The employer's verbal reasoning test: what it samples and what it skips

0 The short answer

An employer's verbal reasoning test is a timed reading task, not a verbal IQ score: you read a short passage and judge whether statements are true, false or cannot be said from that passage alone. SHL, a test publisher, describes its Verify Verbal Reasoning test in just those terms, with a maximum of 30 questions and 19 minutes allowed (fact sheet read October 29, 2026). The test samples comprehension knowledge (Gc), reading and writing (Grw) and inference under a clock, which is different from the stored word knowledge behind a verbal IQ score. ACIS has no passage task of this kind; as a normed online battery it can show an adult their own verbal profile, at 15, 30 or 50 dollars for its three forms or 25 dollars for the five verbal comprehension subtests alone (prices read October 29, 2026), but it is not preparation for, a predictor of or a substitute for any employer test.

1 What Is an Employer's Verbal Reasoning Test?

An employer's verbal reasoning test is a short, timed, usually unsupervised online assessment in which the candidate reads text and decides what follows from it, and its publishers describe that task in plain terms. SHL's fact sheet for Verify Verbal Reasoning, a document carrying a 2018 copyright line that we read on October 29, 2026, says the test measures "a candidate's ability to evaluate written statements." It adds that the candidate draws conclusions from written evidence, using content based on workplace experiences. The task, as SHL words it, is to read a passage and use that information to decide whether a given statement is true, false, or cannot be said from the information given.

The logistics on that sheet matter to a candidate more than the marketing does. Average testing time is 17 to 19 minutes, the allowed time is 19 minutes, and the maximum number of questions is 30, all in multiple choice format. The sheet records that the test is designed for an unproctored environment and involves one sitting plus a verification test. It does not state a norm group, a reliability coefficient, a validity coefficient, a price, or what a candidate receives afterwards, and we do not guess at any of them.

SHL also sells a second, separate test with a similar name. Its Verify Verbal Ability fact sheet, also read on October 29, 2026, describes an adaptive, multiple choice test of 13 to 15 minutes on average, 15 minutes allowed and a maximum of 18 questions. That one asks the candidate to read written passages and comprehend them: to interpret tone and author intent, to identify main ideas and to make inferences. Its fact sheet does not describe the true, false or cannot say format. A candidate told only that the employer uses "SHL verbal" may be facing either, and the invitation email is the only reliable guide to which.

Other publishers vary again. Talogy's page for its Logiks cognitive ability test, read on October 29, 2026, says it gives insight into general, verbal, numerical and abstract ability, comes in intermediate and advanced versions, takes 12 to 25 minutes, has 20 to 50 items depending on version, and requires accreditation of the person who buys it. It does not describe its verbal items, so we cannot say whether they use a passage. The page on the Watson-Glaser test covers another publisher's format, and it is not assumed to match the one described here.

Employers do not always name the format in advance. UBS's page for its Graduate Talent Program, read on October 29, 2026, lists four online assessments (verbal reasoning, numerical reasoning, inductive-logical reasoning and a culture match) and says the ones assigned depend on the business area. It says they must be completed within seven days of the invitation email, only by the applicant, and that they cannot be retaken. It points applicants to practice assessments on another company's website. It does not describe the format of its verbal reasoning assessment, so the true, false or cannot say description above should not be assumed for it.

None of this makes the test an IQ test. The wider category, the short employer ability screens built to sort many applicants quickly, is set out on the page about the pre-employment cognitive test, which owns the employer side of the question. This page takes the candidate's side of one format.

2 Which Abilities Does a Passage and Statements Task Sample?

A passage followed by true, false or cannot say statements draws on three broad abilities in the Cattell-Horn-Carroll (CHC) model at once, comprehension knowledge, reading and writing, and fluid reasoning, and it is a mistake to call it any one of them. The CHC model is a widely used taxonomy of cognitive abilities. In the version McGrew set out in the Journal of Intelligence in 2023, drawing on Schneider and McGrew (2018), the broad abilities are named Gc (Comprehension-Knowledge), Grw (Reading and Writing), Gq (Quantitative Knowledge), Gf (Fluid Reasoning), Gwm (Short-term Working Memory), Gv (Visual-spatial Processing), Ga (Auditory Processing), Gl (Learning Efficiency), Gr (Retrieval Fluency) and Gs (Processing Speed). Three of those are in play here.

The first is Gc. Words, phrases and the background knowledge a workplace passage assumes are all stored language knowledge, and a reader who does not know what a "tender" or a "quarter" means in a business context starts a step behind. The ACIS description of Gc, on its subtest pages, is knowledge "acquired through schooling, reading, and everyday exposure." The page on crystallized intelligence testing covers how that knowledge changes across adult life, which is one reason verbal tests are read against an age reference.

The second is Grw. In the CHC model reading and writing is its own broad ability, separate from Gc, and a passage task is a reading task before it is anything else: the candidate must decode, parse and hold sentences whose meaning depends on small words such as "only", "most" and "unless." Treating reading as part of verbal intelligence in general is common in everyday speech, but the model separates them, and a person with a large vocabulary who reads slowly can score below a person with a smaller one who reads fast.

The third is Gf, reasoning with what is on the page rather than with what is stored. Deciding that a statement follows from the passage, contradicts it, or goes beyond it is an inference step, and the model files novel inference under fluid reasoning. A working memory demand sits underneath, because the reader has to keep the passage's claims available while testing a statement against them, which the model files under Gwm.

One caution on all of this. We opened no factor analytic study of SHL's test or of any other publisher's, so the mapping above is our reading of the task the publisher describes, applied to the CHC names, not a finding about how any particular test behaves. Where a seller states that a verbal reasoning test "measures verbal intelligence", the accurate reply is that it measures a mixture, and the mixture changes with the passage, the statement and the time allowed.

3 How Does a Verbal Reasoning Test Differ From a Verbal IQ Score?

A verbal reasoning test scores how well you draw conclusions from text placed in front of you, under a clock, for a hiring decision, while a verbal IQ score, or its modern replacement the Verbal Comprehension Index, reports how much language knowledge you hold, compared with adults of your age. The two overlap, because a large vocabulary helps with any passage, but they ask different questions and return different numbers.

The first difference is where the information comes from. In a verbal comprehension subtest the answer lives in your memory: you are asked what a word means, which word is nearest in meaning, or how two things are alike, and nothing on the screen gives it away. In the passage task the facts are on the screen, and a reader who brings outside knowledge in is penalized for it, because the correct answer depends only on the text. The page on the Verbal Comprehension Index explains the score the first kind of task feeds, and the page on the verbal IQ test explains why publishers stopped reporting a Verbal IQ at all.

The second difference is time. SHL's fact sheet gives 19 minutes for up to 30 questions, which works out, as our arithmetic and not the publisher's figure, to about 38 seconds per question if a candidate sees all 30 and spends it evenly, with the reading included. A comprehension knowledge subtest is paced differently, and ACIS describes its own Antonyms and Synonyms subtests as 31 words each with a 10 minute limit for the section, which is about 19 seconds per word (our arithmetic) but with a single word to read each time instead of a paragraph.

The third difference is the comparison group and the report. A verbal index from a normed battery is compared with a described norm group of adults. The SHL fact sheet does not say who the Verify Verbal Reasoning result is compared with or what the candidate sees. The Standards for Educational and Psychological Testing say in Standard 5.8 that norms, if used, "should refer to clearly described populations," and a candidate who is told only a percentile should ask against whom. Without that, the number cannot be placed on the scale.

The fourth difference is purpose. A verbal IQ score or index describes a person's ability for the person's own use or for a clinician or researcher. An employer's verbal reasoning score is a selection input, and an employer may set a cut score for a role. The page on vocabulary and its relation to g shows why word knowledge in particular tracks general ability closely; the passage task depends on word knowledge but is a looser measure of it.

4 Why Does the Cannot Say Answer Trip Careful Readers?

The third answer trips careful readers because it asks them to stop reasoning exactly where the passage stops, and careful readers are the ones most inclined to keep going. The SHL fact sheet frames the decision as one made from the information given, and that phrase carries the whole format. A statement can be true in the world, plausible to a specialist and still be a Cannot Say, because the passage does not establish it.

A gloss of the three answers helps, and it is ours, not the publisher's. True means the passage itself supports the statement. False means the passage itself rules the statement out. Cannot Say means the passage neither supports nor rules it out, so the statement may be right or wrong in the world and the page does not tell you. It is easy to treat False as "not supported" and Cannot Say as "unlikely", which merges two different answers and loses points on both.

The third statement is where the trap sits. Nothing in the passage contradicts it, so some readers mark False, and the idea that early members might get perks sounds plausible, so others mark True. The only answer the text supports is Cannot Say, and the error comes from filling a gap with a reasonable guess, not from weak language ability.

There are at least four reasons a careful reader falls into it, and they are our analysis of the format, not findings from a study. The first is outside knowledge. A recruiter or an accountant reading a passage about their own field already knows what usually happens, and the statement that matches what usually happens feels true. The Standards for Educational and Psychological Testing, published jointly by the American Educational Research Association, the American Psychological Association and the National Council on Measurement in Education, list "familiarity with the subject matter of the reading passages" among the possible sources of construct-irrelevant variance on a reading comprehension test, meaning score differences that do not come from the ability being measured.

The second reason is small words. "Most", "only", "all", "at least" and "unless" change what a statement claims, and a reader moving at speed registers the topic of a sentence and not its quantifier. A statement that says "all" is False the moment the passage mentions one exception, and Cannot Say if the passage never addresses the whole group. The third is the pull of the first two answers: True and False each need a sentence you can point to, so they feel more decisive than a hedge under a clock. Fatigue is a plausible fourth.

None of the publisher pages we read states how items are scored beyond multiple choice, whether wrong answers cost anything, or how often Cannot Say is the keyed answer, and a candidate should be wary of any prep claim that does. What the format does reward is a habit. For each statement, find the sentence that supports it or the sentence that rules it out. If you cannot point to one, you do not have True or False.

That habit is a reasoning discipline more than a language skill, and whether it is a separate ability from intelligence is a question the page on critical thinking versus intelligence takes up. A related employer format, in which the candidate evaluates the strength of an argument rather than the support for a statement, is covered on the page about the critical thinking test.

5 What Do the Clock and the Unsupervised Setting Add?

The 19 minute limit and the unsupervised setting turn a reading task into a reading task done at a pace, at home, once, and each of those conditions is stated by the publisher or the employer rather than guessed. SHL's fact sheet gives 19 minutes for up to 30 questions and says the item response theory design allows an unsupervised setting. For its adaptive Verify Verbal Ability test the figures are 15 minutes for up to 18 questions, about 50 seconds per question if all 18 are shown (our arithmetic, not the publisher's).

Whether the clock changes what is measured is a question the Standards address directly. Standard 4.14 says that for a test with a time limit, development research should examine the degree to which scores include a speed component and evaluate whether that component is appropriate for the domain. Its comment adds that where speed is not a meaningful part of the target construct, time limits should be set so that examinees have adequate time to demonstrate the targeted knowledge and skill. The SHL sheet does not report such an analysis, so a candidate cannot tell from it how much of the score is speed. A reasonable inference, and only an inference, is that a slow careful reader and a fast careful reader will not score alike on a 19 minute test even if their comprehension is equal.

Speed of processing is its own broad ability in the CHC model (Gs), and a passage test is not a speed test. But reading is the slowest part of the task, and reading speed is partly a function of practice, of language background and of the format on the screen. A candidate who wants to see their own speed of simple visual processing, separately from reading, can look at the processing speed test, which is a different ability and not a rehearsal for the employer's test.

The unsupervised setting changes what the candidate controls and what can go wrong. The SHL sheet says the test is designed for an unproctored environment and lists the sittings as one plus a verification test, which the text we read does not define. UBS's graduate page says its assessments must be completed within seven days of the invitation, by the applicant alone, with no retake. The candidate chooses the hour, the room and the device, and also bears the cost of a noisy room or a tired evening. The practical steps are plain: choose a quiet hour, close other windows, read the instructions slowly and complete the whole test in one sitting.

Standard 8.2 supports asking for the facts in advance. It says test takers should be given as much information about the test as is consistent with obtaining valid responses, including item formats, and should be made aware of any imposed time limits so they can manage their time. The standard does not tell any employer what to do, but asking for the format, the number of questions and the time allowed is consistent with it. The publisher pages we read say nothing about adjustments.

6 What Does English Proficiency Do to the Score?

A passage test in English measures English as well as reasoning, so a candidate who learned English as a second language should expect the score to understate the reasoning, and the research we opened supports the direction of that effect for vocabulary but not its size on this test. The Standards make the point in their discussion of content evidence: in an employment test, the use of vocabulary more complex than needed on the job "may be a source of construct-irrelevant variance for English language learners or others." That sentence is a statement about test design, not a measurement, but it is the professional bodies' own wording.

The measured evidence is about vocabulary, which is the part of the passage task that language background touches most. Portocarrero, Burright and Donovick (2007) compared college students who were bilingual, born abroad and speaking English, with monolingual English speakers. The bilinguals' English vocabulary was in the average range, yet they had lower receptive and expressive English vocabularies than their monolingual peers despite having enough English to attend a competitive university. The younger the bilingual students were when they arrived in the United States, the better their English vocabulary. Bialystok and Luk (2012) aggregated English receptive vocabulary scores from 797 monolingual and 808 bilingual participants aged 17 to 89 across 20 studies and found that monolinguals obtained significantly higher standard scores than bilinguals.

Those are vocabulary results, not results from job applicants on a timed passage test, and we opened no study that measured how English proficiency changes the result on SHL's test or on any other employer passage test. A school-age result points the same way: Cormier and colleagues (2022), using the Woodcock-Johnson IV normative sample of school-age students, found that language abilities had a significant influence on cognitive test performance, and that test directions did not once language abilities were accounted for. The sample is children, the test is not an employer screen, and the finding is about the examinee's language more than about the test.

Four practical points follow, and the middle two are inference. First, the evidence we found concerns vocabulary, which the passage task needs. Second, a candidate with lower English proficiency can be a strong reasoner and still run short of time, because reading in a second language is plausibly slower. Third, nothing in the publisher pages we read describes language adjustments, so a candidate who needs one should ask the employer rather than assume it exists. Fourth, a low result on this kind of test should be read with the language demand in mind and not as a statement about intelligence.

The ACIS technical manual makes the same caution for its own tests: it says verbal and knowledge tasks can be affected by English proficiency and educational exposure, and the technical manual is the place to read it in context. The page on IQ tests for non-native English speakers owns the wider language question. ACIS is in English only.

7 What Does Retest Research Show?

Retesting on a cognitive ability test raises scores on average, and one latent variable study found that the second score measured something different from the first, which is why a second attempt is not a clean second measurement. Hausknecht, Halpert, Di Paolo and Moriarty Gerrard (2007) pooled 50 studies of practice effects on cognitive ability tests, covering 107 samples and 134,436 participants, and reported an adjusted overall effect size of .26. Their abstract notes that previous studies indicated as many as 25 to 50 percent of applicants in organizational and educational settings are retested. Effects were larger when practice came with test coaching and when identical forms were used. We read the paper's abstract and not its full text, so we report no figure that the abstract does not contain, and we do not convert the .26 into score points.

The second study asks a different question: whether retesting changes what is measured. Lievens, Reeve and Heggestad (2007) studied 941 candidates who took a cognitive ability test in a high-stakes context. Using a latent variable approach they found no measurement bias at the first testing, but retesting induced both measurement bias and predictive bias. The factor underlying the retest scores was less saturated with g and more associated with memory than the factor underlying the first scores, and those changes eliminated the test's criterion-related validity. That is one sample in one setting, and the finding concerns a cognitive ability test, not specifically a verbal one, but it bears directly on what a retest score means.

Two readings of these results are worth keeping apart. One is that practice helps: familiarity with the format and the pace removes friction. The other is that some of the gain is memory and not reasoning, which the Lievens result points to. Hausknecht's finding that identical forms produce larger gains is consistent with the second reading, although the abstract does not test it.

For a candidate the useful points are narrow. Some employers do not allow a retake, as UBS's page shows with its single attempt inside the seven day window, so the first attempt is the only one. Where retakes exist, Standard 9.18 says test users should explain a test taker's opportunities, if any, to retake the examination and whether earlier and later scores will be reported, so asking which score the employer uses is reasonable. And a score that has improved on a retake means the candidate improved on that test under those conditions, which the research shows is not the same as having improved at verbal reasoning.

8 What Do Verbal Tests Predict at Work, and What Do They Not?

Verbal ability tests do predict job performance and training success on average, with the strongest support coming from meta-analyses of many jobs, but that average says nothing certain about any one candidate or any one test. The study to start from is the UK meta-analysis by Bertua, Anderson and Salgado (2005) in the Journal of Occupational and Organizational Psychology. According to its abstract, the authors built a database of 283 independent samples with job performance as the criterion (13,262 people) and 223 with training success as the criterion (75,311 people), coded by seven occupational groups and by type of specific ability test: verbal, numerical, perceptual and spatial. They report that general mental ability and the specific ability tests are valid predictors of both criteria, with operational validities in the magnitude of .5 to .6, and that the validities were moderated by occupational group, with higher values in occupational families of greater job complexity.

Three limits on that result matter to a candidate. First, we read the abstract and not the full paper, so we report no coefficient for verbal tests alone. Second, a correlation across thousands of workers describes a group and not a particular person, and it does not describe a particular publisher's product. Third, the result is about prediction, not about what the test measures. The page on IQ and job performance owns the wider meta-analytic evidence on general mental ability and work outcomes, and this page does not repeat it.

A later paper changes how those numbers should be read. Sackett, Zhang, Berry and Lievens (2022) argued in the Journal of Applied Psychology that the range restriction corrections used in earlier selection meta-analyses often substantially overcorrected, so that the validity of many selection procedures for predicting job performance was overestimated. Their revised mean validity estimates were reduced by .10 to .20 points, although most procedures that ranked high before stayed high in rank. We did not check whether the corrections in Bertua and colleagues fall under that critique, so the .5 to .6 should be read as the value reported at the time and not as a current estimate.

Whether verbal ability predicts differently from numerical ability is not answered by the abstract for the type of test in this article. The page on verbal versus quantitative ability examines how the two tilt in the population, and the sibling page on the numerical reasoning test covers the employer format for numbers, which often sits beside a verbal test in the same battery.

The Standards add a point about how an employer should use any such score. Standard 11.3 says that when test content is the primary source of validity evidence for employment decisions, a close link between test content and the job's requirements should be demonstrated. Standard 11.4 says that when several scores, or scores and non-test information, are combined into a decision, the role of each should be explained and supported by validity evidence. SHL's fact sheet says its content makes use of workplace experiences, but it presents no job analysis or validity coefficient in the text we read. That is a limit of the one page we opened, not a statement that no such evidence exists.

For a candidate the useful reading is modest: a verbal reasoning result is one input the employer weighs, usually with an interview, a record and often other tests, and it is neither a verdict on ability nor a measure of how you read at your desk.

9 How Should a Candidate Read a Verbal Reasoning Result?

A result on a short employer test is a position in a comparison group at one sitting, and reading it well means asking which group, how precise the number is and what the employer does with it. The first question is the group. The Standards say in Standard 5.8 that norms, if used, should refer to clearly described populations that include the groups with whom users will ordinarily wish to compare examinees. Employer reports vary: some give a percentile, some a band, some only an outcome. The SHL fact sheet states no norm group, so whether a percentile of 70 means "above 70 percent of adults", "above 70 percent of graduates" or "above 70 percent of applicants to this firm" is a question the candidate cannot answer from the sheet. The page on how IQ scores are normed shows why the same performance gives different numbers against different groups.

The second question is the scale. A percentile and a standard score are different numbers for the same position. On the usual standard scale for IQ and index scores the mean is 100 and the standard deviation is 15, a convention explained on the page about why standard deviation 15 matters. An employer test may or may not report on that scale, and percentiles from different tests cannot be averaged, a point developed on the page about IQ score versus percentile. If an employer gives only a pass or fail outcome, no conversion to either is possible.

The third question is precision. The Standards state in their chapter on reliability and precision that if an assessment is shortened, for example by decreasing the number of items, its reliability is likely to decrease, and that lengthening it with comparable items is an effective and commonly used way to improve it. A test of up to 30 questions in 19 minutes samples one narrow slice of performance, and any one score on it should be read as a band and not a point. The page on reliability and validity explains why a longer, broader battery is more stable than a screener.

The fourth question is what you will be told. Standard 9.16 says that unless circumstances clearly require that results be withheld, a test user is obligated to provide a timely report to the test taker. Its comment says feedback in the form of a score report or interpretation is not always provided when tests are used for personnel selection or promotion. A candidate can ask the employer or the recruiter whether feedback is given, and what it will contain. That request is reasonable, and the comment shows the practice of withholding feedback is not unusual.

The fifth is the one candidates most want and the research least supports: a number that says what you would score on the real test. Nothing in the sources we opened lets a practice or other test predict a score on an employer's test.

10 What Can Practice Questions, Answers and PDFs Honestly Do?

Practice material can show the format and the logic of the three answers, and it cannot tell you where you stand, because it has no connection to the employer's comparison group. The searches behind this page ask for a verbal reasoning test free, with answers, as a PDF and with examples, and the honest answer to each is about what the format allows. We do not reproduce passages or statements from any test or practice site, we do not link sellers of practice material, and our one example above is written by us and labelled as ours.

What a good practice set can do is narrow. It can show the layout, the number of statements per passage, how the timer behaves and what navigation exists. It can let the candidate try the discipline described above, locating the sentence that supports or rules out each statement, until it is automatic. The Standards support the idea that format information should be available: Standard 8.1 says information about test content and purposes available to any test taker prior to testing should be available to all test takers, and the comment notes that some organizations supplement this with additional training or coaching. Hausknecht and colleagues' meta-analysis found larger practice effects when practice came with coaching, which is consistent with format familiarity mattering.

What it cannot do is limited by what the publishers disclose. The SHL fact sheet gives no scoring rule beyond multiple choice and no norm group, so a practice score is a statement about accuracy on that set. It does not convert into a percentile in any employer's group, and it does not convert into an IQ.

Answers are worth more than the item set. An answer key that says only "True" tells you nothing. A useful key names the sentence in the passage that supports the answer or the sentence that rules it out, or says explicitly that no sentence does. If a key relies on what is generally known about the topic, it is teaching the wrong habit for this format. A PDF has the same strengths and weaknesses as a web page, with one added check: whether it comes from the test's publisher or the employer, which is the version most likely to match the live test, or from an unrelated source, which may not.

Free is a word with two meanings here. A free practice set costs nothing and says nothing about your ability, while a free IQ test measures general ability on adult norms; neither stands in for the other. Employers sometimes provide their own practice, and UBS's page points applicants to practice assessments on another company's website, which is the first place to look because the employer chose it.

11 How Can an Adult See Their Own Verbal Profile?

An adult who wants to know where their verbal abilities stand can take a normed battery of word and concept tasks and read the result by index, which answers a different question from the employer's passage test. The table sets the real options side by side. It holds only what we could read on a publisher's or employer's page on October 29, 2026, and where the source is silent the cell says so, because a guessed cost would be worse than a blank.

The WAIS-5 row comes from Pearson's product page, which calls it an individually administered clinical instrument, lists the ages 16:0 to 90:11 and gives qualification level C. The rows for practice sets and for ACIS are our description, with prices read on October 29, 2026 that can change. The page on where to take an IQ test compares every route in more detail, and the page on the IQ test for adults covers the adult norm case.

ACIS has no passage task, no true, false or cannot say statements, and no item that asks you to read a paragraph and judge a claim. Its verbal tasks work on words and concepts. The Vocabulary subtest shows 45 words one at a time and asks you to type each definition. The Similarities subtest gives 31 items asking how two things are alike, with a typed answer. The Antonyms subtest and the Synonyms subtest each present 31 words and ask for the best opposite or the best match from options, with a 10 minute limit for the section. The Information subtest asks 30 questions with typed answers, and the Complex Relations subtest asks for 30 analogies completed by typing words into two blanks. Each description comes from the subtest's own page, and where a page is silent, such as on reading load, so are we.

Five of them form the Verbal Comprehension Index (VCI): Antonyms, Vocabulary, Information, Synonyms and Similarities. Complex Relations belongs to the Fluid Reasoning Index (FRI) and is part of the Full Scale form, and the Custom Test lets a buyer choose it. The technical manual lists a composite called the VRI, which "focuses on verbal reasoning through Similarities and Complex Relations." A reader who wants only the verbal index can buy the five subtests through the Custom Test for 25 dollars, and custom pricing reaches 50 dollars at 10 subtests, the price of the Full Scale form. ACIS also offers a free trial with no card.

What the result shows is a profile, and the profile is where the information is. A Vocabulary result, produced by typing definitions, can differ from a Synonyms result, produced by recognizing the closest option, and the Similarities page notes that a weaker Similarities score than Vocabulary score can mean many words but more difficulty forming broad categories. The page on index score differences explains how large a gap between two indices is unusual. Scaled scores run from 1 to 19 with a mean of 10 and a standard deviation of 3, and the indices use the standard scale with a mean of 100 and a standard deviation of 15.

12 Where ACIS sits, and what a buyer should do

ACIS is a way to see your own verbal profile on adult norms, and it is not a verbal reasoning test, a preparation for one, a predictor of one or a substitute for one. We sell it, so read this section as a disclosure. Its limits are plain. It is online and unsupervised, which means the setting, the device and the effort are the taker's own. It is not a clinical instrument. It is English only, and the manual says verbal and knowledge tasks can be affected by English proficiency and educational exposure. And its research page states that ACIS "is not intended for clinical diagnosis, hiring decisions, accommodation determinations or admission to high IQ societies," which is the limit stated on the ACIS for research page. Nothing here is a reason to send an ACIS report to an employer, and we do not suggest it.

The Standards for Educational and Psychological Testing (AERA, APA and NCME, 2014) frame how any score should be read, and three points from them recur on this page. A score is valid for a stated use and not in general, so a verbal profile on adult norms supports a statement about your own verbal abilities and not a prediction of an employer test result. Norms should refer to clearly described populations, which is why a percentile on one test cannot be carried over to another. And a shortened assessment is likely to be less reliable, which is why one subtest should be read as part of a profile. The American Psychological Association co-publishes the Standards and also issues professional guidance on assessment; we could not open its pages on October 29, 2026, so we quote none of it. The ACIS technical manual describes how it handles scores, and no figure from it is repeated here.

What should a reader do next? The first step is not an ACIS purchase. Read the employer's invitation for the format, the number of questions, the time allowed and whether a retake exists. If you need a language adjustment, ask before the test and not after. Use the employer's own practice material if it provides some, and treat any practice score as a statement about practice. Keep your reasoning habit simple: for each statement, find the sentence that supports it or rules it out.

The second step depends on what you want to know. If the question is how your verbal abilities compare with adults of your age, and you want a profile you can read across the five verbal comprehension subtests, the Custom Test at 25 dollars returns the Verbal Comprehension Index, and the Optimized form at 30 dollars and the Full Scale form at 50 dollars add more domains, all one time payments read on October 29, 2026. The guarantee 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." If you only want to see the format and the report before paying, start the free trial, which needs no card. The ACIS result is for you. It does not change what the employer's test will ask.

13 Sources Behind This Page

Every figure above is traceable to one of the sources below, and each is linked at the point where it is used. Publisher, employer and product pages were opened on October 29, 2026 and can change. Where a figure is our arithmetic on a published number, the text says so. We read the abstracts, not the full text, of the journal articles that sit behind paywalls (Bertua and colleagues; Hausknecht and colleagues; Lievens and colleagues; Sackett and colleagues).

• SHL. Verify Verbal Reasoning Test, Assessment Fact Sheet. SHL, 2018 copyright line, two pages. Read October 29, 2026.
• SHL. Verify Verbal Ability, Assessment Fact Sheet. SHL, 2018 copyright line, two pages. Read October 29, 2026.
• Talogy. Logiks Assessment, cognitive ability test. Product page, read October 29, 2026.
• UBS. Graduate Talent Program. Careers page, read October 29, 2026.
• Pearson Assessments. Wechsler Adult Intelligence Scale, Fifth Edition (WAIS-5). Product page, read October 29, 2026.
• Bertua C, Anderson N and Salgado J. The predictive validity of cognitive ability tests: A UK meta-analysis. Journal of Occupational and Organizational Psychology, 2005, volume 78, issue 3, pages 387 to 409.
• Hausknecht J, Halpert J, Di Paolo N and Moriarty Gerrard M. Retesting in selection: A meta-analysis of coaching and practice effects for tests of cognitive ability. Journal of Applied Psychology, 2007, volume 92, issue 2, pages 373 to 385.
• Lievens F, Reeve C and Heggestad E. An examination of psychometric bias due to retesting on cognitive ability tests in selection settings. Journal of Applied Psychology, 2007, volume 92, issue 6, pages 1672 to 1682.
• Portocarrero J, Burright R and Donovick P. Vocabulary and verbal fluency of bilingual and monolingual college students. Archives of Clinical Neuropsychology, 2007, volume 22, issue 3, pages 415 to 422.
• Bialystok E and Luk G. Receptive vocabulary differences in monolingual and bilingual adults. Bilingualism: Language and Cognition, 2012, volume 15, issue 2, pages 397 to 401.
• Cormier D, Bulut O, McGrew K and Kennedy K. Linguistic influences on cognitive test performance: Examinee characteristics are more important than test characteristics. Journal of Intelligence, 2022, volume 10, issue 1, article 8.
• McGrew K. Carroll's three-stratum (3S) cognitive ability theory at 30 years: Impact, 3S-CHC theory clarification, structural replication, and cognitive-achievement psychometric network analysis extension. Journal of Intelligence, 2023, volume 11, issue 2, article 32.
• Sackett P, Zhang C, Berry C and Lievens F. Revisiting meta-analytic estimates of validity in personnel selection: Addressing systematic overcorrection for restriction of range. Journal of Applied Psychology, 2022, volume 107, issue 11, pages 2040 to 2068.
• American Educational Research Association, American Psychological Association and National Council on Measurement in Education. Standards for Educational and Psychological Testing. AERA, 2014. Open access edition, read October 29, 2026; Standards 4.14, 5.8, 8.1, 8.2, 9.16, 9.18, 11.3 and 11.4, the chapter on reliability and precision, and the discussion of content evidence in chapter 1.

14 Frequently Asked Questions