Familiarity with the question formats is the only preparation that reliably moves an IQ score, and it moves it quite a lot on a first attempt. People who have never seen a matrix reasoning item perform noticeably worse than people who have, and a single practice session closes most of that gap. Everything else on the list is about not losing points you already have: sleep, anxiety, pacing, and the room you sit in.
The routine, in order
- Learn the item formats. Work through examples of matrix reasoning, number series, verbal analogies and spatial rotation until you recognise each on sight. This is the single largest legitimate gain available.
- Sleep properly for two nights. Not one. Two consecutive nights of seven to nine hours outperforms a single good night, and sleep loss hits working memory hardest, which is what most timed items lean on.
- Take one full timed practice test. A complete run under the clock calibrates your pacing and removes the novelty that costs points on a first attempt.
- Plan your skip strategy. Decide in advance how long you will give a single question before moving on. Roughly 90 seconds is a reasonable ceiling on a 25-question timed test.
- Set up the room. Quiet, warm, phone in another room, no interruptions for the full window. Test conditions affect scores measurably and this is the easiest variable to control.
Why format familiarity works
The first time you meet a matrix reasoning item, part of your effort goes into working out what is being asked. The grid changes along two dimensions at once, one option breaks a rule the others obey, and none of that is explained. Once the format is familiar, that overhead disappears and the whole 90 seconds goes to the actual problem.
This is a real effect and it is also the reason retest scores rise. It does not mean you have become more intelligent between attempts. It means the first score was depressed by unfamiliarity, and the second one is closer to your actual ability. Test publishers account for this by holding minimum intervals between administrations.
The four formats worth knowing before any general test: matrix completion, number series, verbal analogy, and mental rotation. An hour spread across all four is enough to remove the novelty penalty.
Sleep is the largest controllable variable
Sleep deprivation hits exactly the abilities IQ tests lean on hardest. Working memory degrades first, then sustained attention, then processing speed. A single bad night can cost several points on a timed test, and the effect is largest on the items that require holding several things in mind at once.
Two nights matter more than one. Recovery from a sleep debt is not instant, so the night before the night before counts. Aim for seven to nine hours on both.
Caffeine is worth a note. A normal dose helps if you are already a regular consumer and are testing at your usual time. Doubling your intake on test day trades alertness for jitter, and jitter costs more than the alertness gains on items requiring careful reading.
Anxiety costs more points than anything you could revise
Test anxiety measurably suppresses scores, and it does so through working memory: worry occupies the same limited resource the test is trying to measure. Someone anxious enough is effectively taking the test with less capacity available.
The interventions that help are unglamorous. Arrive early enough that you are not rushing. Breathe slowly for a minute before starting, which is enough to bring heart rate down. Start with the questions you find easy, since early success reduces the physiological response more effectively than telling yourself to relax.
Reframing helps too, and there is decent evidence for it. Interpreting the physical signs of arousal as readiness rather than as fear changes performance, because the sensations are largely the same and the interpretation is the part you control.
Pacing and the skip decision
On a 25-question timed test, the most common way to lose points is spending six minutes on one hard item and then rushing five easy ones at the end. Every question is worth the same. A hard one is worth no more than an easy one you did not reach.
Set a ceiling before you start. Around 90 seconds per question is sensible on a 25-minute test. When you hit the ceiling, take your best guess and move. If the test lets you skip and return, use it; if unanswered items are scored as wrong, always guess rather than leaving a blank.
On multiple choice, eliminate before you select. Removing two implausible options doubles the value of a guess, and on sequence items the wrong answers are frequently constructed from the near-miss rules, so identifying why an option is wrong often reveals the right rule.
What does not work
Brain training apps do not generalise. The research is consistent on this: you get better at the trained task and the improvement does not transfer to general cognitive ability. Practising dual n-back makes you good at dual n-back.
Memorising answer keys from practice tests is worse than useless, since any decent test randomises its item pool. Our own tests draw 25 questions from a 30-item bank in random order for exactly this reason.
Cramming general knowledge does not help on fluid reasoning items, which are designed to require no prior knowledge. It does help slightly on vocabulary items, but vocabulary accumulates over years of reading and a week of word lists produces recognition without any sense of how the words are used.
Logical Reasoning Test
25 questions, 28 minutes, scored as soon as you finish.
Take the Logical Reasoning TestQuestions people also ask
How much can practice raise an IQ score?
Format familiarity typically accounts for a first-to-second attempt rise of several points, with most of it arriving on the second attempt and little afterwards. It is not raising your underlying ability; it is removing the novelty penalty that suppressed the first score.
Should I practice the day before or a week before?
A week before, with a light refresher the day before. Practising heavily the day before adds fatigue without adding familiarity, and the sleep you lose to it costs more than the practice gains.