A long IQ test: what more minutes buy and where they stop

0 The short answer

A long IQ test is worth its minutes only up to the point where extra tasks stop adding precision; beyond that point what a longer battery buys is breadth across abilities, not a more exact number. Pearson lists the WAIS-5 at 45 minutes for the 7 subtests of its Full Scale IQ and 60 minutes for its 10 primary index subtests (read on November 4, 2026), so a single hour of examiner administered testing is already a long test, while ACIS Full Scale takes about 175 minutes for 20 subtests (59 dollars), with Quick (about 45 minutes, 25 dollars) and Optimized (about 110 minutes, 35 dollars) as shorter forms. ACIS is one option among several, and its prices can change. The decisive limit is that an unsupervised online session of any length gives a personal estimate, not a clinical, hiring, school or society result.

1 What Counts as a Long IQ Test?

Three different lengths hide behind the phrase "long IQ test", and they do not move together: the number of items, the number of subtests and the minutes on the clock. A test can run for an hour with one kind of task repeated, or for an hour with six different kinds of task, and the two hours say very different things about a person. People who search for a long, full or complete IQ test usually want one of two things: a number they can trust more than a five minute quiz gives, or a picture of more than one ability. Those are separate goals, and the research treats them separately.

Length by items governs how steady a single score is. Length by subtests governs how much of intelligence the battery samples. Length by minutes is the cost to you, and it depends on how many items, how many subtests and how much time each subtest allows. The page on times by instrument lists how long each major test takes; this page asks the prior question of how long a test needs to be.

Real numbers anchor the idea. Pearson, the publisher of the Wechsler Adult Intelligence Scale, Fifth Edition (WAIS-5), lists "45 minutes for the 7-subtest FSIQ; 60 minutes for the 10 primary index subtests" on its product page for the scale, read on November 4, 2026. The same page says that "administration time is determined by the composite scores desired" and that giving further subtests makes testing time somewhat longer. It lists ages 16 years 0 months to 90 years 11 months and paper or Q-interactive administration. A reader who wants the "hour long IQ test" is, in the professional setting, asking for roughly the core of what a psychologist gives.

The ACIS forms list their own times: Quick lists about 45 minutes for 6 subtests, Optimized about 110 minutes for 13 subtests and Full Scale about 175 minutes for all 20. The times of an examiner administered scale and an unsupervised online battery are not like for like, and this page does not claim to explain the gap. What matters for a buyer is what the minutes contain. The page on what a complete test means in coverage covers what complete means in practice, and the page on the Full Scale IQ test covers what that test is and how to take one.

A last distinction keeps the rest of the page honest. A long test is not the same as a hard test or an untimed test. Difficulty decides who can score high; length decides how precisely the score is estimated; timing rules decide what the clock is allowed to measure. We return to timing in the section on fatigue.

2 How Do the Real Options Compare?

The options split into a short list: a psychologist giving a full Wechsler scale, free online screeners, job-screening practice tests, and normed online batteries, and they differ more in who the score is compared with than in how long they run. The table sets them side by side. ACIS sells one of them, so read the ACIS rows as a disclosure and check them against the rest of the page.

Three notes make the table safe to use. First, the Pearson times are the publisher's own estimates for individual administration, and the Swinburne University clinic figure comes from the clinic's low cost assessment page, which lists the WAIS for ages 16 to 89 and does not state a duration. Second, the ACIS prices are those published when this page was written, and they can change. Third, the free and job-screening rows are descriptions of a category, not measurements of any single product, because the point of the row is the question to ask: compared with whom?

Where to find each option is a different job. The page on where to take an IQ test maps the in person and online routes, and the ranking of online IQ tests compares the online batteries. For the free end, ACIS's free trial page is the place to see what the trial offers: a free trial with no card, one per person. Whichever row you choose, the length of the session is only a proxy for what you actually need, which is a score from enough comparable tasks, compared with a described group of adults.

3 Why Does a Longer IQ Test Give a More Stable Score?

A longer test gives a steadier score because random error in single answers partly cancels when more comparable answers are averaged, and that is the logic of the Spearman-Brown relation. Any one answer on any test carries luck: a guess, a slip, a lapse of attention, a word that happened to be familiar. Ask more comparable questions and these small accidents average out, while the part that reflects the ability stays. That is why a three item quiz can swing wildly between two attempts on the same day.

The relation takes its name from two papers published together in 1910. Charles Spearman's "Correlation calculated from faulty data" and William Brown's "Some experimental results in the correlation of mental abilities" appeared in volume 3, issue 3 of the British Journal of Psychology, on pages 271 to 295 and 296 to 322. The formula the textbooks attribute to them predicts the reliability of a test lengthened by a factor k from its current reliability r: the new reliability is k times r, divided by one plus (k minus 1) times r. We give the formula as the standard textbook statement and label every number built on it as our arithmetic.

Reliability is the share of score differences between people that reflects stable differences rather than error. A test with a reliability of .90 is one where about 90 percent of the variation in scores is stable variation. It is tied to a number that matters more in practice: the standard error of measurement, the typical size of the error in a single score. On the IQ scale with a standard deviation of 15, the error is 15 times the square root of one minus the reliability, and a 95 percent confidence interval is about 1.96 times that error on either side of the score. The page on reliability and validity develops both ideas, and the page on why the standard deviation of 15 matters explains the scale.

The testing profession writes the same principle into its core document. The Standards for Educational and Psychological Testing (AERA, APA and NCME, 2014), open access at testingstandards.net, say in the chapter on test design that specifications for test length "must balance testing time requirements with the precision of the resulting scores", with "longer tests generally leading to more precise scores". Read that sentence twice. It makes two claims: longer is more precise, and length is a trade against time. The word "balance" signals that time is a cost to be weighed against precision.

The same document says what the relation assumes. Psychometric models can estimate the precision of a shorter or longer version of a test, but they "generally make assumptions that may not be met", for example that the items added or dropped are all sampled at random from one domain. Real tests add different tasks, not clones, so the formula is a guide to the shape of the gain, not a forecast for any particular battery. The next section uses it for exactly that purpose.

4 Where Do the Extra Minutes Stop Paying?

The gain from length is steep at first and flat later: doubling a short test helps a great deal, while doubling a test that is already long adds little. The table below is an illustration, not a measurement of any test. It starts from an imagined test with a reliability of .70 and applies the Spearman-Brown relation, then converts to the error on the IQ scale. Every figure is our arithmetic on a textbook relation.

Read down the third column. The first doubling removes about two points of standard error. The doubling from 8 times to 16 times removes about one, at eight times the added testing time of the first doubling. If the original test took 10 minutes, the first doubling costs 10 more minutes and the last costs 80. The score is steadier at each step, and each step pays less than the one before.

Two lessons follow, and both are limits on the table. The first is that the table assumes the added items are as good as the original ones and measure the same thing. That is the assumption the Standards flag. In practice a test developer can run out of strong items, and the extra tasks can be weaker. The second is that error in a score is not only item error. A person who sleeps badly, is distracted, or misreads instructions shifts every score in the session in the same direction, and adding items from that same session does not average that shift away.

A publisher's own behavior fits the picture. Pearson's product page says of the WAIS-5 that "the FSIQ, now with only seven subtests, can be obtained about 20 minutes faster than the WAIS-IV." The page lists the saving; it does not say what the publisher weighed to get there, and we do not infer more than that. What it shows is that a major battery can shorten the route to its headline score while the 10 primary subtests remain for the index scores.

For a buyer, the practical reading is simple. If you want one overall number, a battery of a few well chosen subtests from several domains is on the flat part of the curve for many purposes, and adding hours of the same material is on the flatter part still. If you want separate index scores for verbal, reasoning, memory and speed, you need enough subtests in each domain, and that is a reason for the longer forms, not for repeating items. The page on index score differences covers how often index scores differ within one person, and the section on breadth below shows why it differs from precision.

5 What Do Short Forms Keep, and What Do They Lose?

A short form keeps most of the ranking of people on overall ability when it is built and tested well, and loses precision for the individual, the profile and anything the dropped subtests covered. The evidence comes from a methodological review and from a worked example.

The review is by Gregory Smith, Denis McCarthy and Kristen Anderson, On the sins of short-form development, in Psychological Assessment in 2000. Their abstract describes a literature "characterized by overly optimistic views of the transfer of validity from parent form to short form". The authors do not oppose short forms. They argue that the validity standards for short forms "should be quite high", identify 2 general and 9 specific methodological sins, and recommend 6 a priori steps and 9 procedures for building valid abbreviated forms. The lesson for a reader is that a short test is only as good as the evidence that was gathered about it, and that the shortcut deserves its own evidence.

The Standards say the same in two places. Under reliability they note that "the short version of a standardized test often comprises a nonrandom sample of items" from the full version, so a prediction from the long form to the short one "may not provide a very good estimate" of the real value, and both should be evaluated directly. Under score scales, Standard 5.19 asks for evidence that tests created by taking a subset of the items in an existing test show no distortions of scale scores, cut scores or norms. Those are requirements for the developer, and a buyer can ask whether the developer met them.

A worked example gives the size of the cost. Christiany Suwartono and colleagues, in the Journal of Intelligence in 2023, tested 13 short forms of the Indonesian version of the WAIS-IV, from 2 to 7 subtests, in a standardization data set of 1,745 people. The correlation between the short form estimate and the full Full Scale IQ ranged from .77 for the weakest form to .97 for the strongest, and only nine of the thirteen exceeded .90. The four subtest form the authors recommended, built around Information, Matrix Reasoning, Arithmetic and Coding, correlated .95 with the full score and took about 25 minutes in their small validation study. Even so, they wrote that "differences of five IQ points or more" between the estimate and the full score "are not rare", and that short forms are best used for a quick indication, not for index scores or profile analysis.

That is the pattern in one study: a well built four subtest form tracks the full score closely on average and still misses by five points or more not rarely, and its authors advise against using it for index scores or a profile. It is one sample, with one translation of one scale, so we use it to show the shape of the trade, not to set a rule.

This is the honest place to describe the shorter ACIS forms. The page on what a short battery costs you is built around this trade. Quick covers 6 subtests in 3 domains in about 45 minutes and reports a Full Scale IQ with its 95 percent confidence interval. If you only want a rough position, that is a reasonable fit; if you want to know whether your processing speed differs from your reasoning, a battery without a processing speed subtest cannot tell you.

6 What Does a Full Battery Add That More Items Cannot?

A full battery adds breadth, meaning the score reflects several distinct abilities and not one task type repeated, and the research on general intelligence shows why breadth matters more than count. The Cattell-Horn-Carroll (CHC) model, the standard map of cognitive abilities, groups tasks into broad abilities such as fluid reasoning (Gf), crystallized knowledge (Gc), quantitative reasoning (Gq), visual processing (Gv), working memory (Gwm) and processing speed (Gs). The page on the CHC model describes the map, and the page on what IQ measures describes how the overall score relates to it.

Here is the key evidence on breadth. Wendy Johnson and colleagues gave three mental ability batteries to 436 adults and extracted a general factor from each. In Just one g: consistent results from three test batteries, published in Intelligence in 2004, they report that the general factors "were completely correlated (correlations were .99, .99, and 1.00)", so that the general factor does not depend on the specific tasks. A follow up with five batteries and 500 Dutch seamen, Still just 1 g (Intelligence, 2008), reports that all general factor correlations "were at least .95", with one exception: the Cattell Culture Fair Test, which "consists of just four matrix reasoning tasks", and for which the lowest correlation was .77. Read together, broad batteries converge on the same general factor, and a narrow battery of one task type tracks it less well.

The most direct evidence on how many subtests are needed is recent. Martin Breit, Andrew Conway and Kristof Kovacs, in 99 Ways to Obtain a General Factor of Intelligence (Journal of Psychoeducational Assessment, 2025), computed the general factor from every possible combination of subtests of the Woodcock-Johnson Tests of Cognitive Abilities V. They report that correlations among the general factors increased "with the number of subtests or broad abilities included", that most closely matched the estimate from all subtests (r above .9), that low correlations were "mainly tied to the overrepresentation of processing speed (Gs) in small test sets", and that reliable estimates "require broad, balanced batteries covering at least three broad abilities". That study says nothing about ACIS, and we do not read any ACIS property into it. We use it for its reasoning: for an overall score, coverage across abilities counts for more than the volume of one kind of task.

Breadth does a second job that precision cannot: it makes the profile readable. The ACIS Full Scale form covers six domains with 20 subtests. Verbal comprehension includes Vocabulary, a crystallized knowledge task built on word definitions; fluid reasoning includes Matrix Reasoning, which asks for rule discovery in unfamiliar visual patterns; working memory includes Digit Span, which uses audio; processing speed includes Symbol Search, a quick visual comparison task. Each subtest page names the CHC domain it belongs to. The same Full Scale IQ can come from a high Gf and a low Gs, or the reverse, and only a battery that includes both can show it.

Breadth also supports composite scores built from selected subtests. The technical manual lists composites such as CPI, which "combines working memory and processing speed as cognitive proficiency", and VKI, which "combines Antonyms, Synonyms, Vocabulary, and Information". The page on the g factor test takes the general factor itself further. A longer battery makes these composites possible because their subtests are available; a short one cannot offer them.

7 Is an Hour Long Online Test Long Enough?

An hour long online test can be long enough in minutes and still not be a usable IQ test, because the minutes say nothing about who the score is compared with. This is the question behind many searches for a long IQ test free of charge, a long form test or a complete test online, and the answer has two halves that must both hold.

The first half is norms. The Standards set the rule in Standard 5.8: "Norms, if used, should refer to clearly described populations." A score from sixty minutes of tasks is a raw count until it is placed against a reference group, and a reference group you cannot see is a reference group you cannot evaluate. A site that compares you with its own visitors, or states no comparison at all, can be perfectly entertaining and still not tell you where you stand among adults. The page on how IQ scores are normed explains how norms work, and the page on the best IQ test on Reddit covers how people there compare tests. The question to ask of any test is the one the Standards ask: compared with whom, and is that group described?

The second half is coverage. An hour of one task type can be precise and narrow, in the way the Breit and colleagues result and the matrix-only battery in Johnson and colleagues' second study suggest. An hour built from tasks in three or more domains, with a described norm, is a different product. So the useful test of "is it long enough" has three parts, in this order: is the comparison group described, are at least three broad abilities covered, and is the score reported with its error.

How long a test needs to be is therefore partly answered in minutes. The WAIS-5 core in 45 to 60 minutes (Pearson, read November 4, 2026) shows that a professionally given headline score rests on about an hour of well chosen subtests. Our reading is that an hour of one repeated item type is not comparable to it, and a three hour battery is not automatically better than it. The ACIS Full Scale form takes about 175 minutes because it contains 20 subtests in six domains, and in our reading the length serves the profile and the six indices, not a claim that 175 minutes makes the Full Scale IQ more exact than 60 minutes of well chosen tasks would. That is our reading, not a measured finding.

For a reader choosing between an hour long free screener and a longer normed battery, the page on which IQ test to take walks through the decision by goal, and the page on IQ tests for adults covers testing for adults. The trial is the low risk way to see an ACIS subtest before paying.

8 Does Fatigue Ruin a Long IQ Test?

Fatigue is real in long sessions, but the one controlled experiment on test length that we found reported that scores did not fall as the session got longer, so fatigue is a reason to plan breaks and not a reason to avoid a long test. The evidence is mixed and comes from two settings that are not IQ batteries.

The direct experiment is by Phillip Ackerman and Ruth Kanfer, Test length and cognitive fatigue, in the Journal of Experimental Psychology: Applied in 2009. Two hundred thirty-nine first-year university students completed three sessions of different length in a within-participant design: a standard SAT battery of 4.5 hours, a shorter one of 3.5 hours and a longer one of 5.5 hours. Subjective fatigue rose with time on task. But the abstract reports that "mean performance increased in the longer test length conditions, compared with the shorter test length condition". Individual differences in personality, interest and motivation predicted fatigue better than test length did. The authors do not say fatigue is harmless; they report that in this design more time on task did not lower the mean score. The measure is the SAT, not an IQ battery, and the sample is students.

The other evidence pulls the other way. Hans Henrik Sievertsen, Francesca Gino and Marco Piovesan, in the Proceedings of the National Academy of Sciences in 2016, used test data for all children in Danish public schools from 2009/10 to 2012/13. They found that for every hour later in the day, test performance decreased by 0.9 percent of a standard deviation, while a 20 to 30 minute break improved average performance by 1.7 percent of a standard deviation. Those are schoolchildren taking standardized school tests, and the effects are small. They are evidence that breaks and time of day matter at the margin.

The Standards add a caution from the other side. In the comment to Standard 5.19, they say that when items from an initial version are used to build conversion tables or approximate norms for another version, "caution is required in cases where context effects are likely, including speeded tests, long tests where fatigue may be a factor". That is a statement about test development and norming, and it tells us the profession treats fatigue as a design variable: it can change what a score means, and developers are expected to account for it.

Putting these together without overreach: fatigue is an experience of long tests, it does not reliably lower scores in the one controlled experiment on test length that we found, and small time-of-day and break effects appear in large school data. For a three hour IQ battery taken at home, the safe reading is practical. Choose a time when you are usually alert, take the breaks the format allows, and do not read a low score on a late subtest as proof of low ability when you were plainly worn out. We know of no study that measured fatigue inside an ACIS session, and we do not report one.

Fatigue is also a different matter from the clock. Whether each subtest has a time limit, and what the limit measures, is the subject of the page on why IQ tests are timed and the page on what the clock changes in an untimed test. A long test can be timed on every subtest or have no limits at all, and the two properties are independent.

9 How Long Is Long Enough for Each Purpose?

The length you need depends on what you will use the score for, and for most personal purposes a few well chosen subtests from several domains are enough, while a profile asks for more and an official result asks for a psychologist. The table matches purposes to lengths using only the logic above.

The curiosity row is where many readers sit. The shorter ACIS forms exist for it: a Quick or Optimized form gives a Full Scale IQ with its confidence interval without three hours. The profile row is where the Full Scale form earns its length, because it samples every domain. For the gifted question, the page on gifted tests for adults covers testing for giftedness as an adult, and the page on high range IQ tests covers how scores near the top of the scale are measured. For a single ability, the page on processing speed tests shows a one domain route.

The last row matters most. A longer unsupervised session is still unsupervised. The APA Ethics Code, in Standard 9.02(a), asks psychologists to use assessments for purposes that are appropriate in light of the evidence on their usefulness and proper application, and in 9.02(b) to use instruments whose "validity and reliability have been established for use with members of the population tested". Length is not on that list. A clinical or legal purpose needs a qualified examiner, whatever the test.

One more distinction avoids a common mistake. A long test is not automatically a better route to a very high or very low score. If a result near either end of the scale matters to you, read it with its interval and treat it as an estimate.

10 How Should You Sit a Long IQ Test at Home?

Sit a long test as you would a long exam: pick an alert time, remove interruptions, check the sound before you start, and use the breaks the format provides. The advice is practical and it follows from the evidence on fatigue and from what ACIS says about its own format.

Choose the time first. The Danish school data showed performance dropping across the school day and recovering after a break, so a morning or other usual high point is a reasonable default, and you can spread the Full Scale form over several sittings. ACIS allows breaks, saves progress between subtests and gives 30 days to complete a form, so there is no reason to push through a bad hour. The natural place to stop is between subtests.

Check the device and the sound. Digit Span, Alphanumeric Sequencing and Arithmetic present their items as audio and need working sound. Headphones in a quiet room reduce the chance that a noise changes a memory span result. The Digit Span page describes the task as measuring immediate auditory-verbal storage, so noise in the room is a real hazard for it. Layer Rotation, by the technical manual's own description, "carries no words at all".

Keep the conditions consistent across sittings. Use the same device where you can, the same room and a similar time of day, so that a difference between two subtests is less likely to be a difference in conditions. Be honest with yourself about alertness: a subtest taken while ill, upset or very tired is a weaker estimate of that ability, and the right response is to say so and, if needed, take it again under better conditions rather than to treat the result as a verdict.

Do not use outside help. A score is only comparable with the norm group if you took the test as its instructions say, so follow each subtest's instructions and do not ask a search engine or another person. English is the only language, and the manual says verbal and knowledge tasks are affected by English proficiency and educational exposure, so a reader whose first language is not English should expect the verbal subtests to reflect that as well as ability.

Finally, plan what the free trial can do for you. It opens a small set of subtests with no card, and finished trial subtests are kept if you later unlock a form that includes them. Trying a few before you commit shows you the format and how the session feels, before you book the rest.

11 How Do You Read the Score From a Long Test?

A long test narrows the error in a score but does not remove it, so read every score as a range, read subtest scores as less precise than composites, and read percentiles as positions, not points to average. The reading rules are the same for any normed test.

Start with the scales. Indices and the Full Scale IQ sit on the standard scale with a mean of 100 and a standard deviation of 15. Scaled subtest scores run from 1 to 19 with a mean of 10 and a standard deviation of 3. Percentiles say what share of the norm group scored lower. The page on IQ score versus percentile explains why you cannot average percentiles, and any IQ score chart lays out the levels and ranges. Rarity is a related but different question.

Then read the interval. Every ACIS form reports a Full Scale IQ with its 95 percent confidence interval. The interval is the honest statement of the score: the true value is likely inside the band, and the band is wider than most people expect. The arithmetic table earlier shows how the width shrinks with length and how slowly, so a longer test gives a narrower band but never a point.

Next, compare like with like. A single subtest is usually less precise than an index, because it rests on fewer items; an index built from several subtests is steadier; the Full Scale IQ, which draws on the most subtests, is steadier still. The Standards make the matching point in Standard 2.3: estimates of reliability should be reported for each total score, subscore or combination of scores that is to be interpreted, and a total score can be reliable while its subscores are not. So a single low scaled score is a prompt to look at the index it belongs to, and a gap between two indices is a prompt to consider whether the gap exceeds the measurement error. The page on the Full Scale IQ score explains what that composite is.

Finally, keep group statistics where they belong. A norm table describes a group, and a score places you in it; neither is a diagnosis. If your score will drive a decision about a job, school, accommodation or clinical question, take it to a psychologist and do not rely on a self administered session.

For the documentation behind ACIS scores, the technical manual is the place to read, and the report itself shows the interval, the indices and the subtest scores. We do not reproduce the manual's figures here, because the manual is the proper home for them.

12 Where ACIS Sits, and What a Buyer Should Do

ACIS is a normed online battery for adults with forms of three lengths, it is not a clinical instrument, and the right length is the shortest one that covers the abilities you care about. We sell it, so read this section as a disclosure and check it against the rest of the page.

The limits come first. ACIS is online and unsupervised, in English only, and "is not intended for clinical diagnosis, hiring decisions, accommodation determinations or admission to high IQ societies", in the words of its research page. The technical manual says verbal and knowledge tasks are affected by English proficiency and educational exposure. Adult norms cover ages 16 to 90. A session of 175 minutes at home is still a session without an examiner, and no length changes that.

The strengths are real and bounded. Quick (6 subtests in 3 domains, about 45 minutes, 25 dollars), Optimized (13 subtests in 5 domains, about 110 minutes, 35 dollars) and Full Scale (all 20 subtests in six domains, about 175 minutes, 59 dollars) are one time payments, with no subscription. Every form reports a Full Scale IQ with its 95 percent confidence interval. Breaks are allowed, progress is saved between subtests, and you have 30 days to finish. The Custom Test lets you choose subtests at 10 dollars each, which costs more than the Full Scale form from 6 subtests. The refund policy reads: "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." Prices were read on the live site on November 4, 2026 and can change.

The standards for a fair reading are written down. The Standards for Educational and Psychological Testing (AERA, APA and NCME, 2014) ask that norms refer to clearly described populations (Standard 5.8), that reliability be reported for each score that is interpreted (Standard 2.3), and that shortened versions show evidence of no distortion of scale scores or norms (Standard 5.19). They also call for caution where fatigue may be a factor in long tests. The American Psychological Association's Ethics Code, Standard 9.06, tells psychologists interpreting assessment results to take into account "the purpose of the assessment" and the test taker's characteristics, and to "indicate any significant limitations of their interpretations". Those are rules for professionals, and a buyer can apply the same questions to any test: who is the score compared with, how precise is it, what are its limits?

What to do next depends on your goal. If you want a rough overall position, start the free trial with no card and, if you like the format, take Quick. If you want to compare abilities, take the Full Scale form and plan two or three sittings. If you need a result for a school, an employer, a clinic or a society, book a psychologist or the society's own supervised exam and treat ACIS as a way to see an adult normed picture first. Whichever you choose, read the score with its interval, and let the purpose, not the minutes, decide the length.

13 Sources Behind This Page

Every figure above traces to one of the following, and each is linked where it is used; numbers built on the Spearman-Brown relation are our arithmetic and are labeled as such.

• Spearman C. Correlation calculated from faulty data. British Journal of Psychology, 1910, volume 3, issue 3, pages 271 to 295.
• Brown W. Some experimental results in the correlation of mental abilities. British Journal of Psychology, 1910, volume 3, issue 3, pages 296 to 322.
• Smith GT, McCarthy DM, Anderson KG. On the sins of short-form development. Psychological Assessment, 2000, volume 12, issue 1, pages 102 to 111.
• Suwartono C, Hendriks MPH, Hidajat LL, Halim MS, Kessels RPC. The Development of a Short Form of the Indonesian Version of the Wechsler Adult Intelligence Scale, Fourth Edition. Journal of Intelligence, 2023, volume 11, issue 8, article 154.
• Johnson W, Bouchard TJ, Krueger RF, McGue M, Gottesman II. Just one g: consistent results from three test batteries. Intelligence, 2004, volume 32, issue 1, pages 95 to 107.
• Johnson W, te Nijenhuis J, Bouchard TJ. Still just 1 g: Consistent results from five test batteries. Intelligence, 2008, volume 36, issue 1, pages 81 to 95.
• Breit M, Conway A, Kovacs K. 99 Ways to Obtain a General Factor of Intelligence, and They Are Not All Created Equal. Journal of Psychoeducational Assessment, 2025, volume 43, issue 8, pages 918 to 932.
• Ackerman PL, Kanfer R. Test length and cognitive fatigue: An empirical examination of effects on performance and test-taker reactions. Journal of Experimental Psychology: Applied, 2009, volume 15, issue 2, pages 163 to 181.
• Sievertsen HH, Gino F, Piovesan M. Cognitive fatigue influences students' performance on standardized tests. Proceedings of the National Academy of Sciences, 2016, volume 113, issue 10, pages 2621 to 2624.
• American Educational Research Association, American Psychological Association and National Council on Measurement in Education. Standards for Educational and Psychological Testing. Washington, DC: AERA, 2014.
• American Psychological Association. Ethical Principles of Psychologists and Code of Conduct, Standards 9.02 and 9.06. Effective January 1, 2017.
• Pearson. Wechsler Adult Intelligence Scale, Fifth Edition, product page. Administration time, ages and format, read on November 4, 2026.
• Swinburne University of Technology. Swinburne Psychology Clinic, low cost psychological assessments. WAIS fee and age range, read on November 4, 2026.

14 Frequently Asked Questions