Music and Creativity: What the Research Says
Music and creativity have an intuitive connection that most people act on without examining. You put on a playlist, assume it helps, and keep working. The research is more specific — and more useful — than that intuition suggests.
The basic finding is this: music affects creative thinking, but the direction of that effect depends on what type of thinking you're doing, what kind of music is playing, and whether you're a trained musician. Getting those variables wrong doesn't just fail to help. It actively impairs performance.
The Ambient Noise Effect
In 2012, Ravi Mehta, Rui Zhu, and Amar Cheema published a study in the Journal of Consumer Research that has since become the empirical basis for the "work in a coffee shop" movement. They tested creative task performance across four noise levels: 50 dB (near-silent), 70 dB (typical coffee shop), 85 dB (loud), and silence. Performance on creative divergent tasks peaked at 70 dB — the moderate ambient noise condition.
The mechanism is not that noise aids concentration. It's the opposite: moderate ambient noise increases arousal and slightly disrupts focused processing, which forces the mind toward more abstract thinking. Low noise allows hyper-focused detail processing, which is efficient for analytical work but suppresses the loose, broad attention that divergent thinking requires. Too much noise overwhelms cognitive resources entirely.
This study describes a specific sweet spot. 70 dB of diffuse, non-verbal ambient sound — the kind found in a moderately busy café — produces better creative output than silence for tasks that require flexible, associative thinking. Tasks requiring precision and accuracy have the opposite relationship: silence wins.
Why Lyrics Hurt Focused Creative Work
Nick Perham's research at Cardiff Metropolitan University provides the clearest explanation for why music with lyrics impairs reading comprehension and writing tasks. The phonological loop — the component of working memory responsible for language processing — can only hold one stream of verbal information at a time. When song lyrics occupy that loop, there's no capacity left for the verbal processing the task requires.
This is not about distraction in a general sense. Perham's participants working in non-lyrical steady-state noise (like a fan) showed no impairment, while those working with vocal music showed significant decrements in verbal recall and reading comprehension. The impairment was phonological, not attentional.
The practical implication is direct: for tasks involving reading, writing, or verbal reasoning, vocal music costs working memory bandwidth. Instrumental music without lyrics — ambient electronic, classical, or anything without recognizable phonological content — doesn't carry the same cost.
Mood and Creative Breadth
Simone Ritter and Sam Ferguson's 2017 study in PLOS ONE examined how music valence (positive versus negative emotional tone) affected divergent thinking. Participants listening to music with positive valence — happy, major-key, upbeat — produced significantly more original and varied responses on a divergent thinking task than participants in silence or listening to minor-key music. Neutral or sad music produced no benefit.
This finding connects to Barbara Fredrickson's broaden-and-build theory of positive emotions, which shows that positive affect systematically broadens attentional scope. When attention is broadened, more concepts become simultaneously active, and the probability of making remote connections increases. Happy music, to the extent that it reliably induces positive affect, recruits this broadening mechanism.
The effect is specific to divergent tasks. For analytical tasks requiring convergent evaluation — narrowing a field of options to identify a single correct answer — positive affect doesn't help and can modestly impair precision by encouraging excessive exploration. The creative process involves both modes, and the music appropriate for each phase differs.
The Expertise Paradox
Teresa Lesiuk's 2005 study in Psychology of Music found that music improved both quality of work and time-on-task among software developers — but only for developers with moderate musical training. Novices showed no significant benefit, and the highly trained musicians showed reduced performance.
The explanation is habituation and attentional load. For someone with limited musical training, music requires active processing: it's genuinely engaging, stimulating, and slightly arousing in ways that sharpen performance. For highly trained musicians, familiar musical structures become so absorbing that they compete with the task rather than complementing it. The person who studied piano for twelve years is not getting ambient noise from Bach; they're getting a second task.
This is why boredom and creativity research is relevant here: moderate stimulation is often better than either understimulation or overstimulation. Music calibrated to produce mild arousal without active engagement is the target, and that calibration depends on how much cognitive load you bring to the music itself.
What This Looks Like in Practice
The research converges on a set of specific conditions:
For divergent or generative tasks (brainstorming, free writing, exploring possibilities): 70 dB of non-vocal ambient noise or upbeat instrumental music with positive valence. Explicit major-key music appears to work better than neutral ambient noise for originality of output.
For convergent or detail-oriented tasks (editing, analysis, coding, evaluation): silence or white/brown noise. Lyrical music impairs verbal processing and should be removed entirely.
For insight tasks (the kind where you're stuck and need a reframe): low-attention music or ambient sound that prevents rumination without demanding focus. The incubation effect depends on reduced directed attention; music that is engaging enough to prevent active problem-repetition but not demanding enough to block associative processing occupies a useful middle ground.
Musicians should calibrate down. If you have significant musical training, you need music that is less complex, less structured, or less personally meaningful to achieve the same mild arousal that moderate complexity provides for someone without that training.
A playlist built for creative work is not a single playlist. It's at minimum two: one for generating, one for evaluating. The common mistake is to use the same music throughout, which means you're either suppressing divergent thinking with silence during the generative phase or suppressing precision with stimulating music during evaluation.
Ready to train your creativity? Try science-backed exercises that measure and improve your creative thinking. Start a Free Exercise