Can retaking an IQ test raise your score?

These are Table 2 results. The first comparison included 122 studies and 153,185 participants; later comparisons involved smaller subsets. They should not be treated as a single cohort's learning curve or extrapolated indefinitely. The outcome is standardized mean change across varied cognitive tests, not a direct prediction of your next IQ report. Scharfen et al., 2018

What can improve on a retest?

Familiarity with instructions, item formats, strategies and the testing situation can affect performance. The retest review examines moderators because the size of gains depends on the procedure and interval. A better second result does not isolate which process changed or prove a general intellectual increase. Scharfen et al., 2018

Matrix items illustrate the issue: solving them involves discovering and coordinating rules. Knowing a task family or remembering an item can alter what a later administration measures. Carpenter and colleagues' process analysis concerns those reasoning demands, while Raven's review emphasises the importance of the relevant norms. Carpenter et al., 1990 Raven, 2000

Can I subtract a fixed practice correction?

No universal personal correction follows from the pooled effect. Applying a group average to one person ignores variation in tests, exposure and retest conditions. Test-use guidance requires interpretation appropriate to the actual instrument and use; an improvised fixed subtraction does not establish validity. International Test Commission, 2001

Why converting the effect to IQ points is risky

Multiplying a standardized effect by 15 may look convenient, but the synthesis pools different cognitive outcomes and uses its own standardization. The common mean-100, SD-15 IQ convention does not make every effect in every study an equivalent personal IQ-point change. Preserve the effect metric rather than inventing a precise gain. Scharfen et al., 2018 Neisser et al., 1996

An established score report also includes measurement uncertainty. A difference between two reported values should be interpreted with the test's rules, interval and confidence information, not merely by comparing their labels. Pearson, 2020

Does a stable score solve the problem?

Reliability and practice are different issues. A task can have a consistent retest shift while preserving individual ranking, or show noisy individual differences even when its average effect is robust. The reliability-paradox literature is a reminder that these measurement properties have to be demonstrated separately. Hedge et al., 2017

Broader cognitive research also distinguishes different components of performance. A familiar matrix item set does not assess every aspect of thinking, so improvement on one task cannot establish equal change in memory, verbal ability and reasoning. Hampshire et al., 2012

What is realistic when tracking progress?

Record previous exposure, test version and interval. For any consequential assessment, follow its administrator's retest rules. For informal games, describe improvement as improvement on those tasks and keep repeated or memorised material separate from unfamiliar material. That preserves the useful skill gain without overstating what it proves. International Test Commission, 2001 Scharfen et al., 2018

IQ remains age-normed to 100 within age groups; retesting does not change that convention. The question is whether a particular new score validly represents the intended ability under its administration conditions. Neisser et al., 1996 Pearson, 2020