A working memory test for adults: what it measures, and how to take a normed one

0 Quick Answer

A working memory test measures how much information you can hold and manipulate for a few seconds, it is scored as a span against age norms, and the number it produces is only interpretable with its standard error next to it. The standard tasks are digit span, in which digits are read or shown one at a time and repeated forward, backward or re-ordered; letter and number sequencing, in which mixed items are re-ordered by rule; and visual span, in which a spatial sequence is reproduced. The Wechsler Adult Intelligence Scale, Fifth Edition, published by Pearson in 2024, measures working memory in what its product page calls auditory, visual and spatial forms, with a task that mimics the real world rate of speech. ACIS measures it with three subtests, Digit Span, Alphanumeric Sequencing and Visual Sequence, and reports a Working Memory Index with a composite reliability of .9247 and a standard error of 4.12 points, per the technical manual.

The construct is well defined. Baddeley and Hitch proposed it in 1974 as a system with limited capacity that both stores and works on information, distinct from a passive short term store. The capacity is small: Cowan's 2001 review put it at about four chunks when rehearsal and grouping are prevented, against the seven plus or minus two of Miller's 1956 paper, which measured something closer to immediate recall.

What a working memory test does not do is diagnose anything. Adults with attention deficit hyperactivity disorder show working memory deficits on average, as documented below, but a low span is not a diagnosis, and a normal span does not rule one out. The page on ADHD and IQ covers that boundary in detail.

1 What Working Memory Is, and What It Is Not

Working memory is the small amount of information you can keep active and manipulate right now, and it is not the same thing as remembering a phone number or recalling last week. The distinction was drawn in 1974 by Alan Baddeley and Graham Hitch in a chapter of The Psychology of Learning and Motivation, volume 8, pages 47 to 89. They replaced the idea of a single short term store with a system of components: a central executive that directs attention, a phonological loop that holds verbal material, and a visuospatial sketchpad that holds visual and spatial material. Baddeley added a fourth component, the episodic buffer, in Trends in Cognitive Sciences in 2000, volume 4, issue 11, pages 417 to 423, to account for the binding of information across the other stores.

The model matters for testing because it predicts what a good test battery should contain. A verbal span task loads the phonological loop. A spatial span task loads the sketchpad. A task that requires re-ordering loads the central executive on top of storage. A battery that uses only forward digit span measures storage and little else, and a battery that uses only re-ordering tasks cannot separate the executive component from the store it operates on. The page on cognitive domains places working memory among the broad abilities of the Cattell Horn Carroll model, where it is labelled Gwm, and explains why the broad abilities are kept separate in scoring.

The capacity of the system is smaller than most people assume. George Miller's 1956 paper in Psychological Review, volume 63, issue 2, pages 81 to 97, gave the famous figure of seven plus or minus two, but that figure describes immediate serial recall of familiar items that can be rehearsed and grouped. Nelson Cowan's 2001 target article in Behavioral and Brain Sciences, volume 24, issue 1, pages 87 to 114, reviewed studies in which rehearsal and chunking were prevented and concluded that the capacity of the focus of attention is about four chunks, with a range of three to five in adults. A digit span of seven is therefore not evidence of a seven item working memory. It is evidence that the person can rehearse and group digits, which is itself part of what the test measures.

None of this is long term memory. Recalling a fact learned years ago is a retrieval from a different system, measured on ACIS by subtests such as Information and covered on the page on IQ and memory. The everyday complaints that bring adults to a working memory test, losing the thread of a conversation, forgetting why they walked into a room, re-reading a paragraph three times, are plausibly working memory complaints, but they are also fatigue, anxiety and attention complaints, and a test measures capacity under standardized conditions rather than the complaint.

2 How Psychologists Test It

Every professional working memory test is a span task with a discontinue rule, and the differences between instruments are in which spans are used and how the score is normed. The Wechsler adult scale is the reference point. Pearson's WAIS-5 page, read on September 19, 2026, describes comprehensive measurement of working memory across auditory, visual and spatial material, capacity and focus of attention, and lists among the new subtests Running Digits, in which digits arrive at a rate that mimics real speech and the examinee reports the last several, Symbol Span, a visual recognition span, and Spatial Addition, a spatial working memory task. The page on what the WAIS-5 is walks the index structure, and the page on the adult IQ testing process describes what an administration looks like in practice.

The mechanics are simple and fixed. In a digit span task, the examiner presents digits one per second and the examinee repeats them. Forward span measures storage. Backward span requires reversal, and sequencing span requires re-ordering into ascending order; both add manipulation on top of storage. Each condition begins with short sequences and lengthens by one item at a time, and each stops after a fixed number of consecutive failures at a length, which is the discontinue rule. The raw score is the number of sequences reproduced correctly, or in some instruments the longest span achieved. The raw score is then converted, against the norm table for the examinee's age, to a scaled score with a mean of 10 and a standard deviation of 3, and the scaled scores of the working memory subtests are summed and converted to an index with a mean of 100 and a standard deviation of 15. The page on how IQ is calculated follows that chain for any subtest.

ACIS uses the same architecture, and its technical manual documents each task. Digit Span runs forward from 2 to 12 digits, backward from 2 to 10 and sequencing from 2 to 11, and ends a part after two consecutive errors. Alphanumeric Sequencing presents intermixed numbers and letters and asks for them back with numbers ascending and then letters in alphabetical order, ending after three consecutive errors; the manual describes it as active auditory verbal working memory rather than passive storage, because two ordering rules must be applied while the series is held. Visual Sequence shows tiles that flash in sequence and asks for the sequence back by clicking, forward from 2 to 11 and backward from 2 to 10, ending a part after two consecutive errors; the manual is explicit that this task belongs to the Working Memory Index and not to the Visual Spatial Index, because its defining operation is temporary storage and ordered reproduction rather than reasoning about space. The page on subtest types describes the item formats across the battery.

The three ACIS tasks map onto the Baddeley components in the intended way. Digit Span forward is the phonological loop. Alphanumeric Sequencing is the central executive operating on verbal material. Visual Sequence is the sketchpad with an executive load in its backward condition. A person can be strong on one and weak on another, and the value of three tasks over one is that the pattern becomes visible.

3 From a Span to a Score: What the Numbers Mean

A raw span is a count, a scaled score is a rank among people your age, and only the second one carries information, because a span of six means different things at 20 and at 75. The conversion is the whole point of norming. The page on how IQ scores are normed explains how a reference sample turns raw counts into standard scores, and the page on the 15 point standard deviation explains the scale. On the subtest scale, a scaled score of 10 is the median for the age band, 13 is one standard deviation above it, and 7 is one below. On the index scale, 100 is the median, 115 one standard deviation above and 85 one below, and the IQ percentile calculator converts any point on either scale into a percentile.

Every score carries an error, and the error is larger for a single subtest than for an index built from three. ACIS reports the reliability and the standard error of measurement of each working memory subtest in its technical manual. Digit Span has a test retest reliability of .890 and a standard error of 0.99 scaled score points. Alphanumeric Sequencing has a reliability of .800 and a standard error of 1.34. Visual Sequence has a reliability of .820 and a standard error of 1.27. The manual uses test retest reliability rather than internal consistency for these tasks, because the discontinue rule makes item level consistency the wrong statistic for a span. The Working Memory Index that combines them has a composite reliability of .9247 and a standard error of 4.12 points on the IQ scale, which means a reported 100 is best read as a band from about 92 to 108 at 95 percent confidence. The page on reliability and validity explains why the index is more precise than any of its parts.

Two narrower composites exist for readers who need the modality split. The Auditory Working Memory Index combines Digit Span and Alphanumeric Sequencing, with a composite reliability of .9054 and a standard error of 4.61. The Visual Working Memory Index is a single indicator score built on Visual Sequence, with a reliability of .8200 and a standard error of 6.36, and the manual attaches a single score caution to it for that reason. A reader who sees a 20 point gap between the two should first check whether the gap exceeds the two standard errors added together before reading anything into it.

The table below shows how the standard error changes what a score can support, using the three ACIS working memory scores and the standard formula, standard deviation times the square root of one minus reliability. The figures are the manual's; the intervals are our arithmetic on them.

4 Working Memory and Intelligence: How Strong the Link Is

Working memory is one of the strongest correlates of fluid reasoning ever measured, and the two are still not the same thing, which is why both are scored. Engle, Tuholski, Laughlin and Conway showed in 1999, in the Journal of Experimental Psychology: General, volume 128, issue 3, pages 309 to 331, that a latent working memory factor predicted fluid intelligence while a latent short term memory factor did not once working memory was accounted for. The result established that the executive part of the system, not the storage part, carries the relationship with reasoning.

How strong the relationship is was argued in a single issue of Psychological Bulletin in 2005. Ackerman, Beier and Boyle meta-analysed the literature, volume 131, issue 1, pages 30 to 60, and concluded that working memory and general intelligence were related at about .5 on average and were not the same construct. Kane, Hambrick and Conway replied in the same issue, pages 66 to 71, that at the level of latent variables, where measurement error is removed, the correlation between working memory capacity and fluid intelligence across the available studies was about .7. Both sides agreed on the conclusion that matters for a test taker: the two constructs overlap heavily and are distinct.

ACIS's own structure reflects that finding. In the confirmatory factor model reported in the technical manual, fitted on 2,750 complete records, the Working Memory Index loads .788 on the higher order general factor, below the reasoning and knowledge domains and well above processing speed, which loads .648. The Working Memory Index correlates .699 with the Fluid Reasoning Index and .475 with the Processing Speed Index. The manual describes working memory and processing speed together as a cognitive proficiency cluster that remains clearly distinct from reasoning, which is the same distinction the WAIS makes when it reports a Cognitive Proficiency Index alongside the General Ability Index. The page on the General Ability Index explains why some readers want a reasoning score with working memory and speed removed, and the page on the g factor explains what the general factor is.

For the person reading their own report, the practical consequence is that a working memory score is expected to sit somewhere near the fluid reasoning score, and a large gap in either direction is the informative finding. A working memory index far below reasoning is the pattern the clinical literature associates with attention disorders and with fatigue and anxiety at the time of testing. A working memory index far above reasoning is less common and usually reflects a specific strength in serial storage. Neither pattern is a diagnosis, and the next section explains why.

5 Working Memory and ADHD in Adults

Adults with ADHD score lower on working memory tasks on average, the deficit is largest on tasks that load the central executive, and none of that turns a span score into a diagnostic test. The children's literature was synthesized by Martinussen, Hayden, Hogg-Johnson and Tannock in the Journal of the American Academy of Child and Adolescent Psychiatry in 2005, volume 44, issue 4, pages 377 to 384, in a meta-analysis of 26 studies that found moderate to large deficits, larger for spatial than for verbal working memory and larger for tasks requiring manipulation than for simple storage. Later syntheses of the children's literature found that the size of the deficit depended on task design, with more demanding tasks reporting larger effects.

The adult evidence is more limited and points the same way. Alderson, Kasper, Hudec and Patros tested adults with and without ADHD on tasks designed to separate the central executive from storage and rehearsal, and reported in the Journal of Abnormal Psychology in 2013, volume 122, issue 2, pages 532 to 541, that the adults with ADHD showed working memory deficits that were most consistent for the central executive component. In test terms, the re-ordering tasks, backward and sequencing span and letter number sequencing, are where the group difference is expected to appear, and forward span is where it is expected to be smallest.

What follows for a person taking a working memory test with ADHD in mind is a set of limits. A group difference is a statement about averages, and the distributions of adults with and without a diagnosis overlap substantially; a span score cannot sort an individual into either group. A low score has many causes other than attention, including poor sleep, anxiety at the time of testing, a second language, and hearing or screen difficulties in the task itself. And a diagnosis of ADHD is made on history and functioning across settings, not on any cognitive score. The page on ADHD and IQ reviews the wider evidence, and the page on neuropsychological testing for adults explains what a clinical evaluation adds that a cognitive battery cannot.

What a normed test can do is describe the profile precisely. A person who scores in the average range on reasoning and knowledge and a standard deviation lower on working memory and processing speed has a pattern worth taking to a clinician, with the numbers and their standard errors in hand. That is a better starting point for an evaluation than a complaint, and it is the honest use of a test like this in that context.

6 Working Memory Across the Adult Years

Working memory declines across adulthood in a nearly straight line from the twenties, and that is why an adult working memory score has to be read against people of the same age. Park and colleagues measured speed, working memory, long term memory and verbal knowledge in 345 adults from 20 to 92 and published the result in Psychology and Aging in 2002, volume 17, issue 2, pages 299 to 320. Working memory, like speed and long term recall, declined linearly across the decades, while verbal knowledge held steady or rose into the seventies. The pattern is the reason a raw span cannot be compared across ages and the reason an instrument's norm frame has to cover the ages it is sold to.

The ACIS adult reference frame covers ages 16 to 90 in age bands, with 3,243 records, and the technical manual states that no out of range case is interpreted in the adult frame. A working memory scaled score of 10 at 70 therefore means median performance for people in that band, and a raw span that would score 8 at 25 may score 11 at 70. The page on whether IQ changes with age works through what that means for composite scores, and the page on processing speed covers the other proficiency ability with the same age profile.

Two consequences follow for anyone comparing scores. A working memory result from one instrument cannot be compared to a result from another unless both were normed on the same kind of sample, because a 10 is a rank within a sample and the samples differ. And a result obtained years ago cannot be compared to a result today without allowing for the age norm, which is why a report states the age band it used. A person who scored a 12 at 30 and a 12 at 55 has held their rank while their raw span has almost certainly fallen, which is the expected and reassuring outcome.

7 What Free Online Working Memory Tests Measure, and What They Leave Out

A free online span test measures your span under uncontrolled conditions and reports it against whoever else took the test, and neither half of that produces a score you can interpret. The typical format shows digits or tiles, lengthens the sequence until you fail, and reports a number, sometimes with a percentile computed from the site's previous visitors and sometimes with a leaderboard. The International Test Commission's guidelines on internet delivered testing, published in the International Journal of Testing in 2006, volume 6, issue 2, pages 143 to 171, classify that administration as open mode: no identification, no supervision and no control of the conditions. The guidelines are explicit that open mode scores support weaker interpretations than controlled or supervised ones, and the page on whether online IQ tests are accurate applies the distinction across the market.

The three specific gaps are norms, reliability and conditions. A percentile against previous visitors compares you to people who chose to test their memory on that site, took it in whatever state and environment they were in, and possibly took it several times; it is not a percentile against adults of your age in a defined population. A single run with no reliability figure gives no basis for the precision the number implies, and the span tasks are exactly the tasks where a single run varies most, because one lapse of attention ends the sequence. And a task that shows the sequence at a speed the user controls, or lets the user replay it, is not the same task as one presented at a fixed rate, which is why the professional instruments fix the rate at about one item per second.

The free tests are not useless. They show what the task feels like, and a person who has never done a backward span learns something from trying one. The page on free IQ tests sets out what an unnormed score can and cannot support, and the page on free versus validated IQ tests gives the checklist. The checklist for a working memory test specifically is short: a fixed presentation rate, a stated discontinue rule, a norm table by age with its sample described, a reliability figure for the score, and a standard error printed on the report. A test that has the five is a measurement. A test that has none of them is a game, and the difference is visible before you start.

8 Can a Working Memory Score Be Trained Upward?

The evidence says that practicing working memory tasks improves working memory tasks and does not improve reasoning, and a score raised by practice is a score you can no longer interpret. The claim that it could came from Jaeggi, Buschkuehl, Jonides and Perrig, who reported in the Proceedings of the National Academy of Sciences in 2008, volume 105, issue 19, pages 6829 to 6833, that training on a dual n-back task raised scores on a fluid reasoning test in proportion to the amount of training. The paper launched an industry. The meta-analyses that followed did not support it. Melby-Lervåg and Hulme synthesized 23 studies in Developmental Psychology in 2013, volume 49, issue 2, pages 270 to 291, and found reliable short term gains on the trained tasks and on closely similar ones, with no convincing evidence of transfer to verbal ability, nonverbal ability, reading or arithmetic. Melby-Lervåg, Redick and Hulme extended the analysis in Perspectives on Psychological Science in 2016, volume 11, issue 4, pages 512 to 534, to 87 publications and reached the same conclusion, with the strongest designs showing the smallest far transfer.

For a person taking a test, the consequence is direct. Practicing span tasks before a working memory test raises the span through familiarity and strategy, and the norm table was built from people who had not practiced, so the resulting score no longer means what the table says it means. The page on how to prepare for an IQ test separates preparation that preserves the meaning of a score, such as sleep, a quiet room and understanding the instructions, from preparation that destroys it. The page on whether you can improve your IQ covers the wider literature on training and transfer.

The legitimate reading of the training literature is about strategy rather than capacity. Chunking, rehearsal and grouping raise measured span without changing the underlying capacity that Cowan estimated, which is part of why a digit span of seven and a capacity of four are both true at once. A test that fixes the presentation rate and uses re-ordering conditions limits how much strategy can add, which is one reason the professional instruments are built the way they are.

9 Taking a Normed Working Memory Test Online

A normed working memory test online is possible, and the way to recognize one is that it publishes the same three things a clinical instrument does: its norm frame, its reliability and its error. ACIS administers the three working memory subtests described above inside a timed, self administered battery. Each subtest is scored against the adult reference frame for the taker's age band and reported as a scaled score with its standard error. A person who wants only working memory can use the Custom form, which lets a buyer select subtests and is priced per subtest on the order page; completing the three working memory subtests returns the Working Memory Index and its two narrower composites. The Quick form, 15 dollars for six subtests including Digit Span and Alphanumeric Sequencing, adds verbal and fluid reasoning for context. The Optimized and Full Scale forms, 30 and 50 dollars, report the full profile in which the working memory result can be compared with the other domains, which is the reading the earlier sections argued for. Five subtests can be taken free without a card, and access to a purchased form stays open for 30 days.

What the report shows is the index, its percentile, its confidence interval and the three subtest scaled scores, alongside whatever other indices the form includes. The technical manual's interpretation rules apply: the manual states that single subtests and one indicator composites should be described with their standard error rather than as exact scores, and that the Visual Working Memory Index in particular carries a single score caution. A reader who follows those rules will treat a Digit Span of 9 with a standard error of 0.99 as a band from 7 to 11, and a Working Memory Index of 96 as a band from about 88 to 104, and will look at the pattern across the three tasks before drawing any conclusion.

What the test is not, stated plainly: it is not a clinical evaluation, it is not accepted as one by any institution, and it is self administered rather than supervised, which the manual addresses through completion requirements, retake control and integrity screening rather than through a proctor. The manual states that ACIS is not designed to treat repeated practice attempts as equivalent to first valid completions, which is a safeguard for the norm frame and a limit for anyone hoping to retest until the number improves. The page on IQ tests for adults and the page on accurate IQ tests place the battery among the alternatives.

10 How to Read a Working Memory Result

Read the index against the rest of the profile, read each subtest against the others, and read every number as a band, in that order. The first comparison is between the Working Memory Index and the reasoning indices. The manual's correlations, .699 with fluid reasoning and .704 with quantitative reasoning, mean that the scores are expected to be similar and that a difference larger than the combined standard errors is the finding. A working memory index 15 or more points below fluid reasoning, with both intervals not overlapping, is the pattern worth attention; a difference of 5 points is noise.

The second comparison is inside the index. Forward digit span measures storage and rehearsal. Sequencing and Alphanumeric Sequencing measure the executive component. Visual Sequence measures the spatial store. A profile in which the auditory composite is average and the visual composite is low, or the reverse, is more informative than the index alone, with the caution that the visual composite rests on one subtest and carries a standard error of 6.36 points. A working memory weakness pulls the Full Scale IQ down by less than it pulls the index down, because the composite averages over twenty subtests.

The third comparison is with the conditions. The manual records testing conditions and flags quality issues such as unusually fast responses or high non response, and a working memory result obtained while tired, interrupted or in a second language is a result about that session. A reader who suspects the session was compromised should treat the score as a lower bound and, if the question matters, retake after the interval the retake policy allows rather than immediately. The page on IQ and academic achievement shows where working memory sits among the predictors of learning outcomes, which is the context in which most adults want the number in the first place.

11 When to Seek a Clinical Evaluation Instead

A normed online test answers the question of where your working memory stands; a clinical evaluation answers the question of why, and the second question needs a person who can ask about your history. Three situations call for the clinical route. The first is a suspected diagnosis, because ADHD, learning disorders and acquired conditions are diagnosed on history and functioning, and a cognitive profile is one input among several. The second is a documented need, because schools, employers' accommodation processes and disability boards accept a signed evaluation and nothing else, as the page on official IQ tests sets out. The third is a change over time, because a fall in working memory in later adulthood relative to earlier ability is a clinical question, not a psychometric one.

A clinical working memory assessment is usually part of a larger evaluation. The WAIS-5 working memory subtests will be administered alongside the rest of the scale, often with the Wechsler Memory Scale, Fifth Edition, with which Pearson's page says the WAIS-5 was co-normed, and with attention, executive and personality measures as the referral question requires. The page on who can administer an IQ test explains the licensing behind such an evaluation, and the page on how much an IQ test costs prices it.

The two routes are not in competition. A normed online profile, taken first, tells a person whether the pattern they suspect is there and gives a clinician numbers with standard errors to start from. It costs a fraction of an evaluation and takes an hour. What it cannot do is interpret the pattern in the light of a life, and nobody should pay for a test expecting it to.

12 Where ACIS Sits, and What a Buyer Should Do

We sell the working memory subtests described on this page, so treat this section as a disclosure and check it against the rest of the page. ACIS measures working memory with three subtests, Digit Span, Alphanumeric Sequencing and Visual Sequence, scored against an adult reference frame of 3,243 records in age bands from 16 to 90, with test retest reliabilities of .890, .800 and .820 and standard errors of 0.99, 1.34 and 1.27 scaled score points respectively. The Working Memory Index has a composite reliability of .9247 and a standard error of 4.12 points on the IQ scale, and it loads .788 on the general factor in a confirmatory model fitted to 2,750 complete records with a comparative fit index of .9761. Every one of those figures is printed in the technical manual, version 1.4, updated August 3, 2026, and the manual states which reliability coefficient produced each standard error.

The prices are one time: 15 dollars for Quick, 30 for Optimized, 50 for Full Scale, and a per subtest price for the Custom form shown on the order page. Five subtests are free without a card. Purchased access stays open for 30 days. The quality guarantee is a full refund within five days if the report does not deliver the features described at checkout or a technical issue prevents access, and it does not cover disliking the result.

The limits are real. The test is self administered and not supervised, so its scores depend on integrity controls rather than a proctor. It is normed on English speaking adults, and a person tested in a language they do not use daily is outside the frame. The Visual Working Memory Index rests on one subtest and carries a standard error of 6.36 points, which the manual flags. And a working memory score, from ACIS or from anyone, is not a diagnosis of anything.

What a buyer should do depends on the question. If the question is whether working memory is a relative weakness or strength in your profile, take the free subtests, then buy the form whose indices give you the comparison you need, and read every score with its band. If the question is whether you have ADHD or another condition, take the profile to a clinician rather than to a search engine, and treat the numbers as a starting point. If the question is whether a two minute online game measured anything, the answer is that it measured your span on that game under those conditions, and this page has explained why that is not the same thing.

13 Sources Behind This Page

Every figure above is traceable to one of the following, and each is linked at the point where it is used. The 95 percent intervals in the table are our arithmetic on the reliabilities published in the ACIS technical manual, using the standard formula, and are labelled as such. The WAIS-5 description was read on Pearson's product page on September 19, 2026.

• Baddeley A D and Hitch G. Working memory. The Psychology of Learning and Motivation, 1974, volume 8, pages 47 to 89.
• Baddeley A. The episodic buffer: a new component of working memory? Trends in Cognitive Sciences, 2000, volume 4, issue 11, pages 417 to 423.
• Miller G A. The magical number seven, plus or minus two: Some limits on our capacity for processing information. Psychological Review, 1956, volume 63, issue 2, pages 81 to 97.
• Cowan N. The magical number 4 in short-term memory: A reconsideration of mental storage capacity. Behavioral and Brain Sciences, 2001, volume 24, issue 1, pages 87 to 114.
• Engle R W, Tuholski S W, Laughlin J E and Conway A R A. Working memory, short-term memory, and general fluid intelligence: A latent-variable approach. Journal of Experimental Psychology: General, 1999, volume 128, issue 3, pages 309 to 331.
• Ackerman P L, Beier M E and Boyle M O. Working memory and intelligence: The same or different constructs? Psychological Bulletin, 2005, volume 131, issue 1, pages 30 to 60.
• Kane M J, Hambrick D Z and Conway A R A. Working memory capacity and fluid intelligence are strongly related constructs: Comment on Ackerman, Beier, and Boyle (2005). Psychological Bulletin, 2005, volume 131, issue 1, pages 66 to 71.
• Martinussen R, Hayden J, Hogg-Johnson S and Tannock R. A meta-analysis of working memory impairments in children with attention-deficit/hyperactivity disorder. Journal of the American Academy of Child and Adolescent Psychiatry, 2005, volume 44, issue 4, pages 377 to 384.
• Alderson R M, Kasper L J, Hudec K L and Patros C H G. Working memory deficits in adults with attention-deficit/hyperactivity disorder (ADHD): An examination of central executive and storage/rehearsal processes. Journal of Abnormal Psychology, 2013, volume 122, issue 2, pages 532 to 541.
• Park D C and colleagues. Models of visuospatial and verbal memory across the adult life span. Psychology and Aging, 2002, volume 17, issue 2, pages 299 to 320.
• Jaeggi S M, Buschkuehl M, Jonides J and Perrig W J. Improving fluid intelligence with training on working memory. Proceedings of the National Academy of Sciences, 2008, volume 105, issue 19, pages 6829 to 6833.
• Melby-Lervåg M and Hulme C. Is working memory training effective? A meta-analytic review. Developmental Psychology, 2013, volume 49, issue 2, pages 270 to 291.
• Melby-Lervåg M, Redick T S and Hulme C. Working memory training does not improve performance on measures of intelligence or other measures of far transfer: Evidence from a meta-analytic review. Perspectives on Psychological Science, 2016, volume 11, issue 4, pages 512 to 534.
• International Test Commission. International Guidelines on Computer-Based and Internet-Delivered Testing. International Journal of Testing, 2006, volume 6, issue 2, pages 143 to 171.
• Pearson. Wechsler Adult Intelligence Scale, Fifth Edition (WAIS-5), product page. pearsonassessments.com, read September 19, 2026.
• ACIS. Technical manual, version 1.4, updated August 3, 2026: subtest technical table, reliability tables and SEM interpretation, confirmatory factor analysis output. acisiq.com/technical-manual.

14 Frequently Asked Questions