The IQ test on a phone: what holds up and what slips on a small touch screen

0 The short answer

A normed IQ test can be taken on a phone, but untimed reasoning and vocabulary tasks hold up better than tasks scored by how fast you respond, because the phone adds delay and a small target of its own to every response. A test of 26 smartphones measured a total device delay of 35 to 140 milliseconds (Nicosia and colleagues, 2023), and a web study of 59,587 visitors found phone users slower on speed tasks even after adjusting for age, gender, education and vocabulary (Passell and colleagues, 2021). ACIS is one normed option, at 15, 30 or 50 dollars for about 45, 110 or 175 minutes (prices read October 29, 2026), and its own FAQ recommends a larger screen for visual and timed subtests. No study we found shows that scores from a phone and a computer are equivalent, and we do not claim it.

1 What Counts as a Real IQ Test on a Phone?

An IQ test on a phone is real only when it is a battery of ability tasks scored against a stated adult norm group, and many products that are easy to open on a phone fail that test before the device matters. A psychologist giving a Wechsler battery, a free quiz in a browser, a game that returns a number and a normed online battery are all described as IQ tests by someone, and the phone changes none of the questions about what each one measures and against whom it is scored. The page on IQ test apps reads the app listings one by one, with their prices and the norms they name or leave unnamed, so this page keeps to the question the app page does not answer: what the phone itself does to a test that is otherwise sound.

The comparison below sets the real options side by side. The cost column carries only what we opened on October 29, 2026, and where we found no published price we say so. Pearson lists the WAIS-5 for ages 16:0 to 90:11, a completion time of 45 minutes for the 7 subtest Full Scale IQ and 60 minutes for the 10 primary index subtests, and the administration option "Paper or Q-interactive", the latter described on the page as a one to one iPad based system (Pearson product page, read October 29, 2026). That is a tablet in a psychologist's hands with a trained examiner watching, which is a different situation from a personal phone on a kitchen table, and we state it as our reading of the listing and not as a claim Pearson makes.

A free quiz is fine for curiosity, and the page on free IQ tests explains what a free route includes and what it leaves out. The page on the best online IQ tests compares the normed online batteries, the page on where to take an IQ test lists the in person and online routes together, and the page on whether online IQ tests are accurate asks whether any online result can be trusted. The rest of this page assumes the test itself is sound and asks what the phone adds.

2 What Does Research Say About Device Effects on Cognitive Test Scores?

Research shows that the device a person uses can move cognitive test scores, that the movement is largest for tasks scored in milliseconds, and that part of every gap comes from who chooses each device and not from the device alone. The field has been asking the question for decades. Mead and Drasgow (1993), in a meta-analysis in Psychological Bulletin, examined 159 correlations between paper and computer versions of cognitive ability tests, 123 from timed power tests and 36 from speeded tests. The corrected cross-mode correlation was .91 overall, .97 for timed power tests and only .72 for speeded tests. That study compared paper with computers and not phones, but it set the pattern that later work on phones repeats: when speed is the point, the medium matters more.

For personal devices, Germine, Reinecke and Chaytor (2019) give the field's overview in The Clinical Neuropsychologist. They name three barriers that affect both the interpretation of test scores and test norms. The first is variability in the perceptual, motor and cognitive demands of the same test across device classes such as personal computers, tablets and smartphones. The second is hardware and software variability between devices within the same class. The third is rapid change over time in hardware, software and device ownership, which can make particular tests and norms obsolete. Each is a reason a phone score may not mean what a computer score from the same person would mean.

The largest direct study is Passell and colleagues (2021) in Behavior Research Methods. The authors examined differences in measured performance on timed and untimed cognitive tests between users of common digital devices among 59,587 visitors in Study 1 and 3,818 visitors in Study 2 to TestMyBrain.org, a citizen science website. Study 1 used data collected from 2014 to 2019 from participants aged 18 to 80 who reported English as their native language, on three tasks: a simple reaction time task, a digit symbol matching task and a 20 item vocabulary test. Controlling for age, gender, educational background and performance on the untimed vocabulary test, the authors report that users of mobile devices, particularly Android smartphones, showed significantly slower performance on tests of reaction time than users of laptop and desktop computers. They also report that devices differing in user interface, such as screen size or mouse against touchscreen, showed significant differences on tests requiring fast reactions or fine motor movements.

Two limits keep that result in proportion, and the authors state both. First, the device groups differ in who uses them: in the Study 1 sample, iPad users were older on average (45.60 years) than Android (33.23) or iPhone (31.99) users, and Android users reported the fewest years of education. The vocabulary test, which is not scored by speed, also differed by device even after controls, which the authors read as a sign that some of the gap comes from other differences between the people who pick each device. Second, they note that devices change constantly, so the specific differences may not apply to future versions. The group names also hide detail: the Android group in Study 1 included both phones and tablets, and the Windows group included tablets, laptops and desktops.

The practical reading is that "same person, different device" is a real source of difference, that it is bigger for some tasks than others, and that nobody can say from these studies how many points a given phone would add or remove for a given person. The page on what IQ measures and the page on the CHC model define the abilities involved, and the next sections take them one at a time.

3 Why Do Speed Tasks Suffer Most on a Phone?

Speed tasks suffer most because the measured response time includes the delay of the screen and the touch layer, and that delay is a larger share of a short response than of a long one. Nicosia and colleagues (2023), in Behavior Research Methods, tested 26 popular smartphones, priced from under 100 dollars to over 1,000, running Android or iOS. A custom robot measured how long each phone took to display a change and to register a touch. They report considerable variability across smartphones, and that total device latency, the sum of display and touch latency, ranged from 35 to 140 milliseconds. Their point is that response times measured by a device include the latency of the device, so the variability could be misattributed as individual or group differences in response time.

A short comparison shows why that matters. In the Passell sample, the participant level mean simple reaction time was 304.44 milliseconds for Windows users and 374.53 for Android users, and the digit symbol matching task produced a median response of 1,360.03 milliseconds across the whole sample. Set a delay of 35 to 140 milliseconds against a response of about 300 milliseconds and it is roughly 12 to 47 percent of the response; set it against a response of about 1,360 milliseconds and it is roughly 3 to 10 percent. Those percentages are our arithmetic, not a finding of either paper, and they mix a delay measured on phones with response times measured on web users. The authors reach the same conclusion in words: a task that produces longer response times dilutes the effect of device latency, which they illustrate by comparing the two tasks.

The table gives the effect sizes Passell and colleagues report for each device against Windows computers, after controlling for age, gender and educational attainment. Positive numbers mean slower responses than Windows users. For scale, a standardized difference of 0.2 is conventionally called small, 0.5 medium and 0.8 large.

Two readings follow. For a bare reaction task, every touch device was slower than Windows, by a medium to large margin in research terms. For a task whose typical response takes about 1.4 seconds and asks the person to decide and then respond, the phone differences shrink to about a fifth of a standard deviation or less for iPhone and a little more for Android, and the iPad difference vanished. The authors also say that the effect of input type was moderate to large for simple reaction time and relatively small for digit symbol matching. That is why a clock test built on matching and searching is less exposed than a bare click test, though not immune.

ACIS includes two such speed tasks. Coding asks the person to match symbols to numbers using a key, by selecting on-screen number buttons, for 120 seconds after three practice items. Symbol Search asks whether a target symbol appears in a group, with 80 puzzles and a 2 minute limit. Both are processing speed tasks in CHC terms (Gs). The page on a normed processing speed test explains how the two combine into an index, and the sibling page on the Processing Speed Index treats the device, motor and fatigue sensitivity of that index in depth, while the page on the digit symbol substitution test covers the classic paper task. The ACIS FAQ says a physical keyboard can be helpful for Coding, which fits the evidence above.

4 Does a Phone Change a Reasoning or Verbal Score?

For reasoning and vocabulary, the research finds a gap on phones in observational data that shrinks sharply once the person who chooses a phone is accounted for, and it finds no gap in at least one randomized study, but almost all of it comes from job applicant testing and not from IQ tests. The evidence has four pieces, and the order matters because each one corrects the one before.

The first is large and observational. Arthur and colleagues (2014), in the International Journal of Selection and Assessment, analyzed 3,575,207 job applicants who completed an unproctored internet based assessment between January 2011 and April 2012. Only 1.93 percent used a mobile device, still more than 69,000 people. Scores on a personality measure were similar for mobile and nonmobile devices, but scores on a general mental ability test were substantially lower for mobile devices, and tests of measurement invariance indicated equivalence across the two samples. A reader should take both halves: the phone group scored lower, and the test appeared to measure the same thing in both groups.

The second piece asks whether that lower score is the phone or the person. Brown, Grossenbacher and Warman (2023), in the Journal of Applied Psychology, started with 76,948 applicants across more than 400 occupations and found that mobile test takers scored 0.58 standard deviations lower. Mobile use was more common among applicants with lower educational attainment and in jobs of lower complexity. After controlling for selection into device with propensity score weighting, the gap shrank to d = 0.25, and with post-stratification it fell to d = 0.10. The authors conclude that large device effects in earlier operational studies are possibly inflated by selection bias, and they recommend random assignment when evaluating device effects. For scale only, on a scale with a standard deviation of 15, 0.58 is about 8.7 points, 0.25 is about 3.75 points and 0.10 is 1.5 points; that conversion is our arithmetic, and the study measured a general mental ability test for hiring, not an IQ battery.

The third and fourth pieces use random assignment. Traylor, Hagen, Williams and Arthur (2021), in the International Journal of Selection and Assessment, randomly assigned 425 college students to one of three conditions that differed by testing environment and device and gave them a cognitive test. Test scores did not differ between conditions, despite greater distractions reported by those who tested in the distracting environment, and the result was replicated within the study. The authors suggest that self-selection and testing stakes deserve examination as explanations for the device differences seen in operational settings. Arthur, Keiser, Hagen and Traylor (2018), in Intelligence, randomly assigned participants to a desktop (174 people) or a smartphone (173 people) to complete a general mental ability measure, and report initial support for some propositions of the Structural Characteristics/Information Processing framework that Arthur and colleagues had proposed earlier: the relationships of working memory with general mental ability and with completion time were stronger on smartphones than on desktops.

Put together, the evidence supports a cautious sentence: for untimed reasoning, the phone penalty shrinks as the comparison gets better controlled, and one randomized study found none. It does not support a promise that any individual score is unaffected, and it says nothing about timed subtests, which the previous section treats separately. The page on untimed IQ tests explains what removing the clock changes in reasoning and why most batteries keep one.

Verbal knowledge adds a note about screens. Bridgeman, Lennon and Jackenthal (2003), in Applied Measurement in Education, assigned 357 college bound high school juniors to different monitor setups: a 17 inch monitor at a resolution of 1024 by 768, a 17 inch monitor at 640 by 480 and a simulated 15 inch monitor at 640 by 480, each with or without a 5 second delay between questions. They found no significant effects on math scores, and verbal scores about a quarter of a standard deviation higher with the larger, higher resolution display. Those were desktop monitors, so the study cannot give a number for a phone, but it shows that the display alone can move a verbal score when the screen is the thing that varies. In the Passell data, iPhone users did not differ significantly from Windows users on vocabulary accuracy (p = 0.07), while Android users scored lower and Macintosh users higher.

5 What Changes on a Small Touch Screen?

A small touch screen changes three things a test can depend on: how much detail you can see, how precisely you can respond, and how much of the item you can hold in view at once. The Passell authors make the same points. They describe the screen size association as strongest on tests that require fine motor input and clear perception of the full screen, and they write that stimuli which are visually distinct and memorable on a laptop or desktop may be difficult to differentiate on a phone screen, and that if response buttons are placed close together, as in the trail making task they used, it may be difficult to select reliably only the correct one on a small touchscreen.

Their Study 2 gives the clearest evidence. In 3,818 volunteers, they compared iPad and iPhone users, who share an operating system, a maker and a touch interface and differ mainly in screen size. Screen size had a significant effect on the digit symbol matching test and on parts A and B of a digital trail making task, with slower reaction times on iPhones, and no significant effect on vocabulary or simple reaction time. The effect was greatest on the trail making tasks, which require connecting targets without crossing non-target points, and the authors say effects on the other tests, timed and untimed, were small or negligible. Their own caution applies: iPads and iPhones may differ in ways other than screen size.

The International Test Commission and the Association of Test Publishers (2022), in their Guidelines for Technology-Based Assessment, name the same family of problems in general terms. They describe construct-irrelevant variance as occurring when scores reflect factors the test was not intended to measure, with the example of scrolling on a particular device interrupting a test taker's reading fluency. In the section on test development they say that when multiple devices or platforms are used, pilot studies should check that items operate similarly, and that a study of interface differences such as screen size or the amount of scrolling can reveal effects on item timing or performance. In the section on mobile delivery they note that the allowed classes of devices could be limited, by device class, input type or orientation, to offer a comparable testing condition to all end users.

What this means for a reader is easiest to see by task format, without reproducing any item. Some tasks show one large figure and ask for one choice among a few options, and these ask little of the finger. Some tasks ask the person to inspect small differences in a symbol, a pattern or a grid cell, and these ask a lot of the eye. Some ask the person to select several tiles in a remembered order, to type a string of digits, or to move pieces, and these ask a lot of the hand. The response format Passell and colleagues found harder on the smaller screen was tapping targets that sit close together. Dragging a piece to a spot or tapping a small cell makes a similar demand on the finger, but no study we opened tested drag and drop, so that extension is our inference and not a finding.

ACIS pages describe formats of each kind. Matrix Reasoning is a nonverbal fluid reasoning task in which the person picks the option that completes a visual pattern. Layer Rotation is a visual spatial task, and the technical manual says it carries no words at all, so the figures carry all of the information. The page on the cognitive domains maps each ACIS index to its CHC ability. The ACIS FAQ states what the company recommends for visual and timed subtests, and it is quoted exactly in the section on choosing a device below.

6 What About the Audio Tasks on a Phone?

Audio tasks depend on the sound reaching the ear cleanly and on nothing else competing for attention, so the question for a phone is the speaker, the headphones and the notifications, not the screen. In ACIS, Digit Span, Alphanumeric Sequencing and Arithmetic present their items as audio and need working sound. Digit Span has three sections: forward recall, backward recall, and entering the digits from smallest to largest after listening. Alphanumeric Sequencing asks the person to keep listening until the sequence ends before typing a response. These are Gwm tasks in CHC terms, and in Digit Span the audio sets the pace, so the person cannot slow it down.

The sound path is an uncontrolled variable. Woods and colleagues (2017), in Attention, Perception, and Psychophysics, built a short test to determine whether online participants are wearing headphones, because remote experiments lose control over sound presentation. When they ran it online, the scores had a bimodal distribution, which they say suggests that some participants did the task over loudspeakers despite instructions to use headphones. The study concerned hearing research and not IQ tests, but the lesson transfers: instructions to use headphones are followed by some people and not others, and a test publisher may not be able to tell which. The ACIS FAQ recommends headphones for audio-based subtests like Digit Span.

Notifications are the second variable, and a phone is the device most likely to produce them. Stothart, Mitchum and Yehnert (2015), in the Journal of Experimental Psychology: Human Perception and Performance, found that cellular phone notifications alone significantly disrupted performance on an attention-demanding task, even when participants did not directly interact with a mobile device during the task, and that the size of the effect was comparable to that seen when users actively used a phone for voice calls or text messaging. A buzz during a recall task is therefore not harmless even if the person ignores it; the study did not test a recall task, which is also attention-demanding.

The practical steps are in the section on setting up a session below. The page on the working memory test explains what a normed working memory score is built from.

7 Which Tasks Are Safer on a Phone, and Which Need More?

Tasks with large, untimed visual or verbal items are the safest on a phone, and tasks scored by speed, with small targets close together, or by precise tapping are the ones to move to a larger screen or a keyboard. The table summarizes the evidence by task type. It describes formats and does not reproduce any test content, and the last column is our practical advice drawn from the sources in the evidence column, not a published rule of any test.

The safest recommendation for an entire battery is therefore not "never use a phone" but "use the phone for the tasks where the evidence is kindest and the computer for the rest". Because a battery is taken as a whole and a test page rarely lets you choose the device per subtest, the simpler rule is to choose the device by the most demanding task you will meet. For a battery that includes a clocked symbol task, that is the computer. The page on the quick IQ test explains what a short battery costs you in precision, which matters here because the make-up of a short battery decides whether it contains a clocked speed task at all.

8 Should You Use a Phone, a Tablet or a Computer?

A computer with a physical keyboard is the safest device for a normed battery, a tablet is a reasonable middle option for visual items, and a phone is adequate for curiosity, practice and untimed items with clearly visible figures. That order follows the evidence in the previous sections and the ACIS FAQ, which is quoted exactly: "ACIS is compatible with desktop computers, laptops, tablets, and modern smartphones. For the best experience on visual and timed subtests, a larger screen, stable internet connection, and quiet environment are still recommended. A physical keyboard can be helpful for tasks like Coding, and headphones are recommended for audio-based subtests like Digit Span." That answer says compatible, not equivalent, and the page does not claim that a score taken on one device means the same as a score taken on another. We make no such claim either.

A tablet sits between the two, with a mixed record. In the Passell data iPad users were slower than Windows users on the simple reaction task (a difference of 0.53) but not on digit symbol matching (-0.03, not significant), and the larger iPad screen was associated with faster performance than the iPhone on the digit symbol task and on trail making. The group also differs in who uses it: iPad users were the oldest group in the sample. So a tablet is a good fallback when a computer is not available, but it is not a proven substitute.

When is a phone enough? When the score is for curiosity or practice, when the tasks are untimed and the figures are clearly visible, and when you will not compare the number with a number taken on another device. When is it not enough? When you will compare two results on different devices, when the battery includes tasks scored by speed, or when the result will be used for anything beyond personal interest. The page on the IQ test for adults describes the routes and uses for an adult reader, and the page on how long an IQ test takes covers duration by instrument, which matters here only because a long session on a phone brings more chances for interruption and battery trouble.

9 How Do You Set Up a Fair Phone Session?

A fair phone session fixes everything you can control, which is the device, the room, the sound and your own state, so that the score reflects the tasks and not the setup. None of the steps below is a guarantee, and none replaces a supervised session, but each removes a source of variation the sources above identify. Where a step comes from a source we say so, and where it is our own practical advice we say that too.

Choose the largest, steadiest screen you have, and a keyboard for speed tasks. The ACIS FAQ recommends a larger screen for visual and timed subtests and says a physical keyboard can be helpful for Coding. If only a phone is available, prop it on a stand or a flat surface instead of holding it, so that the screen does not move while you respond; that is our advice, not a published rule.

Silence everything. Turn on a do not disturb mode, silence calls and message alerts and close other apps. Stothart and colleagues (2015) found that notifications alone disrupted an attention-demanding task even without any interaction with the phone, which makes the phone the one testing device that carries its own interruptions.

Put headphones in before the first audio subtest. ACIS recommends headphones for audio based subtests like Digit Span, and Woods and colleagues (2017) show that some online participants told to wear headphones did not. Set the volume during the instruction audio, which plays before the scored items, and not during an item.

Charge the phone and use a stable connection. The ACIS FAQ lists a stable internet connection among its recommendations. The Standards ask that technology based administrations avoid distractions such as equipment or connectivity failures, and a low battery warning or a dropped connection in the middle of a timed task is exactly such a distraction.

Pick a quiet room and a time of day when you are alert. The page on online cognitive test data quality explains why unsupervised sessions are a known concern in research: it reports that stricter attention checks excluded about 62 percent of takers on one crowdsourcing panel and 22 percent on another.

Use breaks the way the format allows. ACIS allows breaks, saves progress between subtests and gives 30 days to complete. A long battery on a phone is therefore best split at subtest boundaries, not paused in the middle of a task. We state nothing about mixing devices inside one battery beyond what the FAQ says, so the cautious rule is to keep one device for the whole battery and for any retest. The times by instrument are on the page about duration, which helps plan the sittings.

Use the practice items to learn the response method. The Standards, discussed in the last section, say that test takers should be given instructions, practice and other support to reduce construct-irrelevant variance, and that practice opportunities should be given when equipment is involved. If a page offers practice items, use them to learn the response method on the device you have chosen, before any clock starts.

10 How Should You Read a Score Taken on a Phone?

Read a phone score as an estimate taken under conditions the test could not control, report it with its confidence interval, and compare it only with another score taken the same way. The reading rules are the same as for any normed score, and the device adds a reason for caution on the tasks the earlier sections flagged.

A normed battery reports several layers. A subtest returns a scaled score, on a scale with a mean of 10 and a standard deviation of 3 in ACIS, from 1 to 19. An index or a Full Scale IQ is on the standard scale, with a mean of 100 and a standard deviation of 15, and the report adds a percentile and a 95 percent confidence interval. The page on how IQ scores are normed explains how a raw score becomes any of these, and the page on IQ score and percentile explains why the two cannot be averaged. The confidence interval is the honest summary of precision: it says that the true score is likely to lie in a range and not at a single point, and a device effect is one of the things that can widen the real uncertainty beyond what an interval computed under standard conditions shows.

The comment on Standard 2.3 in the Standards (AERA, APA and NCME, 2014) says that the consistency of total scores may be acceptably high while subscores have unacceptably low reliability, so a score from one speed task on one phone is the number to lean on least. The page on reliability and validity uses examples to show how consistency and meaning are separate questions. The comment on Standard 2.4 adds that when interpretation centers on the peaks and valleys of a score profile, the reliability of score differences is critical. That matters for the comparison most readers will make: the gap between a speed index and a reasoning index. The page on index score differences reports that in a WAIS-IV simulation 49 percent of normal profiles had a 15 point gap between some pair of the four index scores, which means a low speed index beside a strong reasoning index is common even when nothing is wrong, and a phone session is one more reason to avoid reading it as a finding.

Three rules follow. First, do not convert a phone gap into points. The studies above report standardized differences for groups on specific tasks, and none lets anyone subtract a number from a personal score. Second, do not compare a score taken on a phone with one taken on a computer as if the two were equivalent; the sources do not establish that, and we do not claim it for ACIS. Third, treat an unusually low score on a clocked task, taken on a small screen, as a prompt to retake that task on a larger screen with a keyboard, with the same instructions, before drawing any conclusion. A reader who meets a phone product that reports a "brain age" will find the three meanings of that phrase sorted on the sibling page on the brain age test.

11 What Should a Test Publisher Show Before You Trust a Phone Score?

Before you trust a phone score, a publisher should show which devices the test was built and checked for, what evidence says scores hold up across them, and what conditions the norms were collected under. These are not our rules. They follow from the published guidelines on computer based and internet delivered testing, which treat devices as a matter for the publisher and not for the test taker alone.

The International Test Commission (2006), in the International Journal of Testing, sets out guidelines for developers, publishers and users of tests delivered by computer and the internet. Under the guideline on hardware and software, developers are asked to give a clear description of the minimum hardware and software requirements, naming supported browsers for internet testing, and to conduct adequate usability testing on the appropriate delivery platforms to ensure consistency of appearance and delivery. Under the guideline on equivalence, where a computer or internet test is developed from a paper version, developers are asked to document comparable reliabilities, correlations at the expected level, comparable correlations with other tests and criteria, and comparable means and standard deviations or appropriate calibration. For internet based tests, the guidelines say that equivalence and norming studies should be conducted over the internet with participants completing the test under conditions that represent those the intended population will experience, for example unproctored or unstandardized conditions.

The International Test Commission and the Association of Test Publishers (2022) updated this in their Guidelines for Technology-Based Assessment, which add that pilot studies across multiple devices should check that items operate similarly and that construct-irrelevant variance is not introduced. Germine, Reinecke and Chaytor (2019) name rapid change in hardware, software and device ownership as a barrier that can make particular tests and norms obsolete, which is why the question of which devices a set of norms covers belongs on a buyer's checklist.

Six questions turn those guidelines into a checklist for any phone test:

1. Does the page name the devices and browsers it supports, and say which tasks are better on which?
2. Does it name the norm group, with an age range, and say how the norms were collected?
3. Does it say whether the norms were collected under conditions like yours, including on phones?
4. Does it report precision, such as a confidence interval, and not only a single number?
5. Does it say what the test cannot be used for, such as hiring, school accommodations or clinical decisions?
6. Does it avoid promising that scores are equivalent across devices unless it shows evidence?

A page that fails the first three is a quiz, whatever it calls itself. The same test can be applied to the whole market, and the technical documentation of a normed battery is where the answers to questions 2 to 4 should be found.

12 Where ACIS Sits, and What a Buyer Should Do

ACIS is a normed online battery that runs in a phone browser, but it is an unsupervised test whose own FAQ recommends a larger screen, a keyboard for Coding and headphones for audio subtests, and a phone buyer should plan around those recommendations. ACIS reports a Full Scale IQ and six primary indices from 20 subtests in six CHC domains, in three forms: Quick at 15 dollars for 6 subtests in about 45 minutes, Optimized at 30 dollars for 13 subtests in about 110 minutes, and Full Scale at 50 dollars for all 20 subtests in about 175 minutes (prices read October 29, 2026 and subject to change). Every form reports a Full Scale IQ with its 95 percent confidence interval. Payment is one time with no subscription, breaks are allowed, progress is saved between subtests, and a buyer has 30 days to complete.

The device statement is the one quoted earlier: ACIS "is compatible with desktop computers, laptops, tablets, and modern smartphones". Compatible is not equivalent. We publish no ACIS figures by device, and we do not claim that scores are equivalent across devices. The limits that apply to every form, whatever the device, are plain: ACIS is online and unsupervised, it is not a clinical or diagnostic instrument, it is not for hiring, school accommodations or admission to high IQ societies, and it is in English only, with the technical manual noting that verbal and knowledge tasks are affected by English proficiency and educational exposure. The documentation of the battery is in the technical manual, and we state nothing about its figures on this page.

The Standards for Educational and Psychological Testing (AERA, APA and NCME, 2014), which we read in the open access edition on October 29, 2026, bear directly on a phone session. The comment on Standard 6.1 says test developers should specify the standardized administration conditions that support the intended uses of scores, that technology and the internet have made it possible to administer tests in settings where conditions are not strictly controlled, and that those who allow a lack of standardization are responsible for providing evidence that it did not affect test-taker performance or the comparability of the scores. The same comment says that time limits, response formats and complex interfaces should be scrutinized for construct-irrelevant variance. The comment on Standard 6.3 says that testing environments that are not monitored, for example on the internet, should meet the standardized administration conditions, and otherwise the report on scores should note that standardized conditions were not guaranteed. Standard 6.4 asks for a testing environment with reasonable comfort and minimal distractions, and its comment says that in general testing conditions should be equivalent to those that prevailed when the norms were obtained. Standard 6.5 asks that test takers receive instructions, practice and other support to reduce construct-irrelevant variance. The practical reading is that a buyer has to supply the controls an examiner would, and that the result should be read as an unsupervised score.

A concrete plan follows from all of this. If you have a computer with a keyboard, use it for the whole battery, and use headphones for the audio subtests. If a phone is all you have, start the free trial, which needs no card and allows one trial per person, on the device you plan to use, and see how the tasks feel on your screen before you pay; finished trial subtests are kept if you later unlock a form that includes them. Then choose the form by the question you are asking. Keep the same device for every subtest and for any retake, set the room up as the section on a fair session describes, and read the result with its confidence interval. ACIS offers a refund stated this way: "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 guarantee and not an unconditional money back one.

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. Web pages were read on October 29, 2026, and prices can change. Any figure we derived ourselves is labeled in the text as our arithmetic.

• Passell E, Strong R, Rutter L, Kim H, Scheuer L, Martini P, Grinspoon L, Germine L. Cognitive test scores vary with choice of personal digital device. Behavior Research Methods, 2021, volume 53, issue 6, pages 2544 to 2557 (full text read in PubMed Central, PMC8568735).
• Nicosia J, Wang B, Aschenbrenner A, Sliwinski M, Yabiku S, Roque N, Germine L, Bateman R, Morris J, Hassenstab J. To BYOD or not: Are device latencies important for bring-your-own-device (BYOD) smartphone cognitive testing? Behavior Research Methods, 2023, volume 55, issue 6, pages 2800 to 2812.
• Germine L, Reinecke K, Chaytor N. Digital neuropsychology: Challenges and opportunities at the intersection of science and software. The Clinical Neuropsychologist, 2019, volume 33, issue 2, pages 271 to 286.
• Mead A, Drasgow F. Equivalence of computerized and paper-and-pencil cognitive ability tests: A meta-analysis. Psychological Bulletin, 1993, volume 114, issue 3, pages 449 to 458.
• Arthur W, Doverspike D, Munoz G, Taylor J, Carr A. The use of mobile devices in high-stakes remotely delivered assessments and testing. International Journal of Selection and Assessment, 2014, volume 22, issue 2, pages 113 to 123.
• Brown M, Grossenbacher M, Warman Z. Self-selection as an explanation for general mental ability test score differences between mobile and nonmobile devices in observational studies. Journal of Applied Psychology, 2023, volume 108, issue 7, pages 1190 to 1206.
• Traylor Z, Hagen E, Williams A, Arthur W. The testing environment as an explanation for unproctored internet-based testing device-type effects. International Journal of Selection and Assessment, 2021, volume 29, issue 1, pages 65 to 80.
• Arthur W, Keiser N, Hagen E, Traylor Z. Unproctored internet-based device-type effects on test scores: The role of working memory. Intelligence, 2018, volume 67, pages 67 to 75 (abstract read on the IDEAS listing).
• Bridgeman B, Lennon M, Jackenthal A. Effects of screen size, screen resolution, and display rate on computer-based test performance. Applied Measurement in Education, 2003, volume 16, issue 3, pages 191 to 205.
• Stothart C, Mitchum A, Yehnert C. The attentional cost of receiving a cell phone notification. Journal of Experimental Psychology: Human Perception and Performance, 2015, volume 41, issue 4, pages 893 to 897.
• Woods K, Siegel M, Traer J, McDermott J. Headphone screening to facilitate web-based auditory experiments. Attention, Perception, and Psychophysics, 2017, volume 79, issue 7, pages 2064 to 2072.
• 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 (guideline text read in the ITC PDF, version 1.0, July 2005).
• International Test Commission and Association of Test Publishers. Guidelines for technology-based assessment, 2022. Guidelines 1.12 and 3.21 and the chapter introduction, read October 29, 2026.
• American Educational Research Association, American Psychological Association and National Council on Measurement in Education. Standards for Educational and Psychological Testing, 2014. Standards 2.3 and 2.4 and 6.1, 6.3, 6.4 and 6.5 and their comments, read in the open access edition on October 29, 2026.
• Pearson Assessments. Wechsler Adult Intelligence Scale, Fifth Edition (WAIS-5) product page. Age range, completion times and administration options, read October 29, 2026.

14 Frequently Asked Questions