What the Watson-Glaser test measures and how its score is read

0 The short answer

The Watson-Glaser Critical Thinking Appraisal is a Pearson TalentLens test of how well an adult recognizes assumptions, evaluates arguments and draws conclusions from written passages, and in its third edition, the version Pearson's April 2020 report calls current, it presents 40 items under a 30 minute limit that an employer or university sends to the candidate, so it is not a test an individual buys, and the pages we opened list no price. Pearson offers a free practice version, but it is unscored, and nothing from a practice seller is the live test, so the sound preparation is to learn the format and read the instructions you are sent. If you want to see your own reasoning abilities on adult norms, ACIS Quick costs 15 dollars one time and takes about 45 minutes (prices read October 29, 2026), and it is a different test, not a substitute for the Watson-Glaser. The decisive limit is that a Watson-Glaser percentile ranks you inside a norm group someone else chose, and the strongest validity figures are the publisher's own.

1 What Is the Watson-Glaser Critical Thinking Appraisal, and Where Does It Come From?

The Watson-Glaser is a critical thinking test whose publisher dates its development to 1926 and its first publication to 1964, by Goodwin Watson and Edward M. Glaser. That history comes from Pearson's own document, the Watson-Glaser Critical Thinking Appraisal III efficacy report of April 2020, which we opened on October 29, 2026. The report says the test has since become the leading method of assessing critical thinking globally, which is the publisher's claim about itself, and we report it as a claim. The firm side tells the story slightly differently: a Clifford Chance assessments page, also opened that day, says companies have used the test since 1925. The sources disagree on the year, and because the publisher's report is the document that dates its own product, we follow it.

The same report prints a timeline of the editions. It lists two parallel 100 item forms in the original release, two parallel 80 item forms in a revised release, an adaptation for the United Kingdom in 1991, a 40 item short form in 1994, and the Watson-Glaser II in 2010, which introduced the model Pearson calls RED. The third edition is the one that matters now, because it replaced fixed forms with an item bank. The timeline gives two release dates for it, 2012 and 2018, one for each of the United Kingdom and the United States, and the layout of the PDF does not let us tell with certainty which belongs to which, so we give both years and no more.

The edition is the first fact to check, because the older and newer forms differ in length and in scoring. Pearson's report calls the third edition the current version. The TalentLens page for the test, which we opened on October 29, 2026, does not print an edition number, so we cannot tell you from the publisher's pages whether a newer edition exists today, and we do not claim that one does. If an employer's invitation names an edition, that is the one whose rules apply to you.

The test belongs to a family of tests of reasoning from text. The page on critical thinking tests in general covers the whole category, and the page on critical thinking versus intelligence owns the question of whether such a test measures something different from IQ.

2 What Is on the Watson-Glaser III: Subscales, Number of Items and Time?

The Watson-Glaser III gives each test taker 40 items in three areas, 12 on recognizing assumptions, 12 on evaluating arguments and 16 on drawing conclusions, under a time limit of 30 minutes. The item split comes from the efficacy report, which says the structure was informed by the RED model, short for recognize assumptions, evaluate arguments and draw conclusions. The 30 minute limit comes from Pearson's one page efficacy summary, a PDF we opened on October 29, 2026, which also says an untimed version is available in US English for development or reasonable accommodations. The TalentLens page for the Watson-Glaser, read the same day, prints "30 to 40 mins timed" instead, so the publisher's pages state the time slightly differently, and the invitation you receive is the authority on your own sitting.

The older structure still appears in the vocabulary. The efficacy report says the test scores five subscales: Recognition of Assumptions, Evaluation of Arguments, Inference, Deduction and Interpretation. It adds that Pearson ran factor analyses of the original test and found that Inference, Deduction and Interpretation factored together, so since the 2010 release those three have been combined under the heading Draw Conclusions. The TalentLens page lists the five as subtests and describes each in a line, while its profile report lists only the three areas. Guides that disagree about five sections or three are describing different layers: five task types, three reported areas.

In Pearson's terms, recognizing assumptions means noticing what a statement takes for granted without evidence, evaluating arguments means judging how strong and relevant an argument is, and drawing conclusions means reaching only what follows from the evidence given. This page describes the task types and no item.

The delivery details matter to a candidate. The efficacy report says the third edition draws 40 items at random from a large bank, so two candidates are unlikely to see the same set, and the report presents this as a guard against items becoming over-exposed. It also says the test presents passages or scenarios of two kinds, neutral and controversial. Neutral scenarios deal with subjects such as the weather, scientific facts or common business situations, and controversial ones refer to political, economic and social issues that frequently provoke emotional responses. The reason, as the report puts it, is that strong attitudes can affect how well people think critically, so the results should show whether a person can reason well whether or not they have strong feelings about the subject. The test can be completed online and, per the one page summary, is suitable for supervised and unsupervised administration. The TalentLens page lists the languages as UK English, US English, French, French Canadian, Dutch, Spanish and Latin American Spanish.

Pearson itself classes the test with verbal reasoning. Its report, discussing a comparison with a numerical test, says the Watson-Glaser is mainly a verbal reasoning test. That matches what a reader would expect from a test made of written passages, and it is why the siblings of this page are the verbal reasoning test page and the deductive reasoning test page: the Watson-Glaser sits between them, with deduction as one of its tasks and written language as its medium.

3 How Is the Watson-Glaser Scored, and What Does a Percentile Mean?

A Watson-Glaser result starts as a raw score from 0 to 40, and its meaning comes entirely from the norm group the raw score is compared with, so the percentile on your report is a rank inside a group that someone chose. Pearson's efficacy report explains that each response is correct or incorrect and that the number of correct responses gives a raw score between 0 and 40. Raw scores can rank test takers but little else can be inferred from them, so administrators compare a raw score with a large sample of others who took the same test, a normative group. The report gives an illustration and not a norm: a test taker with 35 out of 40 compared with a manager norm might have a percentile rank of 86, which would mean scoring equal to or higher than about 86 percent of managers. Treat the 86 as a teaching example from the report, since it states no real norm table.

The choice of group is the part a candidate cannot see. The report says the ideal norm group represents those taking the test in the context at hand, and that an administrator selecting candidates for employment might compare them with peers in the profession, while one interested in intelligence testing might compare with the general population. It says the test comes with 18 norms constructed by Pearson, including occupations such as accountant, engineer and human resource professional and organizational levels from entry level to executive, that administrators can build their own norms and that a sample of at least 200 test takers is ideal for that. The TalentLens page adds that custom company norms can be mapped for a fee. The same raw score can therefore sit at different percentiles on different norms, and a percentile against graduate applicants is not a percentile against managers.

That is the same logic as in any norm referenced test, and the pages on how IQ scores are normed and on IQ scores versus percentiles explain it for IQ. The Standards for Educational and Psychological Testing (AERA, APA and NCME, 2014), in the open access edition at testingstandards.net, say in Standard 5.8 that norms, if used, should refer to clearly described populations, including those with whom test users will ordinarily wish to compare their examinees. Standard 5.9 asks that norming reports specify the population sampled, the sampling procedures, the dates of testing and descriptive statistics. Pearson's public pages do not print the norm tables or the dates of each norm, because the technical manual that holds them is, in the report's words, available to current or prospective users on request. We did not read that manual, and so we report no norm statistics.

Three reports come with the test according to the TalentLens page: a Profile Report with an overall percentile and percentiles for the three areas, with the number correct, a T-score, a stanine and a sten in its technical section; a Development Report that supports a learning plan; and an Interview Report that supports a structured interview about critical thinking behavior. The efficacy report adds that when the test is used for selection, administrators might compare only the overall scores, while for development the area breakdown shows where to concentrate. For a candidate this means that the number an employer sees may be the overall percentile alone, and the three area percentiles, being based on 12, 12 and 16 items, are shorter and so less precise than the total, a point the section on reading your result takes up.

The TalentLens page also says that use requires a training qualification for administrators, the British Psychological Society Test User: Occupational Ability course at Level A, which it says is not offered to customers in the United States and Canada, and it lists no price. This is a test an organization buys and a candidate takes.

4 Which Law Firms and Graduate Schemes Use the Watson-Glaser, According to Their Own Pages?

Some large law firms send the Watson-Glaser to applicants early in their selection, and some universities have added it to admissions, but the firm pages we opened state that the test is requested and how long applicants have, not how the result is weighed. On the Clifford Chance assessments page, which is written in Dutch and which we opened on October 29, 2026, the firm says applicants receive an email invitation with a link, that the test lasts 30 minutes and has 40 questions, and that applicants are asked to finish within 5 days of receiving the invitation. The page links a practice test and says practice attempts are not part of the assessment and that the firm does not see their results. It also lists the skills the test covers, among them recognizing stated and unstated assumptions and drawing valid conclusions.

On Linklaters' early careers page for Hong Kong, which we opened the same day, the firm says that after you submit your application form it will also ask you to take a test of your critical and logical thinking, and that candidates have 5 days from the date of their application to complete the Watson Glaser online test. The page strongly advises completing a practice test and points to a practice version on Pearson's website. It contrasts this with another assessment on the same page, for which it says no preparation is required. A page for one region is not a statement about every office, and the firm's own wording applies only to the process it describes.

An older example shows how long the pattern has run. In an ABA Journal report of October 26, 2010, which we opened on October 29, 2026, the journal reported that Simmons and Simmons required applicants to pass written tests, including a critical thinking assessment that the article identifies as the Watson Glaser test, before a second interview, and that Hogan Lovells also used it. The report gives the firm's aim as reducing time spent interviewing. It is 16 years old, so it shows the practice and not the current process of either firm.

Universities and training courses appear in Pearson's own material. The efficacy report describes Ritsumeikan Asia Pacific University in Japan, which in 2019 began to require the Watson-Glaser as part of the application for an initial cohort of 500 international students, as a way to compare applicants from very different educational backgrounds. Pearson says thousands of organizations, colleges and schools use the test, and we found no source that counts them, so we make no claim about how many there are.

What none of these pages gives a candidate is a pass mark. The firm pages describe an invitation, a deadline and a practice option. The publisher says norms depend on the group chosen. If you need to know how your result will be used, the employer is the only source, and the page on pre-employment cognitive tests covers what such screening tests are for and how employers use them in general. The page on the SHL test covers another family of employer assessments that candidates meet at the same stage, and the page on IQ versus the LSAT covers the admission test for law school, which is built on a different design and a different norm group.

5 Does the Bar Course Aptitude Test Still Exist?

The Bar Course Aptitude Test, a Watson-Glaser based test that England and Wales once required before Bar vocational training, is no longer a requirement, because the Bar Standards Board announced that it would abolish the requirement from July 31. The Bar Standards Board is the regulator. On its announcement page "BSB to abolish the Bar Course Aptitude Test from 31 July", which we opened on October 29, 2026, the Board says that after the Legal Services Board approved its rule change it will abolish the requirement that students pass the test before starting vocational training. Its Director General is quoted as saying the Board no longer thinks the test serves a useful purpose. The notes on the page say the test is based on a Watson-Glaser psychometric test and is delivered by Pearson VUE.

The page prints no year. A companion announcement, "BSB seeks to abolish the Bar Course Aptitude Test", also opened that day, records a Board decision on March 31, 2022 to seek that approval. Reading the two pages together, we infer that the abolition took effect on July 31, 2022, and we label that as our inference. The companion page also describes the test: computerized, 55 minutes, 60 multiple choice questions, designed to assess critical thinking and understanding of arguments, delivered by Pearson VUE at 150 pounds in the United Kingdom and 170 pounds abroad. It says the test was introduced in 2013 because too many students with little chance of completing vocational training were being enrolled.

The page says the review found the test no longer necessary, because providers now take a more selective approach to admissions, and that the 2019 Bar Qualification Rules and the new Authorisation Framework require providers to have clear and robust admissions policies, which the Board believes work better. That is a regulator's judgment about its own test and not a finding about the Watson-Glaser. We did not find a current Bar Standards Board page that requires the test, and we do not know whether Pearson VUE still offers it privately, so we say only that the requirement was abolished. The old test is therefore a case study and not a hurdle. Any guide that tells a prospective barrister to book it as a requirement predates the change.

The test also appears in the validity evidence. Pearson's efficacy report cites a correlation of 0.62 between Watson-Glaser scores and final course grades for 123 legal training course students, and describes the course as the Bar Professional Training Course, the postgraduate vocational course for aspiring barristers in England and Wales. The next section reports that figure with the rest of the research.

6 What Does Research Say the Watson-Glaser Correlates With?

The strongest figures come from Pearson's own reports and range from 0.16 to 0.58 against work indicators and from 0.38 to 0.62 against course grades, while the independent studies we opened are few, concern older forms and raise questions about the subscales. A correlation is the measure used throughout: 0 means no linear relation and 1 means a perfect one. The table lists what Pearson's efficacy report states, with the report's own citations. We did not read the cited technical manual and studies, so we cannot say how the correlations were computed or whether they were adjusted for range restriction or unreliability.

Pearson judges these against the US Department of Labor guidelines, under which a coefficient above 0.35 is labeled very beneficial, 0.21 to 0.35 likely to be useful, 0.11 to 0.20 dependent on the circumstances and below 0.11 unlikely to be useful. By that scale, the lowest work figure falls in the dependent band and the highest course figure in the very beneficial one. The labels are the publisher's. Squaring a correlation shows how much variance two scores share, and that is our arithmetic and not the report's: 0.16 shares under 3 percent, 0.33 about 11 percent, 0.53 about 28 percent and 0.62 about 38 percent. Even the highest figure leaves most of the variation in grades unexplained, which is the normal situation for a single test, and it is why the page on IQ and job performance treats cognitive ability as one predictor among several for work outcomes.

The reliability claims follow the same pattern. Pearson's report gives an internal consistency of 0.83 for the third edition, computed on 147 working adults in the United States, test-retest correlations of 0.73 to 0.89 across studies of various versions, and alternate form correlations of 0.82 and 0.88 in two United Kingdom studies of 355 and 318 participants. These are publisher figures. The page on reliability and validity explains why a consistent score is not automatically a valid one.

Two independent sources complicate the picture. In a 2008 meta-analysis in Thinking Skills and Creativity, of which we read the abstract, Robert Bernard and colleagues analyzed 13 sets of subscale intercorrelations and 60 sets of subscale means from published studies of the five subscale versions. Each of two versions produced a one factor solution, accounting for 82.69 percent and 79.55 percent of the variance, and the authors concluded that the test should be viewed as a measure of general competency and that the subscales should not be interpreted individually. That fits the publisher's own decision to fold three of the five subscales into one, and it is a reason to treat the area percentiles with caution.

A 2026 open access study in Frontiers in Psychology tested the test on Slovak university students. In Šeboková and colleagues' paper, which we opened on October 29, 2026, 264 students took the Slovak Form C of what the authors call the WGCTA-II, an 80 item test with five subscales that, by their account, dates to the 1980 revision, so it is not the 40 item Watson-Glaser II that Pearson released in 2010. A second sample of 137 students was used to replicate the shortened version. The authors describe the total reliability as moderate to good, with an alpha of 0.691 and an omega of 0.846, but caution that higher values may reflect test length, and the five subscales had alphas of 0.23 to 0.51, which they call weak. Their abstract reports limited construct validity at the subscale level. They also correlated a nine item shortened version with the Cognitive Reflection Test and found 0.381 with its numerical version and 0.299 with its verbal version, against 0.329 and 0.298 for the full test, and 0.303 and 0.307 for the shortened version in the second sample. The authors state that their results are limited to this older version, the only standardized one available in Slovakia. For the Cognitive Reflection Test itself, see the page on the Cognitive Reflection Test.

One last contrast guards against over-reading. In a 2012 study in Applied Cognitive Psychology, of which we read the abstract, Butler found that community adults and students with higher scores on a different instrument, the Halpern Critical Thinking Assessment, reported fewer negative life events, r(131) = -0.38. That study did not use the Watson-Glaser. Pearson's outcome studies concern job ratings and grades, and we did not find a Watson-Glaser study of life outcomes in the sources we opened. The page on whether critical thinking differs from intelligence gives the wider argument.

7 Can You Take the Watson-Glaser Online for Free, or Practice With Sample Questions?

You cannot sit the live Watson-Glaser on your own for free, because it is delivered through an organization that buys it, but Pearson publishes a free practice version that it says it cannot score, and a meta-analysis of retesting on cognitive ability tests found an adjusted overall effect size of .26. The live test is delivered online, remotely, with instant results, per the TalentLens page, and the invitation comes from the employer or school. That answers the search "Watson-Glaser test online": the online test is the one you are invited to, and a score from any other site is not a Watson-Glaser score.

Pearson's practice tests page, which we opened on October 29, 2026, offers a free practice version in English and Dutch. The page says it is not a live test, that it is meant to show how the test works, what its subtests look like and the skills needed, and that because its questions have not been extensively trialed and analyzed like the live questions, it cannot accurately score the test or share results in a meaningful format. On completion you receive a report with advice on improving your critical thinking and preparing for the live test. The page does not state how many questions the practice version has or how long it takes. This is the one practice resource this page can describe from its publisher. The Linklaters page points to a Pearson practice version, and the Clifford Chance page links a practice test and says practice attempts do not count and are not visible to the firm.

Sample questions, answer keys and pdf files from other sites are a different matter. Many sellers sell them, and none of them is a source for this page or linked from it. Two reasons stand behind that. First, the publisher says it draws live items from a large bank precisely so that items are not over-exposed, which means a memorized list is aimed at a test that no longer works that way. Second, the Standards for Educational and Psychological Testing (AERA, APA and NCME, 2014) say in Standard 1.7 that if test performance is claimed to be essentially unaffected by practice and coaching, the propensity to change with them should be documented, and that materials for test takers should give practical guidance about the value of test preparation, including coaching. This page reproduces no item and describes no answer.

What does retesting research show? In a 2007 meta-analysis in the Journal of Applied Psychology, of which we read the abstract, John Hausknecht and colleagues summarized 50 studies of practice effects on cognitive ability tests. They note that as many as 25 to 50 percent of applicants in organizational and educational settings are retested, and across 107 samples and 134,436 participants they found an adjusted overall effect size of .26. The effects were larger when practice came with coaching and when identical forms were used. That is a finding about cognitive ability tests as a group, not about the Watson-Glaser, and the item bank makes identical forms less likely on the third edition, so the figure is a general guide and not a prediction.

If you want practical guidance, the sources support a short list. Read the instructions you are sent, because the time limit, language and any untimed option are set by the organization. Use the free Pearson version to see the format, knowing it is unscored. Expect pacing to be manageable: in a study of 137 participants that the efficacy report describes, each took the test twice online, once with a 30-minute limit and once with none, there was no significant difference between average scores, and all but one person finished all items both times, the exception completing 39 of 40 under the limit. That is a publisher study, and it suggests time is rarely the binding constraint. Our arithmetic puts 40 items in 30 minutes at about 45 seconds each.

8 Which Real Options Can an Adult Use, and How Do They Compare?

The live Watson-Glaser, Pearson's free practice version, a psychologist's battery and a normed online battery answer four different questions, and none of them can stand in for another. The table compares them on what each measures, who the result is compared with and what it costs. Every Pearson fact comes from a page we opened on October 29, 2026, prices can change, and where a page printed no price the cell says so.

The WAIS-5 row comes from Pearson's page for the WAIS-5, which lists the qualification level, the ages, the times and the kit price above. Three readings follow. First, only the live Watson-Glaser produces a Watson-Glaser percentile, and an employer receives it, so every other row is a different instrument. Second, a psychologist's battery is the route to an official ability score, and the pages on every way to take an IQ test and on IQ tests for adults lay out the options for an adult. Third, a normed online battery shows a profile of general abilities, which is a different thing from a critical thinking percentile, as the next section explains.

9 Which Abilities Sit Around the Watson-Glaser, and Which ACIS Subtests Sample Them?

The Watson-Glaser asks you to reason about written statements, so the abilities around it are comprehension of language, deductive and inductive reasoning, and the integration of several conditions at once, and in the Cattell-Horn-Carroll model those fall, in our reading, mainly under crystallized knowledge (Gc) and fluid reasoning (Gf). Pearson does not use CHC terms, so this mapping is ours. It is a reading of the publisher's task descriptions and of the ACIS subtest pages, and it is not a finding of any study. Pearson itself calls the test mainly verbal reasoning, which is consistent with a Gc component, and the task of judging whether a conclusion follows is deduction, which the page on logic IQ tests treats as one of three narrow abilities within fluid reasoning. The page on the CHC model defines these abilities, and the page on what IQ measures explains the general factor behind them.

ACIS is not the Watson-Glaser, gives no Watson-Glaser score and does not measure the evaluation of arguments or the recognition of assumptions in written passages. What it offers is a look at reasoning abilities that sit near those tasks, on adult norms. Four of its subtests are relevant, each described only as its own page describes it.

• Logic Grid measures, in the page's words, controlled deduction under multiple constraints: the person keeps several rules active, eliminates impossible combinations and preserves the implications of each clue. The page relates it to fluid reasoning. It is in the Optimized and Full Scale forms.
• Similarities measures verbal concept formation, linking two items through a shared category, function or principle, and the page relates it to crystallized ability (Gc). It is in all three forms.
• Complex Relations measures relational integration, solving problems where several elements correspond, change or combine at once, and the page relates it to fluid reasoning. It is only in the Full Scale form.
• Matrix Reasoning measures rule discovery in unfamiliar visual patterns, with content that is usually nonverbal, which the page relates to fluid reasoning. It is in all three forms.

Two cautions keep this honest. The first is that overlap in ability is not overlap in task. Pearson reports a correlation of 0.53 between the Watson-Glaser and Raven's Advanced Progressive Matrices, which is publisher evidence that the test shares variance with matrix style reasoning, but a matrix result does not become a Watson-Glaser result by that route, and no ACIS result converts to or predicts one. The second is that the Watson-Glaser also depends on things an ability battery does not measure, such as how a person handles political or emotional content, which Pearson builds into its controlled scenarios on purpose. The page on verbal IQ tests shows what a verbal score covers.

Why would an adult look at this at all? Because a critical thinking percentile tells you little about why a result came out as it did. A profile of verbal knowledge, deduction and abstract reasoning shows where your abilities sit relative to adults in general, and it is private to you. That is self knowledge, not preparation, and the page on the cognitive ability test explains what such a profile contains.

10 What Conditions Does a Timed Online Session Need?

A timed online session needs the same basics for any test: a working device, a quiet room, the instructions read first and any accommodation arranged before you start, because the Watson-Glaser leaves those arrangements to the organization that sends it. Pearson's one page summary says the test is suitable for supervised and unsupervised administration and online completion, and the TalentLens page says the same, adding that it can be delivered remotely with instant results. The summary also lists an untimed version in US English for development or reasonable accommodations. A candidate who needs an adjustment, such as extra time, should ask the organization that issued the invitation before the deadline, since the organization controls which version is sent and the firm pages show deadlines of days.

Language matters as much as format. The TalentLens page lists seven language versions. Pearson's efficacy report states that its analyses found no evidence of bias against men or women, against people whose primary language is not English or against people with disabilities, and that the test marginally favored younger candidates by an effect too small to matter in practice. These are the publisher's analyses of its own data, which we did not read at the level of the technical manual. The page on whether IQ tests are biased explains how fairness is investigated and why a passed bias analysis is not a guarantee for every person. A reader whose first language is not English should note that a test of written passages draws on reading in the language it is given in, whatever the publisher's bias analyses conclude.

The APA's Guidelines for Psychological Assessment and Evaluation (2020), which we opened as a PDF on October 29, 2026, say in Guideline 8 that psychologists strive to be aware of the need for test selection, scoring and administration to reflect the appropriate normative comparison, situational influences, effort and standardized administration as indicated.

For an ACIS session the conditions are different. ACIS is online and unsupervised, in English only, and Digit Span, Alphanumeric Sequencing and Arithmetic present their items as audio, so a session without working sound cannot complete them. It allows breaks, saves progress between subtests and gives 30 days to complete a form. Its technical manual says verbal and knowledge tasks are affected by English proficiency and educational exposure, which is the same caution as above.

11 How Should You Read a Watson-Glaser Result?

Read a Watson-Glaser result as a rank among a stated group on one occasion, with the overall percentile resting on more items than any area percentile, and never as an IQ or a verdict on how well a person thinks. Three facts from this page support that reading. The percentile depends on the norm group, which the administrator picks. The overall score rests on 40 items, and each area on 12, 12 or 16, so the area scores are shorter. And a meta-analysis of the five subscale versions found that the subscales behave like a single dimension, while a Slovak study of an older form found low reliability at the subscale level. An employer that sees only an overall percentile, as the efficacy report says some do for selection, is therefore working with the more defensible number.

A percentile is not an equal interval scale, and the standard scores that IQ tests use are one way of spreading it out. The table shows how a few percentile ranks correspond to standard scores on the IQ scale (mean 100, standard deviation 15) if a norm group's scores follow a normal curve. That is our arithmetic and an assumption, because Pearson does not publish its distributions and a norm group's scores need not be normal.

The page on why the standard deviation is 15 explains the scale. The percentile of a Watson-Glaser score refers to its own norm group, so the table is a way to see the geometry and does not turn a Watson-Glaser percentile into an IQ. The same warning applies in the other direction: a high percentile among managers does not mean a high IQ, and a lower percentile among graduate applicants does not mean a low one.

Differences between scores need their own caution. If a person's three area percentiles differ, the question is whether the gap exceeds measurement error, and the page on index score differences shows how common sizeable gaps are in the general population on a long battery. For the Watson-Glaser, the public pages we opened print no subscale reliability, and the Slovak study found low subscale reliability on an older form, so a gap between areas is a prompt for questions, not a diagnosis. Group statistics describe samples and never decide what one person can do.

12 Where ACIS Sits, and What a Buyer Should Do

ACIS is an online, unsupervised, English only battery of cognitive tasks with adult norms, and it is not the Watson-Glaser, not a preparation for it and not an instrument for an employer, so a buyer should treat its result as a private look at their own reasoning abilities. The limits come first. ACIS "is not intended for clinical diagnosis, hiring decisions, accommodation determinations or admission to high IQ societies," as its research page states. It cannot tell you how an employer will weigh a Watson-Glaser result. The Standards for Educational and Psychological Testing (AERA, APA and NCME, 2014) say in Standard 11.1 that before an employment test is developed and implemented, a clear statement of the intended interpretations of test scores for specified uses should be made, and that validation should show how well this has been achieved for all relevant subgroups. That standard describes what an employment test and its publisher must show, and ACIS makes no such claim. The APA Guidelines for Psychological Assessment and Evaluation (2020) add, in Guideline 6, that tools should show sufficient validity evidence for their uses and sufficient score reliability, and, in Guideline 7, that psychologists use multiple sources of relevant and reliable information.

The strengths are real and bounded. Every ACIS form reports a Full Scale IQ with its 95 percent confidence interval against adult norms covering ages 16 to 90, and the Full Scale form measures all 20 subtests in six domains, with an index for each domain. It is a one time payment with no subscription: Quick costs 15 dollars, Optimized 30 dollars and Full Scale 50 dollars (prices read October 29, 2026, which can change). A free trial needs no card, and one trial is allowed per person. The Custom Test lets a buyer pick subtests at 5 dollars each. The refund wording is: "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 money back guarantee. The technical manual documents the instrument, and this page quotes no figure from it.

The concrete next step depends on why you searched. If you have an invitation, read it for the time limit, the deadline, the language and the untimed option, try Pearson's free practice version once to see the format, and ask the organization about accommodations before the deadline. If you want to see where your reasoning abilities stand for yourself, open the Full Scale example report, start the free trial with no card, and pick the form whose subtests cover the abilities you care about: Quick has Similarities and Matrix Reasoning, Optimized adds Logic Grid, and Full Scale adds Complex Relations. Do not report any online result as a Watson-Glaser score.

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. Where we read only an abstract, the entry says so. Publisher claims are reported as claims, firm and regulator pages were read as published that day, and prices and processes change, so check the page before you rely on it.

1. Pearson TalentLens (2026). Watson-Glaser, product page. Subtests, items, time, reports, languages, norm groups, user training and validity claims. Read on October 29, 2026. https://www.talentlens.com/watson-glaser.html
2. Pearson (2020). Watson-Glaser Critical Thinking Appraisal III (WG-III): Assessment Efficacy Report, April 2020, PDF. History, RED model, item counts, norms, correlations, reliability, fairness and timed versus untimed study. Read on October 29, 2026. https://www.talentlens.com/content/dam/school/global/Global-Talentlens/uk/CriticalThinking/CTHub/Pearson_Efficacy_Watson-Glaser_2020.pdf
3. Pearson (2020). Watson-Glaser Critical Thinking Appraisal: Efficacy Report Summary, one page PDF (file last modified May 4, 2020). Time limit, untimed version, supervised and unsupervised use. Read on October 29, 2026. https://www.pearson.com/content/dam/one-dot-com/one-dot-com/global/Files/efficacy-and-research/reports/Watson-Glaser_One_Page_Summary.pdf
4. Pearson TalentLens (2025). Practice Tests, web page (copyright footer 1996 to 2025). The free, unscored practice version. Read on October 29, 2026. https://www.talentlens.com/practice-tests.html
5. Clifford Chance (2026). Assessments, recruitment web page, in Dutch. Questions, time, deadline and practice. Read on October 29, 2026. https://jobs.cliffordchance.com/assessments
6. Linklaters (2026). Your Application, early careers page for Hong Kong. Deadline and practice advice for the Watson Glaser online test. Read on October 29, 2026. https://www.linklaters.com/en-hk/careers/early-careers/your-application
7. Weiss, D. C. (2010, October 26). UK law firm requires critical thinking test for job seekers. ABA Journal. Read on October 29, 2026. https://abajournal.com/news/article/uk_law_firm_requires_critical_thinking_test_for_job_seekers
8. Bar Standards Board. BSB to abolish the Bar Course Aptitude Test from 31 July, announcement (page prints no year). Read on October 29, 2026. https://www.barstandardsboard.org.uk/resources/bsb-to-abolish-the-bar-course-aptitude-test-from-31-july.html
9. Bar Standards Board. BSB seeks to abolish the Bar Course Aptitude Test, announcement of the March 31, 2022 decision, with the test's description and fees. Read on October 29, 2026. https://www.barstandardsboard.org.uk/resources/bsb-seeks-to-abolish-the-bar-course-aptitude-test.html
10. Bernard, R. M., Zhang, D., Abrami, P. C., Sicoly, F., Borokhovski, E., and Surkes, M. A. (2008). Exploring the structure of the Watson-Glaser Critical Thinking Appraisal: One scale or many subscales? Thinking Skills and Creativity, 3(1), 15 to 22. Abstract read. https://doi.org/10.1016/j.tsc.2007.11.001
11. Šeboková, G., Ráczová, L., Uhláriková, J., Lukačková, E., and Sollár, T. (2026). Psychometric evaluation of the full and shortened versions of the WGCTA-II in Slovak university students. Frontiers in Psychology, 17, article 1764712. Open access, read. https://doi.org/10.3389/fpsyg.2026.1764712
12. Butler, H. A. (2012). Halpern Critical Thinking Assessment predicts real-world outcomes of critical thinking. Applied Cognitive Psychology, 26(5), 721 to 729. Abstract read. https://doi.org/10.1002/acp.2851
13. Hausknecht, J. P., Halpert, J. A., Di Paolo, N. T., and Moriarty Gerrard, M. O. (2007). Retesting in selection: A meta-analysis of coaching and practice effects for tests of cognitive ability. Journal of Applied Psychology, 92(2), 373 to 385. Abstract read. https://doi.org/10.1037/0021-9010.92.2.373
14. Pearson Assessments (2026). Wechsler Adult Intelligence Scale, Fifth Edition (WAIS-5), product page. Qualification level, ages, times and kit price. Read on October 29, 2026. https://www.pearsonassessments.com/en-us/Store/Professional-Assessments/Cognition-%26-Neuro/Wechsler-Adult-Intelligence-Scale-%7C-Fifth-Edition/p/P100071002
15. American Educational Research Association, American Psychological Association, and National Council on Measurement in Education (2014). Standards for Educational and Psychological Testing. Standards 1.7, 5.8, 5.9 and 11.1, read in the open access edition. https://www.testingstandards.net/open-access-files.html
16. American Psychological Association (2020). APA Guidelines for Psychological Assessment and Evaluation, PDF. Guidelines 6, 7 and 8. Read on October 29, 2026. https://www.apa.org/about/policy/guidelines-psychological-assessment-evaluation.pdf

14 Frequently Asked Questions