Scharfen, J., Peters, J. M., & Holling, H. (2018). Retest effects in cognitive ability tests: A meta-analysis. Intelligence, 67, 44-66.
• Hausknecht, J. P., Halpert, J. A., Di Paolo, N. T., & Moriarty Gerrard, M. O. (2007). Retesting in selection: A meta-analysis of coaching and practice effects for tests of cognitive ability. Journal of Applied Psychology, 92(2), 373-385.
• Hausknecht, J. P., Halpert, J. A., Di Paolo, N. T., & Moriarty Gerrard, M. O. (in press at the time). Retesting in selection: A meta-analysis of practice effects for tests of cognitive ability. Accepted manuscript, Cornell University eCommons.
• Calamia, M., Markon, K., & Tranel, D. (2012). Scoring higher the second time around: Meta-analyses of practice effects in neuropsychological assessment. The Clinical Neuropsychologist, 26(4), 543-570.
• Estevis, E., Basso, M., & Combs, D. (2012). Effects of practice on the Wechsler Adult Intelligence Scale-IV across 3- and 6-month intervals. The Clinical Neuropsychologist, 26(2), 239-254.
• Canivez, G., Watkins, M., McGill, R., & Dombrowski, S. (2026). Construct validity of the WAIS-5: Complementary exploratory and confirmatory factor analyses of the 20 primary and secondary subtests. Assessment, online first.
• Pearson. (2024). WAIS-5 sample score report (report dated September 16, 2024). NCS Pearson.
• American Educational Research Association, American Psychological Association, & National Council on Measurement in Education. (2014). Standards for Educational and Psychological Testing. AERA. Open access edition, testingstandards.net.
• American Psychological Association. (2017). Ethical Principles of Psychologists and Code of Conduct (including 2010 and 2016 amendments), standard 9.08.
• American Mensa. Retest page, read October 29, 2026.
• American Mensa. (April 8, 2024). Change allows candidates to retake American Mensa Admission Test. Press release.
• American Mensa. Testing page (regular prices and qualifying score), read October 29, 2026.

14 Frequently Asked Questions