What makes an IQ test actually scientific

0 The short answer

A scientific IQ test is one whose publisher documents, in a form you can read, who the norms came from, how large the measurement error is for every score you are shown, what evidence supports each interpretation and under what conditions the test must be given, and no label, including the word "scientific", can replace that documentation. The real routes are a psychologist giving an individually administered battery such as the WAIS-5, which Pearson lists at about 45 to 60 minutes of testing, and a normed online battery such as ACIS, which costs 15, 30 or 50 dollars one time for about 45, 110 or 175 minutes (prices read on October 29, 2026). The decisive limit is that an online test can publish its norms and its error but cannot control your room, so ACIS is online, unsupervised, not clinical and in English only, and it is built for personal use, not for hiring, school accommodations or a high IQ society.

1 What Does "Scientific" Mean for an IQ Test?

"Scientific" is not a property a test can claim for itself; it describes a test whose scores come with checkable evidence for a stated use, and the Standards for Educational and Psychological Testing say so by refusing to speak of the validity of a test in general. The Standards for Educational and Psychological Testing (AERA, APA and NCME, 2014), which we read in their open access edition on October 29, 2026, ask in Standard 1.0 for a clear statement of each intended score interpretation for a specified use and for validity evidence supporting each one. The comment on Standard 1.1 goes further. It calls the unqualified phrase "the validity of the test" incorrect, and it says no test permits interpretations that are valid for all purposes or in all situations. A test is therefore never scientific in the abstract. It is documented well or badly for a purpose, a population and a set of conditions.

That reading changes the question a reader should ask. "Is this IQ test scientific?" has no answer. "Can I read the evidence that this test's scores mean what they claim, for people like me, under the conditions I will take it in?" does. When we searched for "scientific iq test" on October 29, 2026, the first page of results included press releases, app store listings and vendor pages that describe their own tests as scientific. None of that is evidence, because a description of a test by its seller is a claim, and the Standards separate claims from the rationale and the evidence behind them. Standard 1.2 asks for a rationale for each interpretation together with a summary of the evidence and theory behind it, and its comment adds that no type of evidence is inherently preferable to the others: the quality and relevance of the evidence for the intended use decide its value.

The International Test Commission (ITC) wrote the same expectation into its internet testing guidelines, and it put the burden on both sides. In the International Guidelines on Computer-Based and Internet-Delivered Testing, adopted in July 2005 and published in the International Journal of Testing in 2006, test publishers are asked to offer online only tests with appropriate psychometric evidence (guideline 20.3) and to advise users how to tell tests with and without documented psychometric properties apart (guideline 20.4). Test users are asked to be able to make that distinction themselves (guideline 21.3). The checklist on this page is that distinction, written out so that a layperson can apply it. The American Psychological Association says much the same to psychologists in its Guidelines for Psychological Assessment and Evaluation (2020): Guideline 6 asks them to select tools that show "sufficient validity evidence for their uses, sufficient score reliability, and sound psychometric properties", and it names standardization, reliability and validity among the psychometric characteristics that form the basis of interpretation.

How strict a reader should be depends on the stakes, and the sources agree on that too. Standard 2.0 says a higher degree of reliability and precision is required for uses with more significant consequences for the test taker, and a lower degree may be acceptable where a decision is reversible or depends on corroboration from other sources. The ITC guidelines distinguish high stakes testing, where a third party uses the result for an important decision, from low stakes testing, where the test taker obtains the information for his or her own interest. Someone curious about their own profile is in the second case. Someone who needs a score to document a learning difference, obtain an accommodation or join a society is in the first, and no online test is the right tool. The pages on what IQ actually measures and how IQ tests are made cover the construct and the construction process that the rest of this page takes for granted.

2 What Are the 15 Checks of a Scientific IQ Test?

Fifteen questions, each traceable to a numbered Standard, an ITC guideline or an APA guideline, separate a documented IQ test from one that only claims to be, and a reader can put all of them to any test publisher in about ten minutes. The table lists them in the order the rest of the page follows. The last column names the document and number behind each question, so a reader can check it. The Standards are the 2014 edition, the ITC document is the 2005 guidelines (published 2006) and the APA document is the 2020 guidelines.

Three notes on how the list was built. It is ours: no standard says to use these fifteen, so we grouped the requirements that bear on a person reading an IQ score and kept each item close to the wording of its source. It is about documented evidence and not about how clever the items are, so a test with excellent items can still fail checks 7 and 10. And the checks are not equally heavy: a test that fails checks 1, 7, 10 or 12 gives you a number with nothing under it, while one that fails only check 5 or 6 may still serve personal curiosity if its page says plainly that conditions were not controlled.

3 What Should a Test Say About Its Purpose, Population and Construct?

Before any number matters, a scientific IQ test states what its scores are for, which adults it was built for and which abilities each score claims to measure, and a page that cannot answer those three questions has not yet made a claim that anyone can check. These are checks 1 to 3, and they come first because every later check, from norms to error, is judged against them.

Check 1, intended use. Standard 1.1 asks the developer to set forth how scores are intended to be interpreted and used, and its comment asks for the population, the construct, the contexts in which scores are to be employed and the processes by which the test is administered and scored. Standard 4.1 asks that the test specifications carry a rationale supporting the interpretations and uses of the results, and Standard 7.1 asks for cautions against misuses that can reasonably be anticipated. In practice, look for a sentence of the form "this score is for X and not for Y". A page that sells a personal IQ estimate and, in the same breath, a certificate for employers or societies has made a second use claim, and Standard 1.5 says that when a recommended interpretation is said to lead to a specific outcome, the basis for expecting that outcome should be presented with evidence.

Check 2, population and language. The same standard asks that the population a test is intended for be delimited clearly, and Standard 7.2 adds that known limitations of a test for certain populations should be clearly delineated in the test documents. Language belongs here. Standard 3.0 asks that every step in the testing process minimize construct irrelevant variance, which is score variation that has nothing to do with the ability being measured. When a verbal task is given in a language the test taker learned later, the score blends ability with familiarity with that language, and that is a limit to state in advance. For tests used in more than one language, Standard 7.6 asks for documentation of the procedures used to translate and adapt the test, and the ITC guidelines ask developers of internationally used internet tests to assess differential item functioning (guideline 31.2) and to avoid language, drawings and content that are country or culture specific (guideline 31.4). ACIS states its own range on the page about taking an IQ test as an adult.

Check 3, construct. A scientific test says which abilities it measures and on what theory. Standard 1.2 asks for the theory as well as the data, and Standard 4.1 asks for a definition of the construct or domain measured that is detailed enough to show what is included and what is excluded. For IQ tests the common vocabulary is the Cattell-Horn-Carroll (CHC) model, which names broad abilities such as crystallized knowledge (Gc), fluid reasoning (Gf), quantitative knowledge (Gq), visual processing (Gv), working memory capacity (Gwm) and processing speed (Gs). The page on the CHC model explains the structure and the page on cognitive domains in IQ testing lists what each domain samples. A test that samples one narrow ability should say so, and a timed task blends speed into the score, since the result reflects how well and how fast a person worked (Standard 1.12 asks for evidence when a score's argument depends on assumptions about cognitive operations). A reader who wants speed measured as its own ability, with its own norms, can read about a normed processing speed test, and a reader choosing among general instruments can start from the page on adult normed cognitive ability tests.

4 How Should an IQ Test Be Given So That Scores Are Comparable?

A score is comparable only if the instructions, timing, equipment and conditions were the same in kind for the people in the norm group and for you, so a scientific test writes its procedure down and says what it assumes about your room. These are checks 4 to 6, and they are where an online test differs most from an examiner led one.

Check 4, standardized procedure. Standard 6.0 asks that assessment instruments have established procedures for administration, scoring, reporting and interpretation, and Standard 6.1 asks administrators to follow the standardized procedures carefully. Standard 7.8 says test documentation should include detailed instructions on how the test is administered and scored, covering qualifications of administrators, equipment, timing and accommodations, whether the test is on paper, on a computer or oral. APA Guideline 8 adds that standardized conditions support valid interpretation of norms and minimize confounds that lead to misinterpretation, and Standard 6.5 asks for instructions and practice that reduce construct irrelevant variance. For an online test the software plays the examiner, so the question becomes whether its instructions, practice items and timers are documented.

Check 5, conditions and supervision. The ITC guidelines give the vocabulary. They describe four modes of administration. In open mode there is no direct human supervision and no means of authenticating the test taker, and the guidelines give an internet test with no registration as an example. In controlled mode there is still no supervision but the test is made available only to known test takers. In supervised mode a person oversees the session, and in managed mode, such as a dedicated testing centre, a high level of human supervision and control applies. Guideline 37.1 asks test developers to document the level of supervision a test requires. Standard 3.4 raises two further issues for technology based testing: examinees must have access to technology so that the technology itself does not influence scores, since people working on older, slower equipment can be unfairly disadvantaged, and unproctored administrations, where standardization cannot be ensured, call for caution when results are interpreted. Standard 6.4 lists conditions to avoid: noise, disruption, extremes of temperature, poor lighting, inadequate work space and malfunctioning computers.

A reader can turn those sources into a short routine for an unsupervised session. These are our inferences from Standards 3.4 and 6.4 and ITC guideline 35.3 (breaks in lengthy tests), not rules from any one publisher. Choose a quiet room and a device you trust. Confirm that sound works before you begin, because some subtests present their items as audio: on ACIS, Digit Span, Alphanumeric Sequencing and Arithmetic need working sound, while Layer Rotation, in the technical manual's words, carries no words at all. Close anything that can interrupt you. Where a test allows breaks, rest between subtests and not in the middle of one, because ACIS saves progress between subtests.

Check 6, integrity and exposure. Standard 6.6 asks for reasonable efforts to prevent scores obtained by fraudulent or deceptive means, and its comment says that where results carry important consequences, score integrity should be supported by active efforts to prevent, detect and correct such scores. Standard 6.3 asks that changes or disruptions to standardized administration be documented. The ITC guidelines ask developers to contemplate delivery strategies that deter memorization of test content, such as unique tests drawn from item banks (guideline 40.5), and ask users not to coach test takers with actual test content (guideline 42.2). For an internet test taken in open or controlled mode in a moderate or high stakes setting, guideline 44.2 says checks against cheating can include a later validation assessment under proctored conditions, with scores compared. Standard 1.7 covers practice: if performance is claimed to be essentially unaffected by practice and coaching, the propensity of scores to change with practice should be documented. A publisher that says nothing about this has left check 6 unanswered, and the page on retaking an IQ test is the place to start for repeat testing.

5 What Makes a Norm Group Good Enough, and Do Norms Expire?

A norm group is good enough when the publisher describes exactly who was sampled, how and when, explains what each reported number means, and renews the norms as the population changes, because a score is only a comparison with that group. These are checks 7 to 9. The page on how IQ scores are normed explains the mechanism, and this section lists what a publisher must show.

Check 7, the norm group. Standard 5.9 says reports of norming studies should include a precise specification of the population that was sampled, the sampling procedures and participation rates, any weighting of the sample, the dates of testing and descriptive statistics, and that technical documentation should indicate the precision of the norms themselves. Standard 7.2 asks that, where normative data are provided, the procedures used to gather them be explained, the norming population be described in terms of relevant demographic variables and the years of data collection be reported. Standard 5.8 asks that norms refer to clearly described populations that include the individuals or groups with whom users will ordinarily wish to compare their own examinees. The APA guidelines tell psychologists to read the technical manual and review the standardization samples and their descriptive statistics. A lay reader can do the same: find the norm group section and see whether it answers who, how many, how recruited and when. Pearson's WAIS-5 sample score report shows how explicit a professional publisher is about this. It prints scaled scores against the examinee's own age group and a second column of scaled scores against a reference group that it defines as examinees aged 20 years 0 months to 34 years 11 months.

Check 8, the score scale. Standard 5.1 asks that users receive clear explanations of the characteristics, meaning and intended interpretation of scale scores and of their limitations, and Standard 5.2 asks that the procedures for constructing the reporting scales and their rationale be described clearly. ACIS reports indices and Full Scale IQ with a mean of 100 and a standard deviation of 15, and subtest scaled scores with a mean of 10 and a standard deviation of 3, which are the conventional scales for this kind of test. The page on why the standard deviation is 15 explains what a point on that scale means. A score that is described only as an IQ, with no stated scale, fails this check, because the same number means different things on different scales. Standard 5.3 adds that when specific misinterpretations of a scale are likely, users should be explicitly cautioned.

Check 9, currency. Standard 5.11 says that as long as a test remains in print, the publisher is responsible for renorming it often enough to keep score interpretations accurate. The APA guidelines list the agedness of the norms among the considerations in test selection. The reason is empirical. In a meta-analysis of 285 studies with 14,031 participants, Trahan and colleagues (2014) found that IQ scores rose by 2.31 standard score points per decade on average (95 percent confidence interval 1.99 to 2.64), and by 2.93 points per decade (2.3 to 3.5) in the modern Stanford-Binet and Wechsler comparisons they examined. That works out to about 4.6 points over two decades at the all studies average (our arithmetic), an average across tests and not a prediction for any one test. It is why a scientific test reports when its norms were collected and says how it keeps them current.

6 How Much Error Should a Scientific IQ Test Publish?

A scientific IQ test publishes an estimate of error for every score it asks you to read, expressed in the units of that score, and it treats the difference between two scores as a separate and noisier quantity than either score. These are checks 10 and 11.

Check 10, error for every score. Standard 2.0 asks for appropriate evidence of reliability and precision for each intended score use. Standard 2.3 asks that estimates be reported for each total score, subscore or combination of scores that is to be interpreted, and its comment says it is not sufficient to report reliabilities and standard errors only for total scores when subscores are also interpreted, because total scores may be highly consistent while subscores have unacceptably low reliability. Standard 2.13 asks that the standard error of measurement be given in units of each reported score, and the ITC guidelines ask developers to document the validity, reliability and fairness of the internet testing process (guideline 19.1).

Pearson's WAIS-5 sample score report, which we opened on October 29, 2026, shows what a professional publisher prints. Every composite carries a composite score, a percentile rank, a 95 percent confidence interval calculated with the standard error of estimation, and a standard error of measurement. For its invented examinee the Full Scale IQ is 111 with an interval of 106 to 116, the verbal comprehension index is 124 (115 to 130) and the processing speed index is 108 (99 to 116). The composite standard errors run from 3.00 to 5.61 index points, and the subtest standard errors run from 0.67 to 1.47 on the scaled score scale. In our arithmetic the Full Scale band spans 10 points, the verbal comprehension band 15 and the processing speed band 17. Every score has its own band, and a score without a band is incomplete.

A result has four parts. The subtest scaled score (1 to 19, mean 10) describes one task, an index (mean 100, standard deviation 15) describes a broad ability, a percentile says what share of the norm group scored lower, and the confidence interval says how far the obtained score could sit from the underlying value. The page on why IQ scores and percentiles cannot be averaged explains the difference between the second and third, and the IQ percentile calculator converts a score into its percentile. For a closer look at how a report is laid out, see a worked IQ test report example.

Check 11, differences between scores. Standard 2.4 asks that when an interpretation emphasizes differences between two observed scores of an individual, reliability and precision data, including standard errors, be provided for the differences. Its comment notes that the reliability of change scores can be much lower than the reliabilities of the separate scores. Standard 1.14 asks that when interpretation of subscores, score differences or profiles is suggested, the rationale and evidence be provided, and its comment says that evidence for two separate scores would not necessarily justify interpreting the difference between them. We can make the point with the printed numbers. The standard error of a difference between two independent scores is the square root of the sum of their squared standard errors. For verbal comprehension (4.50) and processing speed (5.61) in the sample report that is about 7.2 points, and 1.96 times that is about 14 points. This is our arithmetic, it ignores any correction a publisher applies, and a critical value from the test's own manual should replace it when one exists. A gap of 8 points between two indices is well inside it. The page on how common large gaps between index scores are shows why ordinary people have gaps, and the page on reliability and validity explains why a consistent score is not automatically a meaningful one.

7 What Counts as Validity Evidence for an IQ Score?

Validity evidence is the reason to believe that a score means what its publisher says it means for a particular use, and the Standards sort it into evidence from test content, response processes, internal structure and relations to other variables, none of which a high reliability can replace. These are checks 12 and 13.

Check 12, a rationale for each interpretation. Standard 1.0 requires evidence for each intended interpretation, and Standard 1.2 requires a rationale combining logical analysis with empirical evidence. Standard 1.4 puts the burden on the user who adopts a new interpretation: if a score is interpreted for a use that has not been validated, the user must justify it with a rationale and, if necessary, new evidence. Standard 1.5 covers outcome claims. A page that says its IQ score predicts career success, relationship outcomes or academic results needs evidence for that specific claim, and a page that offers only a definition of intelligence has not supplied it.

Check 13, structure and relations. The Standards describe the kinds of evidence a reader should expect to find in a manual. Evidence based on test content analyzes the relationship between what the items sample and the construct (Standard 1.11). Evidence based on response processes asks whether the cognitive operations people use match the intended ones (Standard 1.12). Evidence based on internal structure asks whether the relationships among items and parts of the test fit the construct, for example through factor analysis showing that one major dimension accounts for most of the shared variance when a test is claimed to be essentially unidimensional (Standard 1.13). Evidence based on relations to other variables asks whether scores relate to external measures in the way theory predicts (Standard 1.16), and Standards 1.17 and 1.18 ask that the criteria used be of known technical quality and that the levels of criterion performance associated with given scores be reported when a prediction is asserted. Standard 1.8 asks that the validation sample be described in as much detail as is practical, and its comment names self-selection, attrition, linguistic ability and exclusion criteria as factors that can limit representativeness.

Composites need their own case. Standard 1.14 asks that where composite scores are developed, the basis and rationale for arriving at them be given. A composite is a different cut of the same subtests, so the reader should ask for the rationale and the precision of that composite and not assume them from the Full Scale IQ. The word "validated" belongs under check 12 as well. Standard 1.1 says validity statements should refer to particular interpretations and uses, so "validated" without a population, a score and a use is a label, not a finding.

8 What Changes When a Scientific IQ Test Is Delivered Online?

Delivery over the internet does not change what the standards require, but it removes the examiner, so the evidence that matters shifts toward documented procedures, an honest statement of the administration mode and a clear limit on what the score may be used for. The ITC guidelines say that current psychometric standards apply even though the way tests are developed and delivered may differ (guidelines 19.2, 20.2 and 21.2). They also say what is lost. Guideline 29.1 asks publishers to tell users that, for open or controlled modes of internet testing, test takers may have been tested in non-standardized, unproctored or variable conditions while score interpretations are based on administration in proctored, standardized conditions, and that some modes make it impossible to guarantee the true identity of the test taker. Guideline 30.1 repeats the warning for users and adds that tests alone, however administered, may not provide a complete assessment of an individual.

The size of the cost is an empirical question, and the evidence says it is not zero and not uniform. In a meta-analysis of ability assessments, Steger, Schroeders and Gnambs (2020) combined 109 effect sizes from 49 studies with a total of 100,434 participants and found a pooled standardized mean difference of 0.20 (95 percent confidence interval 0.10 to 0.31), with higher scores in unproctored assessments. They found significantly smaller effects for measures that are difficult to search on the internet, and they concluded that unproctored ability assessments are biased by cheating. That is a finding about ability assessments in general, not about any one IQ battery, and it is an average. On the other side, Germine and colleagues (2012) compared web and laboratory samples on cognitive and perceptual experiments and concluded that collecting data from anonymous, unsupervised, self selected participants need not reduce data quality. Their setting was research, where participants have little to gain from cheating, which is our inference and is the opposite of a person who wants a high number.

Put together, the two findings say that unsupervised does not mean worthless, and that it does not mean equivalent either. A reader should look for three things on an online IQ page. The first is a statement of the mode, in the ITC sense: does the page say plainly that it is unsupervised? A test that can be taken with no registration at all is the ITC's own example of open mode, and the page on taking an IQ test with no sign up is the place to look if that is what you want. The second is a statement of use: the ITC guidelines say tests for post-sift selection or post-hire assessment normally would not be available in open mode (guideline 38.2), and a publisher that sells the same open mode score for hiring has contradicted its own mode. The third is a position on practice and repeat testing. Calamia, Markon and Tranel (2012) summarized nearly 1600 effect sizes from tests used in neuropsychological assessment and reported that practice effects, defined as score increases from memory for items, learned strategies or test sophistication, complicate the interpretation of change. Their evidence comes from neuropsychological tests, so it shows the mechanism and not the size for any IQ battery. The broader question of how trustworthy online tests are is covered on the page on whether online IQ tests are accurate.

9 Which Real Options Pass the Checklist, and What Do They Cost?

Two routes document enough to be judged against most of the 15 checks, a psychologist giving an individually administered battery and a normed online battery, while free quizzes and job screening practice tests have to be judged one product at a time. The table uses Pearson's WAIS-5 product page, opened on October 29, 2026, and the ACIS home page facts for the same date. Prices can change.

Read the table by the use you have in mind. If a professional must interpret the pattern or the result must be documented, the first row is the route, and the pages on every place to take an IQ test and what makes an IQ test professional cover the process. If you want a normed estimate for yourself, the ACIS rows are candidates, and a comparison of the best online IQ tests puts them beside others. For free tests, the page on free versus validated IQ tests explains what validation adds.

10 How Does ACIS Score on the Same 15 Checks?

ACIS documents its purpose, its limits, its score scales and its reports publicly and sends the evidence about norms, error and validity to a technical manual, so this page can say what ACIS states and what you must still verify, and it quotes no figure from the manual. The middle column uses only what ACIS publishes on its home page and in the technical manual's own descriptions. The right column is the work left for the reader.

Three rows deserve plain words. Row 5 is the largest limit: ACIS cannot control your room, and it allows breaks and up to 30 days, so conditions differ from person to person, which a supervised battery given in one sitting avoids. Row 6 follows from it. Row 9 is a gap on this page and not a claim about the test: we state no norm date because we do not quote the manual, so read it there. The technical manual is the document for rows 7 and 9 to 14, and it should be read by anyone who plans to rely on a score.

11 How Do You Run the Checklist in Ten Minutes, and What Are the Red Flags?

Score each check as found, partly found or not found using only the publisher's own pages, count anything you cannot find as not found, and judge the four heavy checks (1, 7, 10 and 12) first, because a number without them has nothing under it. The method is ours, built on the sources above, and it assumes a low stakes personal use. First, find the technical manual or the equivalent page. If there is none, checks 7 to 14 are not found, and Standard 7.0 says information relating to tests should be clearly documented so that users can make informed decisions. Second, for each score in the sample report, find its interval or error. Third, read the norm group paragraph and look for who, how many, how recruited and when. Fourth, read every use claim and ask what evidence stands behind it. Fifth, find the sentence that says which mode the test assumes.

A test that fails any heavy check is best treated as entertainment, however it is worded. A test that passes them and fails only checks 5 and 6 can serve as a personal estimate if its page says plainly that conditions were not controlled. Some wording should make a reader slow down. "Official" or "certified" with no body named is one: the page on what makes an IQ test official lists the routes that count, and the page on what an IQ test certificate is worth explains why a certificate proves little. A promise to place you in a rare tier raises Standard 5.9, which asks for the precision of the norms themselves, and the page on what a top 5 percent IQ means shows how that tier is defined. A single number from a handful of questions with no interval fails check 10. For a fuller list of signs that a score is genuine, see the page on telling an accurate IQ test from a weak one.

12 Where ACIS Sits, and What a Buyer Should Do

ACIS is a normed online battery for adult personal use, and its strengths and its limits are the same fact: it publishes its scales, its reports and a technical manual, and it is online, unsupervised, not clinical and in English only, so it is a reasonable way to see a profile for yourself and the wrong tool for documentation, hiring, school accommodations or a high IQ society. The facts below were read on the ACIS home page on October 29, 2026, and prices can change. It has 20 subtests in six domains, forms at 15, 30 and 50 dollars one time with no subscription, a free trial with no card, and example reports for each form. ACIS states: "Full refund within 5 days if the report does not deliver the features described at checkout or a technical issue prevents you from accessing your results." That is a conditional refund, not an unconditional one.

The Standards for Educational and Psychological Testing (AERA, APA and NCME, 2014) give the reading rules this page applied. Standard 1.1 says no test permits interpretations valid for all purposes or situations, Standards 2.3 and 2.4 ask for error on every score and difference that is interpreted, Standard 5.9 and Standard 7.2 ask for a described norm group, and the comment on Standard 3.4 says unproctored administrations, where standardization cannot be ensured, call for caution. Standard 12.10 adds that in educational settings a decision with major impact should consider more than a single test score. The APA's Guidelines for Psychological Assessment and Evaluation (2020) ask psychologists to choose tools with sufficient validity evidence, reliability and sound psychometric properties and to adhere to standardized conditions. If a school, employer or society asks for a score, it will name the test it accepts, and an online result does not replace that test.

The next step is short. Open the technical manual and read the norm group and reliability sections against checks 7 and 10. Open the example report for the form you are considering. Then start the free trial, which needs no card and is limited to one per person, or choose the form that covers the domains you care about. If a decision depends on the score, book an assessment with a psychologist instead.

13 Sources Behind This Page

Every third party fact and figure on this page comes from the documents below, opened or verified on October 29, 2026, and any figure we calculated ourselves is labeled in the text as our arithmetic. Prices and product pages change, so check them before you pay. The numbered standards and guidelines cited in the text were read in the full documents listed here.

1. American Educational Research Association, American Psychological Association, and National Council on Measurement in Education (2014). Standards for Educational and Psychological Testing. Washington, DC: AERA. Standards 1.0 to 1.18, 2.0 to 2.13, 3.0, 3.4, 4.1, 5.1 to 5.11, 6.0 to 6.10, 7.0 to 7.8 and 12.10 and their comments. Open access edition: https://www.testingstandards.net/open-access-files.html
2. International Test Commission (2006). International Guidelines on Computer-Based and Internet-Delivered Testing. International Journal of Testing, 6(2), 143-171. Guidelines adopted July 2005 and read in the ITC document, guidelines 19 to 21, 28 to 31, 34 to 44. https://doi.org/10.1207/s15327574ijt0602_4
3. American Psychological Association (2020). Guidelines for Psychological Assessment and Evaluation. Approved by the APA Council of Representatives, March 2020. Guidelines 6 and 8. https://www.apa.org/about/policy/guidelines-psychological-assessment-evaluation.pdf
4. Pearson (2024). WAIS-5 Wechsler Adult Intelligence Scale, Fifth Edition: Sample Score Report. NCS Pearson. Composite scores, percentile ranks, 95 percent confidence intervals and standard errors of measurement. https://www.pearsonassessments.com/content/dam/school/global/clinical/us/assets/wais-5/wais-5-sample-score-report.pdf
5. Pearson Assessments. Wechsler Adult Intelligence Scale, Fifth Edition (WAIS-5), product page: age range, qualification level, completion time and format. Read October 29, 2026. https://www.pearsonassessments.com/en-us/Store/Professional-Assessments/Cognition-%26-Neuro/Wechsler-Adult-Intelligence-Scale-%7C-Fifth-Edition/p/P100071002
6. Trahan, L., Stuebing, K., Fletcher, J., and Hiscock, M. (2014). The Flynn effect: A meta-analysis. Psychological Bulletin, 140(5), 1332-1360. https://doi.org/10.1037/a0037173
7. Steger, D., Schroeders, U., and Gnambs, T. (2020). A meta-analysis of test scores in proctored and unproctored ability assessments. European Journal of Psychological Assessment, 36(1), 174-184. https://doi.org/10.1027/1015-5759/a000494
8. Germine, L., Nakayama, K., Duchaine, B., Chabris, C., Chatterjee, G., and Wilmer, J. (2012). Is the Web as good as the lab? Comparable performance from Web and lab in cognitive/perceptual experiments. Psychonomic Bulletin and Review, 19(5), 847-857. https://doi.org/10.3758/s13423-012-0296-9
9. Calamia, M., Markon, K., and Tranel, D. (2012). Scoring higher the second time around: Meta-analyses of practice effects in neuropsychological assessment. The Clinical Neuropsychologist, 26(4), 543-570. https://doi.org/10.1080/13854046.2012.680913

14 Frequently Asked Questions