How Stable Are IQ Scores Over Time? What Long-Term Studies Show
IQ scores become surprisingly stable by late adolescence. “Stable,” however, means that people tend to keep a similar position relative to their peers—not that every retest produces the same number.
IQ and other broad cognitive-ability scores show high long-term stability, especially from late adolescence through adulthood. A person who scores relatively high compared with peers at one point is likely to remain relatively high years later. Individual scores can still move, sometimes by enough to look dramatic on paper.
That distinction is the key to understanding the research. There are two different questions hiding inside “Is IQ stable?”: Does the group keep roughly the same ranking over time, and does one person receive nearly the same score every time?
What Researchers Mean by “Stable”
Most long-term studies focus on rank-order stability. Imagine 100 people taking an intelligence test today and again five years later. If the people near the top mostly remain near the top, and the people near the middle mostly remain near the middle, rank-order stability is high.
The exact numbers do not have to stay frozen for that to happen. Someone might move from 118 to 114 while another moves from 102 to 106. Their scores changed, yet their broad relative positions may remain similar.
How Stable Is IQ in Adults?
Very stable at the group level. A major 2024 meta-analysis combined 205 longitudinal studies, 87,408 participants, and 1,288 test-retest correlations. At age 20 with a five-year interval, the estimated rank-order stability across cognitive abilities was about .76, while general intelligence itself was about .80.
The same review found that stability rises rapidly through childhood and is consistently high from late adolescence through late adulthood. Longer gaps between tests reduce the correlation, although that decline becomes much flatter after roughly five years. Read the 2024 meta-analysis of cognitive-ability stability.
A correlation near .80 is strong. It means people’s relative positions are highly consistent across years. It still leaves room for individual movement.
Some People Really Do Move by 10 Points or More
A useful example comes from a study of 344 students who were assessed with the WISC-IV twice, an average of 2.84 years apart, during special-education reevaluations. The Full Scale IQ test-retest correlation was .82, which looks highly stable.
At the same time, 25% of those students had Full Scale IQ scores that differed by 10 points or more between assessments. Index scores moved even more often. The study therefore shows how a group can have a very high stability correlation while meaningful score changes still occur for individual people. The sample was a special-education population, so that percentage should not be treated as a general-population rate. See the WISC-IV long-term stability study.
This is also why a single score should be read with its confidence interval and the rest of the profile. Our guide to Full Scale IQ explains what the overall composite represents, while the IQ score chart shows how standardized scores relate to percentiles.
What Happens Across an Entire Lifetime?
Some of the most striking evidence comes from Scotland. Children born in 1921 and 1936 completed intelligence testing at age 11, and researchers later traced subsets of those cohorts into old age.
In the Lothian Birth Cohorts, scores on the same Moray House Test correlated .67 from age 11 to 70, .66 from age 11 to 79, and .51 from age 11 to 87. The correlations were higher after statistical correction for the restricted range of the follow-up samples. Those are remarkable levels of continuity across roughly six to seven decades. The Lothian cohort study is one of the clearest demonstrations that childhood cognitive differences can echo far into later life.
This does not mean an 87-year-old thinks exactly as they did at 11. The companion article on how IQ-related abilities change with age covers mean-level shifts in speed, reasoning, memory, and knowledge. Stability asks a different question: whether people tend to retain their relative standing compared with others.
Why Can the Same Person’s IQ Score Change?
Several things can move a retest score without requiring a fundamental transformation in intelligence.
- Measurement error: every psychological test has some imprecision, which is why professionally interpreted scores come with confidence intervals.
- Practice and familiarity: knowing the format can make a second attempt easier.
- Temporary state: fatigue, distraction, stress, illness, motivation, and testing conditions can affect performance.
- Real developmental or cognitive change: especially during childhood, after major life events, or when health changes affect cognition.
Practice effects are large enough to deserve special attention. In one small study of 54 young adults who repeated the WAIS-IV after three or six months, Full Scale IQ increased by about 7 points on average. The authors attributed the gains to prior exposure to the test rather than genuine intellectual change. See the WAIS-IV practice-effects study.
Is IQ Less Stable in Children?
Yes, especially in the early years. The 2024 meta-analysis found low rank-order stability in preschool children, followed by a rapid rise through childhood. By late adolescence, stability had reached the high levels that characterize adulthood.
That makes intuitive developmental sense. Young children are changing quickly, test formats have to change with age, and early measures capture abilities that are still organizing. An IQ score at age 4 therefore carries a different long-term meaning from an adult score at age 30.
Does a High IQ Guarantee You Stay High?
No individual score comes with that guarantee. High rank-order stability is a probability statement about groups. It says that substantial reshuffling becomes less common as people mature.
A major individual change can still occur. The 2024 review specifically notes that serious illness, neurological trauma, major psychosocial stress, and large educational or social changes can justify reassessment sooner than population-level stability estimates would suggest.
What Should You Make of a Retest Difference?
A small difference is usually more interesting in context than by itself. Was the same test used? Was the retest soon enough for familiarity to matter? Were conditions comparable? Does the change exceed the test’s expected measurement error? Did one index move while the overall score stayed similar?
Those questions are more useful than treating either score as the permanent “true” number. If you are looking at a clinical assessment, interpretation belongs with the psychologist and the test’s technical guidance. For ordinary curiosity, online scores are better treated as snapshots of performance on the specific tasks used.
Try an IQ-Style Snapshot
Cognitive Train’s IQ-Style Test uses letter series, number series, analogies, matrix reasoning, mental rotation, and cube folding. It reports a score on its own IQ-style scale plus a six-skill breakdown. It is useful for seeing how your performance can vary across reasoning formats; it does not provide a clinical IQ score.
The larger lesson from the research is simple: intelligence scores become highly stable in relative terms, while individual numbers retain some movement. For the basic meaning of the construct, start with what IQ is. You can also explore our brain tests and cognitive assessments or browse Cognitive Train’s brain training and cognitive training tools across reasoning, memory, attention, and speed.