IIF Certified Assessment
Start IQ Test
IIF Certified Assessment
The 1993 experiment behind three decades of "classical music makes babies smarter" ran for ten minutes, used college students, measured a single spatial task, and never mentioned IQ at all.
The Mozart effect is one of the few psychology findings to become state policy. It is also one of the most thoroughly misdescribed. The original result was narrow, short-lived and honestly reported. Almost everything attached to it since was invented downstream.
Frances Rauscher, Gordon Shaw and Katherine Ky, then at the University of California, Irvine, published a one-page research letter in Nature in 1993. Thirty-six college students each sat through three ten-minute conditions on separate occasions: a Mozart sonata for two pianos, a relaxation tape, and silence. After each one they completed a spatial-reasoning task taken from the Stanford-Binet — the kind where you work out what a folded and cut sheet of paper looks like once it is opened out.
Performance after the Mozart condition was better than after the other two. The authors expressed the difference as the equivalent of eight to nine points on that subtest's scale, and reported that it faded within about ten to fifteen minutes.
Read that again for what it is not. Thirty-six adults, not infants. One spatial task, not an intelligence test. Ten to fifteen minutes, not a lasting change. Eight to nine points on a single subtest's scale, which is a different object from a full-scale IQ score. The authors did not claim to have raised anyone's intelligence, because they had not measured it.
The jump from a ten-minute spatial task in undergraduates to a claim about infant brain development happened in the press and then in politics, without any new evidence appearing in between. By 1998 the governor of Georgia had proposed sending a classical music recording home with every newborn in the state, and Florida had adopted a requirement that state-funded early-childhood programmes play classical music daily. A publishing industry of recordings and toys followed.
None of that was in the paper. It is a clean example of a general pattern: a small, careful finding is restated one step more strongly at each retelling, and by the fourth or fifth step nobody is quoting the study any more — only the previous retelling.
The paper described a fifteen-minute change on one spatial task. The babies, the recordings and the legislation were all added later, by people who had not read it.
Attempts to reproduce the effect produced a mixed and eventually decisive picture. A 1999 meta-analysis in Nature by Christopher Chabris found the pooled effect small and consistent with what researchers call enjoyment arousal: anything that lifts alertness and mood for a few minutes nudges performance on a task attempted immediately afterwards. A separate 1999 failure to replicate appeared in the same journal from Kenneth Steele and colleagues.
The cleanest test came from Kristin Nantais and Glenn Schellenberg, also in 1999. They compared the Mozart sonata with a Stephen King story read aloud. Listeners who preferred the story showed the same boost after the story that Mozart listeners showed after Mozart. The predictor was preference, not the composer — which is precisely what an arousal-and-mood explanation predicts and precisely what a Mozart-specific explanation cannot accommodate.
A 2010 meta-analysis by Jakob Pietschnig, Martin Voracek and Anton Formann pooled around forty studies and found little evidence for a specific Mozart effect. It also found effect sizes noticeably larger in studies from laboratories affiliated with the original authors than in independent ones — a pattern that, wherever it shows up, argues for caution about the finding rather than for it.
One related claim did hold up, and it is not the one people repeat. In 2004 Glenn Schellenberg randomly assigned six-year-olds to a year of keyboard lessons, a year of voice lessons, a year of drama lessons, or no lessons, testing them before and after. The music groups gained a small amount more on a full IQ battery than the comparison groups — a few points, from a year of instruction, on a proper test rather than a single subtest.
That is a real finding and a completely different claim. Learning an instrument across a year involves sustained practice, instruction, attention and effort. Having a recording playing in the room involves none of them. The evidence supports lessons. It does not support playing music at a baby.
The general lesson is larger than the myth. A score means only as much as the conditions and the scale behind it, and a claimed change in a score means only as much as what was actually measured. If you want a number you can reason about, take a properly built assessment under proper conditions and read the result as a percentile on a named scale with a confidence range around it — not as a quantity that something playing in the room can move.
Find your IQ score now! →No. The original 1993 study found a short-lived improvement on one spatial-reasoning task, lasting about ten to fifteen minutes, in college students. It did not measure IQ, and replications indicate the boost comes from being more alert and in a better mood rather than from the music being Mozart.
A small, temporary performance boost after listening to something you enjoy is real and reasonably well established. A specific effect of Mozart is not: a 2010 meta-analysis of around forty studies found little support for it, and a 1999 experiment produced the same boost from a Stephen King story among people who preferred it.
There is no good evidence for it. The 1993 study involved thirty-six college students and no children at all. The infant claim was added by press coverage and by legislation, not by research.
Modestly, in one well-designed trial. A 2004 randomised study assigned six-year-olds to a year of keyboard or voice lessons, drama lessons, or nothing, and the music groups gained a few more IQ points than the comparison groups. That is a claim about a year of instruction, not about listening.
Corrections: spotted an error? Email corrections@iqmetrics.org and we will update this story and note the change here.
The numbers attached to famous names travel for decades without ever acquiring a test record, a date or a scale. Where the trail can be followed at all, it usually ends at a magazine paragraph.
A bad night does not lower intelligence. It reliably lowers the things a timed test happens to measure most closely — sustained attention, working memory and speed — while leaving vocabulary and general knowledge comparatively intact.
Analyses that compare siblings inside the same family do find a firstborn advantage on measured intelligence — roughly one to two points on a mean-100, standard-deviation-15 scale. It is real, its size is disputed, and it is no larger than the error on a single test.
Our IIF-certified assessment reports your score with its scale, percentile and confidence range — and a breakdown of the cognitive domains behind it.
Start IQ Test →
