How to Prepare for an IQ Test
What Helps, What Does Nothing, What Ruins It

0 Quick Answer

Updated August 16, 2026 by Structural. Useful preparation for a cognitive assessment consists of sleeping properly, testing at a time of day when you function well, understanding the format in advance, and removing interruptions. That is close to the whole list.

Direct answer: preparation splits cleanly into two categories. Removing noise means eliminating things that would depress your score for reasons unrelated to ability, and it is entirely legitimate. Manufacturing signal means practising the specific items so you perform above your actual level, and it produces a score that is wrong.

The distinction matters more than it sounds. A score inflated by item practice is not a better result, it is a broken one, and if the assessment is for any purpose beyond curiosity the broken version is worse than useless. Everything below is organised around which side of that line a given action falls on.

1 Removing Noise Versus Manufacturing Signal

Every measurement contains the thing being measured plus everything else that affected the reading. On a cognitive assessment the everything else includes fatigue, anxiety, an unfamiliar interface, distraction, illness, and simply not understanding what the task is asking.

Preparation that removes those does not change your ability. It changes how accurately the test observes it. Somebody who sleeps properly and understands the format produces a score closer to their actual level than the same person tested exhausted and confused. That is a better measurement, not a better performance, and there is nothing questionable about pursuing it.

Preparation that manufactures signal works differently. Practising the specific item types until you recognise them raises the score without raising the ability, so the number no longer describes the person. The score is inflated in a way that does not transfer to anything the test was supposed to predict.

The reason this matters practically is that inflated scores fail exactly where scores are supposed to be useful. A gifted placement based on a coached score puts a child in a programme that does not fit. An accommodation decision built on a distorted profile addresses the wrong problem. And for the individual, an inflated result answers the question you asked with a number about someone else.

There is a middle case that causes genuine confusion: understanding a task type is not the same as practising its items. Knowing what a matrix reasoning item asks you to do removes a misunderstanding that would otherwise cost you items you could have solved. Having worked through fifty matrix items until you know the common rule families is different, because it converts a novel-problem task into a recognition task, and novelty is the whole point of a fluid reasoning measure. Section 4 draws that line precisely.

2 Sleep, the Largest Single Factor

If you do one thing, sleep properly the night before. Nothing else on this page has a comparable effect, and it is the intervention most often skipped because it does not feel like preparation.

The evidence base here is large and consistent. A meta-analysis of short-term sleep deprivation across many studies found that sustained attention is the domain most affected, with substantial degradation from a single night of restriction. That matters for cognitive assessment because sustained attention is precisely what processing speed and working memory subtests demand, and it is the first thing to fail.

The resulting profile has a characteristic shape. Crystallised knowledge holds up well, since retrieving vocabulary does not require sustained vigilance. Reasoning holds up moderately. Processing speed and working memory fall noticeably. Somebody tested after a poor night therefore produces a profile with an apparent relative weakness in speed and memory, and that pattern is indistinguishable from a genuine one on the score sheet.

Two practical points follow. Sleeping well the night before matters most, but a single good night does not fully repair a week of restriction, so if you have a choice about when to test, choose a period when you are not accumulating debt. And do not attempt to compensate with extra caffeine, which restores subjective alertness considerably more than it restores performance, leaving you feeling ready while measuring worse.

If you cannot sleep before a scheduled supervised assessment, tell the examiner. It is exactly the kind of information a validity statement exists to record, and an examiner who knows will either note it or reschedule. Concealing it produces a documented score that understates you, which is a much worse outcome than a rescheduled appointment.

3 Food, Caffeine, and Exercise

These get more attention than they deserve, and the honest summary is that they matter mainly through avoiding deficits rather than by producing gains.

Food. Being hungry is distracting and hunger competes for attention. Eat something before testing. Beyond that, no particular food improves cognitive performance in a way that shows up on an assessment, and claims about specific brain foods do not survive contact with the evidence. A normal meal you would eat anyway is correct, and a large unfamiliar one is a mild risk rather than an advantage.

Caffeine. The rule is to do what you normally do. A habitual coffee drinker who skips it will be in mild withdrawal, which is worse than the caffeine. A non-drinker who takes some for the occasion gets jitteriness and elevated heart rate without a reliable cognitive benefit. Caffeine restores performance in sleep-deprived people toward their rested baseline, which is a real effect, and it does not push a rested person above their own ceiling.

Exercise. Regular exercise over months is associated with better cognitive function, which is a genuine finding and irrelevant to next Tuesday. Vigorous exercise immediately before testing is more likely to leave you depleted than sharpened. A short walk to settle is fine and does nothing measurable either way.

Alcohol. Avoid it the night before. The effect on sleep architecture persists after the subjective effects have gone, and the resulting sleep is less restorative even when its duration is normal. This is one of the few items on this list where a modest amount the previous evening has a measurable cost the next morning.

The pattern across all four is that you are looking for your ordinary baseline rather than an enhanced state. Cognitive assessment measures a stable characteristic, and attempts to acutely elevate performance mostly introduce variance rather than gain.

4 Learning the Format Without Practising the Items

This is where the noise and signal line runs, and it is worth being precise because the useful version of this is genuinely useful.

What you should know in advance: which subtests exist and roughly what each asks, whether limits are per item or per section, whether wrong answers can end a subtest early through a discontinue rule, and how responses are entered. All of that is procedural. Not knowing it costs you items for reasons that have nothing to do with ability, and every good instrument publishes it.

The specific case of discontinue rules deserves emphasis because it changes optimal behaviour and is rarely explained. On a subtest that ends after several consecutive failures, a wrong answer is not neutral. Guessing rapidly to save time on a hard item can terminate the subtest and cost you the easier items that would have followed. Knowing the rule exists changes what you do, and it is procedural knowledge rather than an advantage over other takers who could equally have read it.

What crosses the line is working through banks of items of the same type until the underlying rule families become familiar. On a fluid reasoning measure this defeats the construct directly. The task is to solve a novel problem, and a problem you have seen a hundred variants of is not novel. The score afterwards measures your familiarity with matrix puzzles, which is a real skill and not the one being assessed.

The practical test is whether what you learned transfers to a different instrument. Knowing that matrix items ask you to infer a rule transfers everywhere and is procedural. Knowing that a particular publisher's matrix items favour certain rule types transfers nowhere and is coaching.

Official practice items, provided by the instrument itself and administered before scoring begins, are the correct amount of this. They exist precisely to remove the format-learning cost from the scored portion, and taking them seriously is the single highest-value five minutes available.

5 What Practice Effects Actually Do

Retesting on a cognitive measure produces a higher score. This is well documented, has been quantified across many studies, and is worth understanding because it explains both why coaching works in the narrow sense and why it is worthless in the broad one.

Meta-analytic work on practice effects in neuropsychological assessment finds gains that are substantial on a second administration and larger when the interval is short and the identical form is used. Comparable work on retesting in personnel selection finds meaningful gains from repeated testing and from coaching specifically, with the largest effects where the same test is repeated.

Three properties of these gains determine what they are worth.

They are largest on the tasks most dependent on strategy and smallest on those closest to general ability. A person who works out an efficient approach to a spatial construction task carries that approach into the retest. There is no comparable trick for a vocabulary item, so verbal measures show smaller practice gains than performance ones.

They decay with the interval. A retest a week later shows a large effect. A retest a year later shows much less. The gain is largely memory for the specific material and strategy rather than a durable change.

And critically, they do not transfer. A higher score obtained through practice does not come with the improved real-world performance the original score predicted. This is why practice effects are treated as a threat to validity in the assessment literature rather than as improvement, and why clinical guidance specifies minimum retest intervals and the use of alternate forms.

The implication for preparation is direct. Practising raises the number and moves it away from the thing the number was measuring. If you want the score to be about you, do not do it. The separate and genuinely interesting question of whether cognitive ability itself can be changed is a different topic, addressed in Can You Improve Your IQ?.

6 Test Anxiety

Anxiety about being assessed is common and has a measurable effect on performance. It is also one of the few sources of noise that can be reduced with a small amount of deliberate effort.

The mechanism most supported by the evidence is competition for working memory. Worry occupies the same limited capacity the task needs, so performance falls most on tasks that were already memory-demanding, which in a battery means working memory subtests first and reasoning second. A large meta-analytic review of test anxiety across three decades found consistent negative associations with performance across settings and age groups.

What reduces it, in rough order of how much they help:

Knowing what is coming. Most test anxiety is anticipatory and much of it is about the unknown. Reading the format in advance, as in section 4, removes a specific and unnecessary source of it.

Understanding that the score is not a verdict. A cognitive assessment measures a set of specific abilities with known error, against a reference group, on one occasion. It does not measure worth, potential, or character, and treating it as a judgment is both inaccurate and the main driver of the anxiety.

Expecting to fail items. Every adaptive or difficulty-graded subtest is designed so that you reach items you cannot solve. That is how the ceiling is found. People frequently report the moment items became impossible as the moment they panicked, and knowing in advance that it is designed to happen defuses it entirely.

Not treating a single administration as final. Retesting is possible, results carry confidence intervals, and one session is one estimate. That framing is accurate rather than reassuring, which is why it works.

If anxiety is severe enough to be clinically significant, it is worth raising before a supervised assessment rather than after. An examiner can accommodate pacing and will record it in the validity statement, which is the correct handling of a factor that genuinely depressed the measurement.

7 Environment and Hardware

For an online assessment the conditions are entirely your responsibility, and this is where most avoidable score loss happens.

Interruptions. A speeded subtest can run for two minutes. A single interruption during it costs a substantial fraction of that subtest's score, and nothing in an unsupervised session prevents one. Close notifications, silence the phone, and tell anybody present that you are unavailable. This is a two minute action with a larger effect than anything else in this section.

Input device. The motor component of speeded subtests is timed along with the cognitive part. A mouse is generally faster than a trackpad for pointing tasks, and both are faster than a touchscreen for precise targets. Use whatever you genuinely use daily rather than something borrowed for the occasion.

Screen. Visual and spatial subtests present figures that need to be seen clearly. A larger screen at a normal working distance is better than a phone, and this is one of the few hardware choices with an obvious direction.

Audio. Any subtest presenting material aurally requires working sound, and a failure here does not look like a technical problem on the score sheet, it looks like a memory deficit. Confirm audio works before starting rather than discovering it partway through.

Connection. A slow or unstable connection affects timing on speeded subtests. Prefer a wired or reliable connection, and avoid testing on a shared network under heavy load.

One sitting. A full battery should be completed in one session. Splitting it across days means the parts were taken by a person in different states, and the profile becomes a composite of two occasions rather than a picture of one.

8 When to Test

Time of day has a real if modest effect, and the direction is individual rather than universal.

People differ in when their alertness peaks, and the difference between morning and evening types is large enough to matter for a speeded measure. The useful rule is not to test in the morning but to test when you would choose to do demanding work if the choice were yours. For most people that excludes the early afternoon dip and the last hours before sleep.

Beyond the daily cycle, avoid testing during periods of acute stress, immediately after a long flight, while ill, or while dealing with something that will occupy your attention regardless of intent. None of these is a small effect, and all of them are avoidable by choosing a different week.

Medication timing is worth a specific note. Anyone taking a medication that affects attention or alertness should test in their usual medicated state rather than deliberately unmedicated, unless a clinician has specifically requested otherwise for diagnostic reasons. Testing off medication does not reveal a truer baseline for most purposes, it produces a measurement of somebody in an unusual state, and the interpretation of the result then requires knowing that, which nobody will if you do not say so.

Illness deserves the same treatment. A cold depresses processing speed and attention measurably, and there is no adjustment that recovers it afterward. Rescheduling is nearly always the correct answer, and where the assessment is supervised it is a routine request rather than an imposition.

9 A Practical Timeline

Everything above condensed into what to do and when.

The last row is the one people ignore. An immediate retake produces the largest practice effect available and the least valid score, for the reasons in section 5.

Notice how little of the table is about the test itself. Two rows out of eight concern what happens during the assessment, and one of those is about expectations rather than technique. That distribution is the honest shape of preparation for a cognitive measure, and it is why the topic supports so little content compared to preparation for an examination of knowledge. There is no syllabus, so there is nothing to revise, and the entire useful intervention is arranging for the measurement to observe you under conditions that represent you.

10 Preparing for a Supervised Assessment

A formal assessment with a psychologist has additional considerations, and most of them are about information rather than performance.

Bring your history. A clinician will ask about education, employment, medical background, medication, and what prompted the assessment. Vague answers here reduce the quality of the interpretation more than a few points on any subtest would, because the report's value lies in the clinician relating scores to context.

Bring your glasses and hearing aids. This sounds trivial and is not. Uncorrected sensory impairment depresses scores on specific subtests in ways that look like cognitive findings, and it is among the more common causes of a spuriously low profile in older adults.

Bring prior reports if you have them. Comparison against a previous assessment is often the most informative part of a new one, and the examiner needs the earlier instrument and date to interpret the difference correctly, particularly given practice effects and differing norms.

Say what you want from it. An assessment conducted to answer a specific question produces a report that answers it. One conducted without a stated purpose produces a description. If you are seeking documentation for accommodation, clarification of a suspected difficulty, or a baseline before a medical intervention, say so at the start.

And expect it to take longer than you think. A full battery runs two to three hours, sometimes across two appointments, plus the interview. Arriving with a hard stop scheduled afterward is a genuine source of degraded performance in the final subtests. The full sequence is described in The Adult IQ Testing Process.

11 Preparation Myths

Several widely repeated pieces of advice are either false or actively harmful.

That brain training apps will raise your score. Training on a task improves that task. Transfer to different tasks is weak and transfer to general ability is not established by the short, low-intensity protocols these apps use. The nuanced version of this question, including where meaningful gains do occur, is in Can You Improve Your IQ?.

That you should study general knowledge beforehand. Knowledge subtests sample broadly from material accumulated over a lifetime. A week of cramming cannot move a measure built on decades of exposure, and the attempt mostly produces anxiety.

That answering faster is better. Only on subtests where speed is the construct. On a reasoning subtest with a per-item limit, unused time does not transfer to other items, so rushing costs accuracy and gains nothing.

That you should never guess. Where there is no penalty for a wrong answer and no discontinue rule in play, an educated guess is strictly better than leaving an item blank. The rule reverses when a discontinue rule is active, which is why knowing whether one exists matters.

That a supplement will help. No commercially available supplement has evidence of acutely improving cognitive performance in a healthy, adequately nourished person. Correcting an actual deficiency helps, and that is a medical matter rather than a preparation strategy.

That you can talk your way to a better score. On subtests with graded scoring, giving a fuller answer that demonstrates the concept does earn more credit than a minimal one, which is not a trick but a scoring rule worth knowing. Padding an answer you do not understand earns nothing, because the rubric rewards the specific quality of the relation stated rather than length.

12 Preparing for ACIS Specifically

ACIS is unsupervised and computer administered, so the environment and hardware section carries most of the weight.

Every subtest states its time limit and discontinue rule before it begins, so the procedural knowledge described in section 4 is available without seeking it out. Reading those screens rather than clicking past them is the intended preparation.

Three subtests present material aurally and will not start until sound is confirmed working. That gate exists because a muted session on an audio subtest produces a score that looks like a severe memory deficit and is a technical failure, which is a specific and previously observed failure mode rather than a hypothetical one.

The battery runs twenty subtests across six domains and is designed for one sitting. Splitting it produces a profile assembled from a person in two different states, which is worth avoiding for the same reason you would not measure your height twice with different postures.

Practice items precede scored items where the task type warrants it. They are not scored and exist purely to remove the interface-learning cost, so the correct approach is to use them until the task is clear rather than to move through them quickly.

Results are reported per domain with confidence intervals attached, which is worth knowing in advance because it changes how to read the outcome. A single domain sitting lower than the rest is not a diagnosis, and a difference of a few points between domains sits inside the error of both. The pattern worth attending to is a large and consistent divergence, and the report presents the intervals so that distinction is visible rather than leaving it to inference.

13 FAQ: Preparing for a Cognitive Assessment

14 Best Next Step

Useful preparation is short: sleep, test when you function well, read the format, and remove interruptions. Everything beyond that either does nothing or moves the score away from the thing it was measuring.

If you want the format knowledge described in section 4, read The Subtests Inside an IQ Test and Why IQ Tests Are Timed. If you are heading toward a supervised assessment, read The Adult IQ Testing Process. If you are ready now, take the assessment under the conditions described above.

15 Sources Behind This Page

Claims about sleep, practice effects, and anxiety come from meta-analytic work rather than single studies. Claims about administration come from publisher manuals and professional standards.

• Lim, J. & Dinges, D.F. (2010). A meta-analysis of the impact of short-term sleep deprivation on cognitive variables. Psychological Bulletin, 136(3), 375-389. Establishes that sustained attention is the domain most degraded by sleep restriction.
• Calamia, M., Markon, K. & Tranel, D. (2012). Scoring higher the second time around: meta-analyses of practice effects in neuropsychological assessment. The Clinical Neuropsychologist, 26(4), 543-570. Quantifies practice gains and how they vary with interval and test type.
• Hausknecht, J.P., Halpert, J.A., Di Paolo, N.T. & Moriarty Gerrard, M.O. (2007). Retesting in selection: a meta-analysis of coaching and practice effects for tests of cognitive ability. Journal of Applied Psychology, 92(2), 373-385. Effects of coaching and repeated testing on cognitive ability scores.
• von der Embse, N., Jester, D., Roy, D. & Post, J. (2018). Test anxiety effects, predictors, and correlates: a 30-year meta-analytic review. Journal of Affective Disorders, 227, 483-493. Consistent negative associations between test anxiety and performance across settings.
• American Educational Research Association, American Psychological Association & National Council on Measurement in Education. Standards for Educational and Psychological Testing. Standardised administration conditions, and the documentation of factors affecting score validity.
• Pearson (2008). WAIS-IV Technical and Interpretive Manual. Test-retest data, retest interval guidance, and the role of practice items in administration.
• Pearson (2014). WISC-V Technical and Interpretive Manual. Documents discontinue rules and the empirical criteria used to set them, which is why guessing strategy interacts with them.
• Pearson Clinical Assessment Scientific Council (2023). Standardized Clinical Assessment for Practitioners: A Primer. What examiners observe and record about testing conditions and examinee state.
• American Psychological Association. Understanding psychological testing and assessment. What to expect from a professional assessment and what information to bring to it.
• Simons, D.J., Boot, W.R., Charness, N., Gathercole, S.E., Chabris, C.F., Hambrick, D.Z. & Stine-Morrow, E.A.L. (2016). Do brain-training programs work? Psychological Science in the Public Interest, 17(3), 103-186. Systematic review finding strong evidence of task-specific improvement and weak evidence of transfer.
• Voncken, L., Albers, C.J. & Timmerman, M.E. (2019). Improving confidence intervals for normed test scores. Behavior Research Methods. Open access. Why a single administration is one estimate with error rather than a definitive value.
• Crawford, J.R., Garthwaite, P.H. & Slick, D.J. (2009). On percentile norms in neuropsychology: proposed reporting standards. The Clinical Neuropsychologist, 23(7), 1173-1195. How results should be reported so that a single session is not read as definitive.