A timed reasoning test measures your ability under specific conditions. Change the conditions and the number changes, sometimes by several points, without your underlying ability shifting at all. This guide covers what genuinely affects a result and what to ignore.
Reasoning tests lean heavily on working memory, and working memory is unusually sensitive to interruption. A single notification can cost you the thread of a multi-step problem, and rebuilding it takes far longer than the interruption itself.
Before starting, deal with the obvious:
None of this is dramatic, and that is rather the point. These are cheap changes that remove sources of avoidable error.
Sleep loss affects sustained attention and working memory more than it affects knowledge, which is exactly the wrong profile for a reasoning test. A tired person can usually still recall facts while performing noticeably worse at multi-step problems.
Time of day matters less than people expect, with one caveat: take the test when you are normally alert rather than at your personal low point. For most people the trough is early afternoon.
Food matters mainly in the negative. Taking a timed test hungry adds a distraction you do not need. There is no meal that raises reasoning ability, and claims otherwise should be treated sceptically.
Most people lose more points to poor pacing than to any gap in ability. The failure mode is predictable: an early hard question consumes several minutes, time pressure builds, and the last few questions get rushed or skipped.
The fix is a rule decided in advance, before you are under pressure:
On a test with no penalty for wrong answers, never leave a question blank. A blind guess on four options has a 25% expected return; a blank has zero.
Better still, guess after eliminating. Removing one option takes your odds to one in three, and two takes you to a coin flip. Across a whole test, disciplined elimination on the items you cannot solve is worth more than one extra hard question solved.
Even carefully administered tests carry a margin of error, typically several points in either direction. Reputable reports give a confidence range rather than a single figure for this reason.
Two consequences worth internalising:
This is also why domain breakdowns built on only a few questions each should be read as rough indications rather than precise measurements. One unlucky question can shift a small subtest noticeably.
Some widely repeated advice has little support behind it:
The honest summary is that preparation removes obstacles rather than adding ability. Good conditions, adequate sleep, and sensible pacing let you score at the top of your genuine range. That is a real gain, and it is the one available.
Yes. Sleep loss affects sustained attention and working memory in particular, which is what timed reasoning tests depend on most. A tired person can often still recall facts while performing noticeably worse on multi-step problems.
On a test with no penalty for wrong answers, always guess rather than leaving a blank. Eliminate any options you can first: removing one takes your odds to one in three, and removing two makes it a coin flip.
Several points either way is normal. Test results carry a margin of error, which is why good reports give a confidence range rather than a single number. Differences of a few points between attempts are not meaningful.
Usually yes, but mostly through familiarity with the format rather than improved reasoning. This is known as the practice effect, and it makes repeat scores less informative about your actual ability.
Improvements from brain training tend to be specific to the trained tasks and rarely transfer to different kinds of problems. Getting better at an app mostly means getting better at that app.
Take the free IQ Centre test — 18 questions across 5 cognitive domains, with your score and percentile breakdown.
Take the Free IQ Test →