Question
Are IQ tests reliable?
Short answer
Clinical IQ tests are among the most reliable psychological measures. The WAIS-IV Full Scale IQ showed internal consistency of .98 and a test-retest correlation of .96 after about three weeks; the WAIS-5 FSIQ averages .97, with .93 stability. Narrower index and subtest scores are less reliable, and online tests vary widely.
Want your own number? The test is 25 visual puzzles, about 12 minutes, no sign-up.
What reliability means
The APA Dictionary of Psychology defines reliability as the consistency of a measure: how free a test is of random error, so that it gives the same results across repeated applications (APA Dictionary, reliability). Two kinds matter most for IQ tests. Internal consistency asks whether the items hang together. Test-retest reliability is the correlation between scores from a first and second administration of the same test (APA Dictionary, retest reliability).
The numbers for clinical tests
| Score | Internal consistency | Retest |
|---|---|---|
| WAIS-IV Full Scale IQ | .98 | .96 (mean 22 days, 298 people) |
| WAIS-5 Full Scale IQ | .97 (average across ages) | .93 (mean 29 days, 201 people) |
| WISC-V Full Scale IQ, clinical sample | - | .86 (about 2.6 years) |
Sources: Collinson et al. for the WAIS-IV (Collinson et al. 2016), Winter et al. for the WAIS-5 (Winter et al. 2024) and Watkins et al. for the WISC-V (Watkins et al. 2022). Coefficients above .90 are high for any psychological measure.
Try one · Logical Reasoning
Which tile completes the second pair?
Tap the option you think is right, or show the answer.
Show the answer
The missing panel shows a small outlined triangle pointing right on its own (option 2).
How to solve it
- Compare the first two panels and name exactly what changed.
- The whole figure is flipped left to right, like a mirror image.
- The large shape becomes outlined.
- Nothing else changes.
- Apply the same changes to the third panel: a small hatched triangle pointing left on its own.
- So the missing panel shows a small outlined triangle pointing right on its own.
That was 1 puzzle. The full test has 25.
Where reliability drops
Full-scale scores pool many subtests, which is why they are the most consistent part of a report. Narrower scores are noisier. Watkins and colleagues note that of the 15 possible WISC-V scores, only the Verbal Comprehension Index and the Full Scale IQ met the dual standard of at least .90 for internal consistency and short-term retest reliability (Watkins et al. 2022). Read single index or subtest scores with more caution than the overall score.
Reliable still means a margin of error
Even a very reliable score is an estimate. Clinical manuals report a confidence interval, often around plus or minus 5 points at 95% (Wechsler, 2008). Retests also run a little high from practice: on the WAIS-5 retest study the average Full Scale IQ gain was 3.7 points for ages 16 to 69 (Winter et al. 2024). See how accurate an IQ score is and can IQ tests be wrong.
And online tests?
Published reliability figures like these come from large norm studies of clinical tests. Many online quizzes publish nothing. A careful online test should at least explain its scoring and report a range rather than one exact number. See are online IQ tests accurate and the reliability glossary entry.
Sources
- American Psychological Association, APA Dictionary of Psychology, entry "reliability" dictionary.apa.org/reliability
- American Psychological Association, APA Dictionary of Psychology, entry "retest reliability" dictionary.apa.org/retest-reliability
- Collinson, Evans, Wheeler et al., Confirmatory Factor Analysis of WAIS-IV in a Clinical Sample: Examining a Bi-Factor Model, Journal of Intelligence 5(1):2, 2016 doi.org/10.3390/jintelligence5010002
- Winter, Trudel, Kaufman, Bratsberg et al., Wait, Where's the Flynn Effect on the WAIS-5?, Journal of Intelligence 12(11):118, November 2024 doi.org/10.3390/jintelligence12110118
- Watkins, Canivez, Dombrowski et al., Long-term stability of Wechsler Intelligence Scale for Children, fifth edition scores in a clinical sample, Applied Neuropsychology: Child, 2022 doi.org/10.1080/21622965.2021.1875827
- Wechsler, D. (2008). WAIS-IV Technical and Interpretive Manual. Pearson.