Home › Personality Types › What an IQ Test Measures and What It Includes › Cognitive Ability Tests Explained

Cognitive Ability Tests: What They Are and What They Predict

A cognitive ability test is the short, timed reasoning screen employers use early in hiring. This page explains what is in one, what the evidence now says about how well it predicts job performance, and what preparation is genuinely worth your time.

6 sources cited10 min read
Cognitive Ability Tests Explained: a job applicant at a desk watches a wall clock, hand on a face-down sheet.

In short

A cognitive ability test is a short, usually timed reasoning assessment used to screen job applicants. It typically mixes verbal, numerical and abstract or logical items, is often adaptive, and takes somewhere between ten and forty minutes, and it is not a clinical IQ assessment and not a personality questionnaire: there are right answers, nobody is watching you one to one, and in most cases no score report or interpretation follows. The evidence on how well these tests predict job performance was revised sharply downward in 2022, when a re-analysis found earlier estimates had been systematically overcorrected, and while familiarity with the format helps a little, nothing substitutes for being suited to the job. Psychology.com's reasoning quizzes are practice for the format, not predictors of a hiring outcome.

What it is
A short timed reasoning screen used in hiring
Usual content
Verbal, numerical and abstract or logical items
Usual length
Roughly 10 to 40 minutes, often adaptive
Predictive validity
Revised down to about .31 from an earlier .51
What helps
Format familiarity, timing practice, sleep
Test
IQ Test: Free Reasoning Quiz

What a cognitive ability test is

A cognitive ability test is a standardized reasoning assessment used to compare job applicants. It has right and wrong answers, it is usually timed, and it is almost always taken online and unsupervised at the application stage, sometimes with a supervised retest later for candidates who progress. Employers use it as a screen because it is cheap, fast and applies the same yardstick to everybody.

The term covers a wide range of products. Some are general reasoning tests that mix question types; some are single-domain tests of numerical or verbal reasoning; some are adaptive, choosing each next item based on how you answered the last one, which lets them finish in fifteen minutes with the same precision a longer fixed test would need. Vendors give them different names, including aptitude test, reasoning test, general mental ability test and ability screen, and the differences between those names are mostly marketing.

What they share is the logic of the exercise. You are being placed in a rank order against other applicants on how quickly and accurately you handle unfamiliar material. Nothing about the result is fed back to you in most cases, and nothing about it is intended to tell you anything about yourself.

How it differs from an IQ assessment

A clinical intelligence battery is a different kind of object. It is given one to one by a qualified clinician over one to two hours, covers five domains including spoken and visual material, uses routing and stopping rules, and produces a written interpretation with percentiles and confidence intervals. Its purpose is to answer a question about one person, usually a referral question about learning, development or a change in functioning.

A hiring test shares the reasoning content and almost nothing else. It samples a narrow slice, mostly fluid reasoning plus acquired numerical and verbal skill. At the application stage it is almost always unsupervised. It produces a number or a band used to rank candidates, not an interpretation. And it is not a clinical instrument: it is built and sold to rank applicants, not to assess one person.

It is also not a personality questionnaire, which is a distinction worth holding onto if you are sitting several assessments in one application. A personality questionnaire has no right answers and asks how you typically are; our own work styles page covers what those look like. A cognitive test has right answers and asks what you can work out. Approaching one as if it were the other is a common and costly mistake.

What you will actually sit

Verbal reasoning items give you a short passage and ask what follows from it. The trap is that most people answer from general knowledge rather than from the passage, and the correct answer is almost always the one supported only by the stated text, even when the text describes something implausible.

Numerical reasoning items give you a table or chart and ask for a calculation: a percentage change, a ratio, a projection. The arithmetic is rarely hard. The difficulty is reading the right figure out of a cluttered table under time pressure, and knowing when a question can be answered by estimating rather than computing.

Abstract or logical reasoning items give you shapes or sequences and ask you to infer the rule. A matrix with a missing cell, a series that continues by a pattern you have to find, a set of figures with one odd member. This is the purest fluid reasoning section and the one most sensitive to practice, because the families of rules used are limited and recognizing them quickly is a learnable skill.

Timing is the design feature people underestimate. Many tests are built so almost nobody finishes, so the score reflects speed as much as accuracy and the right strategy is to move rather than to perfect. Check before you start whether wrong answers are penalized; most are not, in which case a blank is strictly worse than a guess.

What employers think they are measuring

The rationale employers were given for decades came from a 1998 review by Frank Schmidt and John Hunter, which put the corrected validity of general mental ability tests for predicting job performance at around .51, higher than almost any other single method. That figure is the reason cognitive screens spread through hiring, and it is quoted in vendor materials to this day.

In 2022 Paul Sackett, Charlene Zhang, Christopher Berry and Filip Lievens published a re-analysis in the Journal of Applied Psychology showing that the corrections applied in those earlier meta-analyses had systematically overcorrected for range restriction. Their revised estimates were lower by roughly .10 to .20 across the board. Cognitive ability came down to about .31. Structured interviews, at about .42, moved to the top of the table.

The same group then went further. Their 2024 analysis of 153 samples run between 2000 and 2021, covering 40,740 people, found a mean corrected validity of .22 for this century's data against an observed correlation of .16. That estimate sits well below the .42 the same group reported for structured interviews, which leaves cognitive ability a long way down their revised ranking of predictors. The authors were careful to say selection methods remain useful rather than useless: the point is that the relationships are considerably weaker than the field believed. Translated into the applicant's situation, a reasoning score is a weak signal that employers are treating as a strong one, so a bad result means you have failed a filter with modest predictive power rather than been measured and found wanting.

The adverse impact problem

Cognitive ability tests produce average score differences between demographic groups that are large enough to screen out protected groups at noticeably different rates. In United States employment law this is the adverse impact problem, and it is the central practical difficulty with these tests regardless of what anyone believes causes the differences.

The legal structure is specific. Under the Uniform Guidelines on Employee Selection Procedures, at 29 CFR 1607.4(D), a selection rate for any race, sex or ethnic group that is less than four fifths of the rate for the group with the highest selection rate is generally regarded by the federal enforcement agencies as evidence of adverse impact. Smaller differences can still count where they are statistically and practically significant. A test with adverse impact must then be shown to be job related and consistent with business necessity, and even then a challenger can point to a less discriminatory alternative that predicts performance equally well.

What employers do about it is a research field of its own, sometimes called the diversity-validity dilemma. Reviewed strategies include using the test as one input in a composite rather than as a cutoff, adding structured interviews and work samples that predict at least as well with smaller subgroup differences, removing items that depend on cultural knowledge rather than reasoning, and dropping unnecessary time pressure. None of these fully resolves the tension, which is one reason the 2022 revision mattered so much: if a cognitive screen is a weaker predictor than everyone assumed, the case for accepting its adverse impact gets correspondingly weaker.

What preparation actually helps

Practice produces a real but modest gain. A meta-analysis by John Hausknecht and colleagues of retesting on cognitive ability tests found an adjusted overall effect of about .26 of a standard deviation, with larger gains when the identical form was reused and when coaching preceded the second attempt. A quarter of a standard deviation is a small move inside a candidate ranking: worth having if you are near a cutoff, and not transformative.

The gain comes mainly from removing avoidable losses. Knowing the instruction format, knowing whether wrong answers are penalized, knowing that verbal items must be answered from the passage alone, having used an on-screen calculator before, and having experienced the pace all save time you would otherwise spend learning the rules while the clock runs.

Everything else on the list is unglamorous and genuinely effective. Sleep, because working memory and processing speed are the first things to go without it. A quiet room, a stable connection and a device you are used to. Starting when you are not hungry. Doing it at the time of day you think best, if the window allows. These sound trivial next to practice tests and they protect more points.

What does not help

Memorizing answers does not help, because the item banks are large and adaptive tests select different items for different candidates. Paid courses promising large score increases are selling against an effect size of roughly a quarter of a standard deviation, which is a claim the evidence does not support.

Cramming the subject matter is mostly wasted too. Numerical items use arithmetic almost everyone already has, and the difficulty is extraction and timing rather than mathematics. Repeating one practice test until your score rises is worse than wasted, because the improvement is familiarity with those specific items. If you practice, use different sources and treat the score as a measure of your pacing rather than of yourself.

Accommodations, and what to ask for

In the United States, employers are required to administer tests so that the result reflects what the test is meant to measure rather than an applicant's impairment, and to provide reasonable accommodation in testing for a known disability unless doing so would impose undue hardship. In one case cited in the Equal Employment Opportunity Commission's own guidance, applicants with learning disabilities who needed reading support received either a reader or an audio version of the test.

Practically, that means extra time, a screen reader, an audio version, a paper version, additional breaks or a quiet separate room can all be reasonable requests, depending on the disability and the test. The request is made to the employer, before the test window if possible, and it does not require you to disclose a full medical history, only the limitation and the accommodation that addresses it. Rules differ outside the United States, so check the local framework.

Scheduling accommodation is also covered where a sincerely held religious belief conflicts with the testing requirement. If a timed assessment lands on a day you cannot sit it, asking for a different window is a request employers are expected to consider.

How our reasoning quizzes relate

Psychology.com's IQ Test is 30 original text-only questions, untimed, reported as an estimated range with a five-category breakdown and an optional in-depth report. It is self-administered, unproctored and unnormed, so it is neither an IQ score nor a prediction of how you would do on an employer's screen. Being untimed, it deliberately removes the feature that matters most in hiring tests.

For format practice specifically, the single-domain quizzes are closer to the real thing: our Verbal Reasoning Test, Numerical Reasoning Test, Abstract Reasoning Test and Logical Reasoning Test each cover one of the question types you are likely to meet, and the Working Memory Test (Digit Span) covers a task some batteries include. Use them to learn what the instructions look like and how the items behave, then do your real practice against the test the employer actually named, if they named it.

What none of them can give you is a percentile or a hiring prediction. No free online test can, because that requires a reference sample of candidates for that job, which only the vendor has. Anyone offering you one is guessing.

What the research says

The well-supported part is that these tests measure something real and measure it reliably. Reasoning performance is stable enough to be worth assessing, and the association with job performance is positive across hundreds of studies. Nobody serious argues that the correlation is zero.

The revision is the news, and it is a methodological correction rather than an ideological one: the standard correction for range restriction had been applied in a way that inflated published validity estimates, and fixing it moved every predictor down, with cognitive ability falling further than structured interviews. The practical consequence is a reordering of what employers should rely on, with structured interviews and work samples ahead of a reasoning screen.

The least supported use is the one applicants fear most: reading a hiring test result as information about yourself. It is not designed for that, it is not fed back in a form that would support it, and its predictive power even for the narrow question it was built for is weaker than the industry that sells it has been saying.

FeatureEmployer's cognitive ability testClinical intelligence assessment
PurposeRank applicants for a roleAnswer a referral question about one person
Who gives itNobody, taken online unsupervisedA qualified clinician, one to one
LengthRoughly 10 to 40 minutesOne to two hours for the appointment
CoverageVerbal, numerical, abstract reasoningFive domains including spoken and visual tasks
TimingUsually tight, often by designSome subtests timed, most not
What you receiveOften nothing at allIndex scores, percentiles, intervals, a written report
Can be appealed or explainedRarely, beyond requesting accommodationYes, discussed with the clinician
What it says about youWhether you passed one filterA profile across domains, with caveats

Key takeaways

  • A cognitive ability test is a short, usually timed reasoning screen used in hiring: verbal, numerical and abstract items, often adaptive, with right answers and no interpretation.
  • It is not a clinical IQ assessment, which is given one to one over one to two hours across five domains, and it is not a personality questionnaire.
  • The validity estimate for cognitive ability was revised from about .51 to about .31 in 2022 after a re-analysis found earlier corrections had systematically overcorrected, with structured interviews now ranking highest at about .42.
  • Adverse impact is the central practical problem: under United States guidelines a selection rate below four fifths of the highest group's rate is generally treated as evidence of it, and the test must then be shown job related.
  • Practice yields roughly a quarter of a standard deviation, about four points on a fifteen-point scale, with the gain coming mostly from format familiarity and pacing rather than from content study.
  • In the United States, reasonable accommodation in testing is required for a known disability, and extra time, a reader, an audio version or a quiet room can all be reasonable requests.

Want to see the question types before they count?

Our free IQ Test gives 30 untimed text reasoning questions with an estimated range and a five-category breakdown. It is practice for the format, not a hiring prediction.

Take the free IQ Test

Cite this source

Fontane Pennock, S. (2026, November 8). Cognitive Ability Tests Explained. Psychology.com. https://psychology.com/types/cognitive-ability-test

Frequently asked questions

What is a cognitive ability test?

A standardized reasoning assessment used to compare job applicants. It has right answers, is usually timed, and typically mixes verbal items based on a short passage, numerical items based on a table or chart, and abstract items where you infer a rule from shapes or sequences. Most are taken online and unsupervised, and many are adaptive, choosing each item based on your previous answer.

Is a cognitive ability test an IQ test?

No. It shares the reasoning content and almost nothing else. A clinical intelligence battery is given one to one by a qualified clinician across five domains over one to two hours and produces a written interpretation with percentiles and confidence intervals. A hiring screen samples a narrow slice, is usually unsupervised, and produces a rank, not an interpretation.

How do employers use cognitive ability test results?

Usually as an early filter, either as a cutoff score or as one component of a composite alongside interviews and work samples. The rationale came from older meta-analyses putting the validity around .51. That estimate was revised to about .31 in 2022, so employers relying on a cutoff are placing considerable weight on a modest signal.

Can you prepare for a cognitive ability test?

To a limited extent. A meta-analysis of retesting found an overall gain of roughly a quarter of a standard deviation, larger when the same form was reused or when coaching preceded the retest. That is about four points on a fifteen-point scale. The usable part is format familiarity, timing practice, knowing whether guessing is penalized, sleep and a quiet room.

What happens if I do badly on a hiring test?

Most often you are screened out of that application, usually without feedback. It is worth knowing that you have failed a filter with modest predictive power rather than been measured accurately. If a disability affected your performance and you did not request accommodation, it is reasonable to raise that with the employer, since in the United States accommodation in testing is required for a known disability.

Why are cognitive ability tests controversial?

Because they produce average score differences between demographic groups large enough to screen out protected groups at different rates, which is the adverse impact problem. Under United States guidelines a test with adverse impact must be shown to be job related and consistent with business necessity, and a challenger can point to an equally predictive alternative with less impact.

Do your free tests predict how I will score for an employer?

No. Ours are unproctored, unnormed and, in the case of the IQ Test, untimed, which removes the feature that matters most in hiring. They are useful for seeing what each question type looks like before it counts. A percentile among candidates for a particular job can only come from the vendor's own reference data.

Related types

References

  1. Sackett PR, Zhang C, Berry CM, Lievens F. Revisiting meta-analytic estimates of validity in personnel selection: Addressing systematic overcorrection for restriction of range. Journal of Applied Psychology. 2022;107(11):2040-2068. doi.org
  2. Schmidt FL, Hunter JE. The validity and utility of selection methods in personnel psychology. Psychological Bulletin. 1998;124(2):262-274. doi.org
  3. Sackett PR, Demeke S, Bazian IM, Griebie AM, Priest R, Kuncel NR. A contemporary look at the relationship between general cognitive ability and job performance. Journal of Applied Psychology. 2024;109(5):687-713. (The published version of the 2022 conference analysis.) doi.org
  4. Ployhart RE, Holtz BC. The diversity-validity dilemma: strategies for reducing racioethnic and sex subgroup differences and adverse impact in selection. Personnel Psychology. 2008;61(1):153-172. doi.org
  5. Hausknecht JP, Halpert JA, Di Paolo NT, Moriarty Gerrard MO. Retesting in selection: A meta-analysis of coaching and practice effects for tests of cognitive ability. Journal of Applied Psychology. 2007;92(2):373-385. doi.org
  6. US Equal Employment Opportunity Commission. Employment Tests and Selection Procedures (fact sheet on the Uniform Guidelines, employer obligations, and accommodation in testing, including the Daimler Chrysler settlement providing a reader or an audiotape). eeoc.gov
Important: A personality type, attachment style, or result pattern is an educational way of describing tendencies, not a diagnosis, a fixed identity, or a substitute for professional care. If you are struggling, reach out to a licensed mental-health professional.