Does music improve exercise performance?
Across 139 studies, music produced a small average performance gain and made exercise feel more positive; a newer preference review found clearer strength and power effects than speed or aerobic endurance.
Music can improve exercise performance, but the most reliable answer is smaller
and more useful than the usual perfect playlist claim.
Across 139 studies, music produced a small average improvement in physical performance. Its larger effects were psychological: exercise felt more positive, and perceived effort shifted modestly. A newer review of preferred music found clearer gains for strength endurance, maximal strength and power than for speed or aerobic endurance.
That pattern makes music a context tool, not a performance guarantee. What you listen to, when you hear it, what the task requires and what the comparison is can all change the result.
The average performance effect was small
The broadest quantitative anchor is a 2020 Psychological Bulletin meta-analysis of 139 exercise and sport studies. It included 3,599 participants and 598 effect sizes.
Music was associated with a standardized performance effect of g = 0.31
(95% CI [0.25, 0.36]). That is a small average difference across many kinds of
outcomes. It is not a 31% boost, and it does not mean every participant performed
better. The authors also detected possible publication bias for performance and
said this estimate may be slightly inflated.
The effect was larger in exercise studies than sport studies and with fast rather than slow-to-medium tempo. Competition, tactics, rules and divided attention may leave less room for music to change output than a controlled exercise test.
The average also conceals nulls. Some studies changed motivation or perceived effort without changing time, power or work. Others found an output effect in one interval or set that did not persist across the full task.
Feeling better was the more consistent outcome
In the same meta-analysis, the effect on affective valence was g = 0.48
(95% CI [0.39, 0.56]), larger than the physical-performance effect. Perceived
exertion shifted by g = 0.22 (95% CI [0.14, 0.30]).
Music was also associated with a small beneficially coded oxygen-consumption,
or physiological-efficiency, effect, but not a statistically significant change
in heart rate. The body did not move as one single music response.
This separation matters. A session can feel better without becoming faster. It can feel equally hard while producing slightly more work. Motivation can rise without changing peak power. If those outcomes are collapsed into one claim, almost any study can be made to sound positive.
The practical value may therefore begin before a measurable performance gain. A more positive session is meaningful, but the acute studies do not establish that music produces long-term adherence or training adaptation.
Preferred music helps some outputs more than others
A 2025 meta-analysis updated the question by comparing preferred music with non-preferred music and no music. It screened 3,146 records and included 41 studies of adolescent and adult athletes.
The direction of the pooled results favored preferred music for motivation, positive affect, perceived exertion, strength endurance, maximal strength and power output. There was no pooled advantage for speed or aerobic endurance. Exact estimates are not repeated here because the version of record reports materially inconsistent values between its abstract, results and sensitivity analyses.
The review also reported high heterogeneity, and comparisons specifically
against non-preferred music were less consistent. Its bias analyses also
indicated possible inflation for motivation, strength endurance, power output
and aerobic endurance. Preferred is not a precise acoustic treatment: it can
bundle familiarity, identity, memory, expectation, tempo and perceived control.
A separate 2025 review of 12 small studies found no significant preferred-music effect on mean heart rate or perceived exertion. Those are internal-load outcomes, not a comprehensive test of objective performance, but they narrow confidence in any universal psychological or physiological benefit.
Small studies show why one playlist cannot fit every task
A bench-press crossover in 12 resistance-trained young men found more repetitions and higher early velocity and power with preferred rather than non-preferred music. That is an interesting within-person result, but it is a narrow sample and does not compare music with silence.
A 2,000-m rowing study in active women found a different split. Preferred volume increased motivation and lowered perceived exertion, but it did not change time, relative power or heart rate.
A repeated-sprint cycling study in 14 active men found the same psychological pattern without a performance gain: preferred music raised motivation and lowered exertion, while mean power, anaerobic capacity and total work remained unchanged.
These studies do not cancel one another. They show that music may change the state around performance more reliably than it changes every output.
Tempo is a moderator, not a universal prescription
Fast music outperformed slow-to-medium music on average in the 2020 synthesis. That does not create one best BPM.
Rhythmic entrainment may help in repeated movement, while very fast or loud music can be irrelevant or distracting in a technical task. Preferred songs may increase arousal before a maximal effort, yet a calmer track may better support pacing in a long session. Music during the task, during warm-up and between rounds are different conditions worth testing; the newer review did not find a significant timing advantage.
The listener is part of the condition. A researcher-selected track that one participant finds motivating may be noise to another. A song’s personal meaning can change even when tempo and volume are identical.
What the evidence supports
| Claim | Current answer |
|---|---|
| Can music improve acute physical output? | Yes, by a small amount on average |
| Does it work for every task and person? | No |
| Are psychological effects larger? | Often; affect and motivation show clearer shifts |
| Does preferred music always beat no music? | No; results vary by outcome and comparison |
| Does fast music always work best? | No; fast tempo is an average moderator, not a rule |
| Does an acute benefit improve long-term fitness? | Not established by these studies |
The strongest use is therefore experimental and specific. Define the output you care about—repetitions, pace, power, time or simply session experience—then hold the task stable and see whether the same music condition repeatedly moves it.
Original sources
- Terry et al., 2020 meta-analysis
- Niering et al., 2025 preferred-music meta-analysis
- Herodek et al., 2025 internal-load meta-analysis
- Ballmann et al., resistance exercise
- Preferred-volume rowing crossover
Independent editorial summary. The authors are not affiliated with LifeScore.
The lifescore take
What matters is not whether music `works`, but which part of the performance system it moves. The average physical gain is small; the shift in affect and motivation is larger. That matters because a better-feeling session and a better result are related but not interchangeable. Preference, tempo and task fit can be treated as variables instead of playlist folklore; timing defines another test condition, not an established universal advantage. What this unlocks is a cleaner personal experiment. Measure one output across repeated comparable sessions, preserve silence or a stable control condition and keep a song only if the benefit survives the excitement of novelty.
Primary source
Terry et al., 2020 meta-analysis
DOI 10.1037/bul0000216.
Independent editorial summary. The authors are not affiliated with LifeScore.
Edge delivery
Get new Edge articles in your inbox.
Evidence-led insights, delivered when published.