What is my IQ? Choose a test that answers your actual question

0 The short answer

Your IQ cannot be established from a conversation, school grades or a personality description. You need a suitable cognitive test, scored against a stated reference group under appropriate conditions. Choose the route from the purpose: an institution may require a qualified professional and particular documentation; Mensa uses its own admission rules; personal exploration allows a wider choice of documented tests. A result should identify its scale, norm group and uncertainty. A test being free or paid does not establish its quality. This page compares the routes and explains what a score can actually answer.

1 Why Nobody Can Tell You Your IQ Without a Test

An IQ is a position, not a possession: it states where your performance on one instrument fell relative to a reference sample of known composition, and no position can be reported until the comparison has been made. A modern IQ is a deviation score. A test is given to a norm sample, the mean raw performance of each age band is set to 100 and the spread to 15 points, and every later score is a distance from that center, as the page on how IQ scores are normed explains. The number therefore depends on three things at once: the items, the sample and the scale. Change any of them and the number changes while the person does not. Mensa International's own testing FAQ, read September 20, 2026, puts the consequence in one line: the term IQ score is widely used but poorly defined, and a result of 132 on one test can be the same as a score of 148 on another. The same page adds that getting a straight answer to the question of what your IQ is may not be so simple.

Self estimates contain information but are much too uncertain to establish an individual score. Freund and Kasten pooled 154 effect sizes from 41 studies in Psychological Bulletin and found an overall correlation of .33 between self estimated and measured cognitive ability. Squaring .33 gives about .11, our arithmetic: the association leaves most measured variation unexplained. Under a simple linear prediction model with a target standard deviation of 15, the residual standard deviation would be about 14 points. That is an illustration of weak predictive precision, not a validated calculator for a person who guesses 120. A usable individual prediction would also need the regression equation, calibration sample and assumptions. The page on whether you can estimate your own IQ explains the distinction.

The question also hides different purposes. Someone may need evidence for a school, an accommodation or a membership; someone else may want personal information. For an institutional decision, obtain the receiving organization's requirements before arranging an assessment. It may specify the professional, instrument, recency, supporting history and format of the report. For personal exploration, assess the test's documentation and conditions rather than its price alone. The page on what an IQ test measures is a useful starting point if the prior question is what the number would mean.

2 The Four Routes and the Questions They Answer

These routes differ in purpose, supervision and the information returned; there is no single quality ranking that fits every reader. The International Test Commission guidelines on computer based and internet delivered testing distinguish open, controlled, supervised and managed administration. Those distinctions concern administration, not whether a test charges money. A free assessment can publish norms and enforce conditions, while a paid assessment can lack adequate evidence. The table describes typical routes, with the final row specifically referring to an undocumented quiz. Provider fees below are dated examples, not a complete market survey.

Two features of the table are counterintuitive. The first is that the route with the most supervision after the clinic, the Mensa session, produces the least information in the United States, because the society is selling admission rather than measurement and says so. The second is that a normed online battery outranks a Mensa session for self knowledge and falls below it for acceptance, because the two routes answer different questions. The page on where to take an IQ test lists the in person options by country, and the page on official IQ tests covers the first three routes from the angle of which paper each one produces. What follows takes them one at a time.

3 Route One: a Clinical Administration

A professional assessment can provide a signed interpretive report, but acceptance depends on the receiving institution and referral question. The WAIS-5 is one adult instrument available in the United States. Pearson's product page identifies its 2024 publication, adult age range and Level C qualification. Its qualification policy concerns access to restricted assessment materials; it is not a substitute for local professional licensing or competence in the intended assessment. The page on what the WAIS-5 is explains its structure. A clinician may instead select another appropriate instrument after reviewing the referral, language, age and assessment history. It is not a test anyone can take alone on a public website: an official WAIS online exists only when a qualified examiner runs the administration, as the guide to taking the WAIS online explains.

What you receive is a written report, not a number on a screen. The examiner states the referral question, describes the conditions of testing, reports each index score with its confidence interval and percentile, comments on the pattern, and answers the question. A professional signature is one element of the documentation; acceptance may also depend on the test, referral question, date and supporting evidence. The examiner also records behavior that no online system can see: whether an answer was reasoned or guessed, whether fatigue set in, whether a hearing or vision problem affected a subtest. That observation is part of what the fee buys, and the page on professional IQ tests covers what the fee buys hour by hour.

The price depends almost entirely on who signs. University training clinics, where doctoral students test under a licensed supervisor, publish the lowest fees. The University of South Florida Psychological Services Center lists a gifted assessment with the Wechsler child or adult scale at 100 to 400 dollars on its clinic policies page, read September 20, 2026. The fee is computed on a sliding scale by income, and half of the quoted amount is due as a deposit at the first meeting. Private practices charge several times that, and a full neuropsychological evaluation that adds memory, attention and executive measures costs more again. The page on how much an IQ test costs tabulates the market by provider type.

The limits of the route are practical. Training clinic availability and waiting lists vary, so confirm the local timetable. The report is written for a referral question, so a person who wants a broad profile for their own interest is paying for a service built to answer a different question. Prior administrations must be considered when planning a repeat assessment. Estevis, Basso and Combs retested 54 healthy adults on the WAIS-IV after three or six months and reported in The Clinical Neuropsychologist, 2012, volume 26, issue 2, pages 239 to 254, that Full Scale IQ rose by about 7 points on retest, with Processing Speed up about 9, and that the interval made little difference. A repeat assessment can still be useful, but its interpretation must account for prior exposure rather than treating it as an independent first administration, which is one reason clinicians ask when you were last tested before they begin.

4 Route Two: a Supervised Mensa Session

A Mensa session is designed for admission, and in the United States the stated testing purpose is to determine whether you reached the admission threshold on an accepted administration. American Mensa's testing page, read September 20, 2026, offers local group sessions proctored by volunteers at 60 dollars and private sessions at approved facilities, administered electronically, at 99 dollars. The test takes between one and two hours, requires a photo ID with proof of date of birth, and may be taken once every eight weeks. The society does not test people under 14. It then states two things in plain words. The Mensa Admission Test is given for the purpose of admission into Mensa and not to quantify intelligence. And the society is not able to provide a detailed report with scores, percentile ranks or an IQ score, citing Psychology Board of Texas regulations that prohibit releasing that information without a detailed interpretation. A person who asks the United States society what their IQ is receives, at most, the answer that it is in the top two percent or that it is not.

British Mensa runs the same threshold with different mechanics. Its IQ tests page, read September 20, 2026, lists a supervised test at 49 pounds, online under proctoring for one and a half hours or in person at a test center for two and a quarter hours, for candidates over the age of ten and a half. The in person option states that once marked you will receive your IQ score, and that anyone in the top two percent is invited to join. The society also sells a 45 minute Home Test at 15 pounds that returns an indicative score and qualifies for nothing, and a free online workout of up to 15 minutes that returns no score at all. Prior evidence, such as a full educational psychologist's report or documentary evidence of a professionally administered test, is assessed for 49 pounds over four to six weeks, and applicants must reside in the United Kingdom or Ireland. The page on the Mensa IQ test covers both societies' sessions in detail.

Mensa International's FAQ explains the rule behind both. Membership is open to persons who have attained a score within the upper two percent of the general population on an approved intelligence test that has been properly administered and supervised. An unsupervised online result is not the same as approved admission testing; British Mensa also offers a specifically supervised remote route. Many tests are not valid for people under sixteen. And because test items can leak online, Mensa International is in the final stages of implementing computerized adaptive testing, which draws items from a large bank matched to the candidate's ability estimate rather than using a fixed form. The same FAQ tells a person who was tested long ago, or who cannot obtain a certified copy of an old score, that the simplest course is to ask Mensa to test them again.

Read as an answer to what is my IQ, the Mensa route must be interpreted country by country. American Mensa's admission test establishes eligibility without providing the detailed numerical profile described above. British Mensa advertises score feedback for its supervised route. Neither should be assumed to provide the full clinical interpretation or supporting documentation that another organization requests. Check the relevant national society rather than transferring the rules from a different country.

5 Route Three: a Normed Online Battery

An online battery can measure, provided it publishes the three things that make any score interpretable, its norm sample, its reliability and its error, but documentation alone does not establish clinical or institutional acceptance. What separates a normed battery from a quiz is checkable before you pay. A normed battery states the size and composition of the reference sample against which your raw performance is converted to a standard score. It reports the reliability of each score it prints, and the standard error that follows from that reliability. It states its scale, meaning the mean and the standard deviation, and it describes how it protects its items and handles invalid attempts. The page on reliability and validity explains what each of those terms buys the reader, and the page on accurate IQ tests turns the list into a checklist. Breadth is the other check: a composite built from many subtests across domains is what separates a full scale IQ test from the single task quizzes behind most online results.

We sell ACIS, so its documentation is a disclosed example, not an independent recommendation. The technical manual, version 1.4, distinguishes an adult English speaking reference frame of 3,243 records aged 16 to 90 from an analysis set of 2,750 complete records. Its Full Scale table reports composite reliability .9886 and SEM 1.6029. Those figures belong to the documented analysis and score definition; they should not be transferred automatically to every shorter form, individual user or revised subtest configuration. The manual also contains index specific estimates. A reader should check the current report's form and version before applying them. The page on IQ tests with detailed results explains the difference between a broad score, an index and a subtest result.

The mode of administration matters alongside the norm sample and intended use. ACIS is self administered online, with its own access, timing and integrity controls. These do not reproduce a clinician observing the assessment. Mensa's published rules require an approved, properly administered test; an ACIS result should not be presented as approved admission evidence. For a school, court or accommodation request, ask the recipient about the required documentation. We have not established a universal acceptance policy across those institutions, and a self administered report is not a clinical diagnosis.

The price and time are the other differences. ACIS charges one time: 15 dollars for the Quick form of 6 subtests in about 45 minutes, 30 dollars for the Optimized form of 13 subtests in about 110 minutes, and 50 dollars for the Full Scale form of 20 subtests in about 175 minutes. Five subtests are free without a card, and a purchased form stays open for 30 days. Those are the home page figures on September 20, 2026, and they can change. The Full Scale form supplies the complete 20 subtest basis for the broadest FSIQ interpretation; shorter form estimates have different coverage, and the page on Full Scale IQ explains why a composite of 20 indicators is more reliable than any single part of it.

6 Route Four: an Undocumented Online Quiz

An undocumented quiz may return an IQ shaped number without enough information to interpret it, whether the quiz is free or paid. Look for the reference sample, conversion procedure, reliability, intended age range and administration conditions. A percentile calculated only from voluntary website visitors describes that participant group unless an appropriate norming procedure supports a wider interpretation. The absence of public documentation does not prove that the publisher used no data, but it prevents the reader from verifying the claimed population meaning. The page on free versus validated IQ tests explains why evidence quality and price are separate questions.

Reliability is another check. Harvill's instructional module explains the classical relationship between reliability, score variability and the standard error of measurement. If a score scale had SD 15 and reliability .70, its classical SEM would be about 8.2 points, our arithmetic. That is a conditional example, not an estimate of any named quiz. Item count alone does not establish reliability: item quality, information, construct breadth and the population also matter. A short, well designed test can outperform a longer poor one for a particular purpose. Ask for the evidence rather than assuming that a long test, a fee or a printed certificate guarantees precision.

Repetition can affect free and paid assessments alike. Scharfen, Peters and Holling pooled 174 samples from 122 studies with 153,185 participants in Intelligence, 2018, volume 67, pages 44 to 66, and found that the mere repetition of a cognitive ability test raised scores by 0.33 standard deviations from the first to the second administration and by 0.50 from the first to the third. There was no further gain after the third sitting. On a scale with a standard deviation of 15 those are gains of about 5 and 7.5 points, our arithmetic. Retest effects were larger by 0.15 standard deviations when the identical form was used again, and smaller when the interval between sittings was longer. These are average effects across the included studies, not a forecast for a particular quiz or person. Repeating familiar items and selecting the highest result can overstate first attempt performance; the change cannot be interpreted as the same amount of improvement in general ability. The clinical figure is consistent: the 7 point WAIS-IV gain that Estevis and colleagues measured came from a single repetition after months, under an examiner, on a full clinical battery.

Timing and identity controls should also be documented. When a task is intended to be timed, an unenforced limit changes the conditions from those used to interpret scores. Uncontrolled repeats, outside assistance and a mismatch between the account holder and the respondent can affect validity. These are risks to examine, not facts to assume about every free website. The page on free IQ tests describes what a free administration can provide, including the ACIS trial's limited subtest coverage. A partial trial is not equivalent to completing the full assessment.

What a free quiz is good for is familiarization. A person who has never seen a matrix item, a number series or a digit span task benefits from seeing the formats before a session that counts, and both national Mensas say so about their own practice materials, which they describe as practice that qualifies for nothing. The page on how to prepare for an IQ test separates the preparation that is legitimate from the preparation that contaminates a score. Use practice materials for their stated purpose. A documented free assessment may provide a meaningful score, but an undocumented entertainment quiz cannot establish the same interpretation.

7 What Will Not Tell You Your IQ

Six things that circulate as ways to find out your IQ without a test share one property: each predicts a measured score with an error large enough to span two or three classification bands, and the figures below are arithmetic on published correlations, not opinion. The method is the same in every row. If a variable correlates r with measured IQ, a simplified linear model with target SD 15 would have residual SD equal to 15 times the square root of one minus r squared. That is our arithmetic on the standard formula, and it assumes a straight line relationship on a scale with a standard deviation of 15. The table gives the correlation, its source, and the error.

Guessing is covered above: the .33 correlation from Freund and Kasten leaves a guess with an error of about 14 points. Grades are better and still not enough. Roth and colleagues pooled 240 independent samples with 105,185 participants in Intelligence, 2015, volume 53, pages 118 to 137, and estimated the population correlation between standardized intelligence tests and school grades at .54 after correcting for sampling error, measurement error and range restriction. A .54 correlation explains about 29 percent of the variance, our arithmetic, and the rest is effort, teaching, interest, the subject and the grading scale. An association across students cannot assign an IQ to one student without a calibrated prediction model and appropriate comparison group.

Personality quizzes tell you less again. Anglim and colleagues meta-analyzed 272 samples with 162,636 participants in Psychological Bulletin, 2022, volume 148, issues 5 to 6, pages 301 to 336, and found that openness, at a corrected correlation of .20, and neuroticism, at negative .09, were the strongest Big Five correlates of intelligence. Openness shares 4 percent of its variance with measured ability, our arithmetic. A quiz that reports your type and then your IQ has measured the first and invented the second, and the page on IQ and personality reviews the relationship in full.

8 Can an SAT Score Answer the Question?

SAT conversions are the strongest shortcut and still fail for most people who use them, for a reason that has nothing to do with the correlation. Frey and Detterman correlated SAT scores with a general factor extracted from the Armed Services Vocational Aptitude Battery for 917 participants in the National Longitudinal Survey of Youth 1979 and reported the result in Psychological Science, 2004, volume 15, issue 6, pages 373 to 378. The correlation was .82, or .86 corrected for nonlinearity, and from it they published equations for converting SAT scores to estimated IQs. Their second study, with 104 undergraduates, correlated the revised and recentered SAT with Raven's Advanced Progressive Matrices at .483, or .72 corrected for restriction of range. The equations belong to the SAT that the NLSY79 participants took, before the recentering that the second study already had to account for, and the test has been redesigned more than once since. A conversion table built on the 2004 paper and applied to a recent score is therefore estimating an IQ from an instrument the equation never saw. Even where it applies, a .82 correlation leaves an error of about 8.6 points, so 1.96 of those illustrative residual standard deviations would span roughly 17 points on either side of a prediction. This is not a validated personal SAT prediction interval: the applicable regression model, calibration and prediction procedure would be required. The pages on IQ versus the SAT and IQ versus the GRE give the version history of each test and what the published correlations do and do not support.

9 Childhood Scores, Norm Years and Age Tables

An old score has two problems, one about norms and one about you. The norm problem is drift. Trahan and colleagues pooled 285 studies with 14,031 participants in Psychological Bulletin, 2014, volume 140, issue 5, pages 1332 to 1360, and estimated the rise in raw performance at 2.31 standard score points per decade, with a 95 percent confidence interval of 1.99 to 2.64, and 2.93 points per decade for modern Wechsler and Stanford-Binet tests. Multiplying those historical mean rates by two gives roughly 4.6 to 5.9 points as an illustration, not a correction to apply to an individual old score. Trends differ by period, population and instrument, and the page on the Flynn effect works through the correction. The personal problem is that ability differences are stable and not fixed. Deary, Pattie and Starr gave 106 members of the Lothian Birth Cohort of 1921 the same Moray House Test at age 11 and again at age 90, and reported in Psychological Science, 2013, volume 24, issue 12, pages 2361 to 2368, a correlation of .54 between the two sittings, or .67 corrected for range restriction. That is remarkable stability across 79 years, and it still leaves a childhood score with an error of about 11 to 13 points as an estimate of the adult one. A score from a school test decades ago is evidence about the child who took it, on the norms of that year. The page on how IQ changes with age separates the stability of rank from the change in raw performance.

Age tables describe a reference framework, not a person. In an age normed test, raw performance is compared with an appropriate age group; the scale is ordinarily centered around 100 in that norm group. A newly sampled population need not have exactly that mean, and cohorts, selection, norm age and scoring methods can affect observed results. A raw score can change with age while an age adjusted standing remains similar. None of these group summaries tells you where a particular 25 year old or 70 year old falls. The page on average IQ by age separates raw performance from age adjusted scores.

10 How to Read the Number Once You Have It

A measured IQ arrives as four things, a score, an error, a percentile and a profile, and the score is the least informative of the four. The error comes first. Harvill's 1991 module makes the argument that the standard error, not the reliability coefficient, is what the reader of an individual score needs, because it is expressed in score points. Under the assumptions of a simple symmetric classical interval, repeated use of the procedure would cover the underlying score about 95 percent of the time. Some manuals use different, including regression based, interval methods, so use the interval printed by the relevant scoring system. Using the version 1.4 published ACIS Full Scale SEM of about 1.60 solely as a symmetric interval illustration gives a margin near 3.1 points: a reported 112 is about 109 to 115. The indices are wider. A Working Memory Index of 112 spans about 104 to 120, and a Processing Speed Index of 112 spans about 102 to 122, from standard errors of 4.12 and 5.22. The page on the 15 point standard deviation shows what a single point is worth on the scale.

The percentile comes second, and it is not linear. An IQ of 100 is the 50th percentile under a symmetric normal model centered at 100. A 115 is about the 84th, a 130 about the 98th, a 145 about the 99.9th, and an 85 about the 16th, all by our arithmetic from the normal curve. The same three points therefore move a person from the 50th to the 58th percentile near the middle and from the 98th to the 99th near the top. That is why the technical manual's own guidance is to report score, percentile and confidence context together rather than treating a percentile as a conclusion. The page on converting a score to a percentile explains the curve, and the IQ percentile calculator does the conversion for any score and standard deviation.

The profile can add useful information to the broad score. An uneven set of indices may suggest different task strengths and demands, but the differences require their own analysis. A simple independent error approximation uses the square root of the sum of the two squared SEMs. That approximation is not a universal ACIS significance rule: correlated errors, score specific reliability, multiple comparisons, base rates and the relevant form matter. Nor does a statistically reliable difference automatically identify a diagnosis or a useful intervention. Consult the manual and interpret the pattern with the referral question. The page on the six cognitive domains explains what the indices sample without treating a numerical gap as a diagnosis.

Practice effects belong in the interpretation of any repeated assessment. The retest studies cited above found average gains under their particular designs, but the mean effect is not a guaranteed increase for a person and does not isolate all causes of an individual change. Record the prior test, form, dates and conditions. A clinician may have a valid reason to reassess, using the appropriate method and retest guidance. Selecting only the highest result or assuming that a different instrument removes all prior exposure can overstate what the evidence shows.

11 Which Route Fits Which Question

Choose from the reader of the score, not from the price, and the decision between the four routes takes one minute. If a school, an employer's accommodation process, a disability board or a court will read the result, confirm which qualified assessment and supporting documents that particular organization accepts, and a university training clinic is the cheapest version of it if one is within reach and its waiting list is tolerable. If the purpose is Mensa, the society's session is the direct route, and in the United Kingdom it also returns a score, while in the United States it does not.

If the purpose is self knowledge, the question changes from which route is official to which measurement is honest. That rules out relying on undocumented scores and leaves both professionally administered and sufficiently documented self administered options, with different limits. The clinical route gives a supervised measurement and a professional's reading of it, at a price in the hundreds of dollars and after a wait. A normed battery gives a controlled mode measurement with a published error, at a price in the tens of dollars, without supervision and without a signature. A person who wants a profile across domains for their own use is well served by the second. A person who wants a document, or who suspects a clinical question behind the curiosity, needs the first. The page on IQ tests for adults works through the choice by situation.

If the assessment concerns a child or teenager, choose an instrument and practitioner suitable for the exact age and purpose. Adult scales cannot be assumed suitable for every child. Some adolescent ages fall within an adult scale's stated range, child scales cover other ranges, and society testing has its own eligibility rules. A school psychologist or pediatric assessment service can help identify the appropriate route. Do not apply the adult examples on this page to a younger child without checking those requirements.

If the purpose is employment, follow the actual employer's process. An employer may use a specific selection assessment whose job relevance, conditions and interpretation are established for that use. A personal IQ certificate does not automatically satisfy such a process, and a website's consumer score should not be substituted for the employer's requested evidence. A person who simply wants to explore their own profile before applying is asking a different, personal information question.

12 Where ACIS Sits, and What a Buyer Should Do

We sell ACIS, so read this section as a product disclosure. ACIS offers a timed, self administered online assessment with a published technical manual. The Full Scale form covers 20 subtests across six domains; shorter forms provide different coverage. Its documented reference frame and analysis sample are distinct, and the manual explains the published score level reliability estimates. An online result supports only interpretations justified by that evidence and the testing conditions. It does not itself establish a clinical diagnosis, disability determination or eligibility for an institution. The fact that a manual is public helps a reader inspect the claims; it does not remove the need for that inspection or make the commercial publisher independent.

The prices are one time, and they were read on the home page on September 20, 2026. The Quick form, 6 subtests, is 15 dollars and takes about 45 minutes. The Optimized form, 13 subtests, is 30 dollars and about 110 minutes. The Full Scale form, 20 subtests, is 50 dollars and about 175 minutes, and it provides the complete 20 subtest basis for the broadest Full Scale interpretation. A Custom form lets a buyer select subtests and is priced per subtest on the order page, and any domain completed in full returns its index score. Five subtests can be taken free without a card, access stays open for 30 days, and a five day quality guarantee covers the purchase. There is no subscription.

Two further limits belong here. The manual is in revision, and version 1.4 remains the reference. One subtest, Spatial Navigation, was retired on September 7, 2026 and replaced by Layer Rotation in the Visual Spatial Index, and the affected coefficients need to be interpreted with that change in mind rather than treated as a new validation of the revised form. A buyer today should read that notice before weighting the Visual Spatial score. And ACIS is English only, so a person tested in a language they do not use daily is being compared with a frame that does not describe them.

What a buyer should do follows from the question. If you need paper, book the clinic and treat this page as a map. If you want Mensa, book the session, and take the five free subtests first if you want to see a profile before paying anyone. If you want to know where you stand across domains at a stated precision, read the manual, take the free subtests, and buy the form whose indices answer your question. Before relying on a number, check the evidence for its intended use. If the reference group, uncertainty and conditions cannot be established, its individual meaning remains unclear.

13 Sources Behind This Page

Every figure above is traceable to one of the following, and each is linked at the point where it is used. Prices, fees and rules were read on the providers' own pages on September 20, 2026 and will change. The standard errors of estimate, the 95 percent intervals, the percentages of variance and the point equivalents of standardized effect sizes are our arithmetic on the published coefficients and are labeled as such where they appear.

• Mensa International. Getting your IQ tested: FAQs. mensa.org, read September 20, 2026.
• American Mensa. Take the Mensa Admission Test. us.mensa.org, read September 20, 2026.
• British Mensa. IQ testing and puzzles. mensa.org.uk, read September 20, 2026.
• Pearson. Wechsler Adult Intelligence Scale, Fifth Edition (WAIS-5), product page with publication date, age range, qualification level, completion time and kit prices, and the qualification levels policy defining level C. pearsonassessments.com and qualifications policy, read September 20, 2026.
• University of South Florida Psychological Services Center. Clinic policies and fees. usf.edu, read September 20, 2026.
• International Test Commission. International Guidelines on Computer-Based and Internet-Delivered Testing. International Journal of Testing, 2006, volume 6, issue 2, pages 143 to 171.
• Freund P A and Kasten N. How smart do you think you are? A meta-analysis on the validity of self-estimates of cognitive ability. Psychological Bulletin, 2012, volume 138, issue 2, pages 296 to 321.
• Roth B, Becker N, Romeyke S, Schäfer S, Domnick F and Spinath F M. Intelligence and school grades: A meta-analysis. Intelligence, 2015, volume 53, pages 118 to 137.
• Anglim J, Dunlop P D, Wee S, Horwood S, Wood J K and Marty A. Personality and intelligence: A meta-analysis. Psychological Bulletin, 2022, volume 148, issues 5 to 6, pages 301 to 336.
• Frey M C and Detterman D K. Scholastic assessment or g? The relationship between the Scholastic Assessment Test and general cognitive ability. Psychological Science, 2004, volume 15, issue 6, pages 373 to 378.
• Scharfen J, Peters J M and Holling H. Retest effects in cognitive ability tests: A meta-analysis. Intelligence, 2018, volume 67, pages 44 to 66.
• Estevis E, Basso M R and Combs D. Effects of practice on the Wechsler Adult Intelligence Scale-IV across 3- and 6-month intervals. The Clinical Neuropsychologist, 2012, volume 26, issue 2, pages 239 to 254.
• Trahan L H, Stuebing K K, Fletcher J M and Hiscock M. The Flynn effect: A meta-analysis. Psychological Bulletin, 2014, volume 140, issue 5, pages 1332 to 1360.
• Deary I J, Pattie A and Starr J M. The stability of intelligence from age 11 to age 90 years: The Lothian Birth Cohort of 1921. Psychological Science, 2013, volume 24, issue 12, pages 2361 to 2368.
• Harvill L M. Standard error of measurement. Educational Measurement: Issues and Practice, 1991, volume 10, issue 2, pages 33 to 41.
• ACIS. Technical manual: structure, norming, reliability and validity, version 1.4 with the revision notice of September 7, 2026. acisiq.com/technical-manual.

The Standards for Educational and Psychological Testing (AERA, APA and NCME, 2014) and the APA Ethical Principles of Psychologists and Code of Conduct, Assessment standards provide the interpretive framework: evidence must support the use, the reference group and limitations must be stated, and assessment results require appropriate explanation. Citing these standards does not imply their authors endorse ACIS or another product.

14 Frequently Asked Questions