The honest summary of decades of research on background music is that the average effect on performance is close to zero, and the average hides everything interesting. Music helps some tasks, hurts others, and the deciding factor is usually whether the music competes for the same mental channel as the work.
The claim to drop first
The Mozart effect, as popularly understood, is not real. The 1993 study found a small, temporary improvement on one narrow spatial reasoning task after listening to Mozart, lasting on the order of ten to fifteen minutes. It said nothing about intelligence, studying, or babies.
Pietschnig and colleagues' meta-analysis of dozens of studies found the effect to be small at best, and notably larger in studies published by labs affiliated with the original authors. The best explanation left standing is arousal and mood: pleasant, energetic music makes you slightly more alert, and slightly more alert people do slightly better on tests for a short while. Any music you enjoy does that. There is nothing special about Mozart.
What reliably hurts: words
This is the finding that survives replication and is worth actually changing your behaviour over.
Background speech and lyrics interfere with verbal work. Salamé and Baddeley's experiments showed that vocal music disrupted short-term recall of items substantially more than instrumental music did, and that unattended speech impairs verbal short-term memory even when participants are told to ignore it.
The mechanism is straightforward: reading, writing, and remembering words all run through a phonological channel, and lyrics occupy that channel whether you attend to them or not. You cannot ignore language by deciding to.
The practical rule follows directly:
| Task | Lyrics | Instrumental |
|---|---|---|
| Reading, writing, studying text | Bad, consistently | Usually fine |
| Learning verbal material | Bad | Neutral to mildly negative |
| Repetitive or mechanical work | Fine, often helps | Fine |
| Physical or routine tasks | Helps, mostly via mood and arousal | Helps |
If you are writing an essay to your favourite album, the album is costing you something. Not much, but reliably.
What music is usually doing for you
Kämpfe and colleagues' meta-analysis found no overall effect of background music on performance, but did find specific effects: negative for reading and memory tasks, positive for emotional reactions and sports performance.
That pattern points at the real mechanism. For most people, music at work is not enhancing cognition. It is doing one of three jobs:
- Masking worse noise. An open-plan office, a café, a housemate on a call. Here music wins easily, because unpredictable speech is more disruptive than steady sound. You are choosing the lesser interference.
- Regulating mood and arousal. Getting started on something you are avoiding, or staying awake at hour three. This is real and useful, and it is about willingness rather than capacity.
- Marking the start of work. A consistent album or playlist becomes a cue that says the session has begun. That is a habit mechanism, not an acoustic one, and it works.
Notice that only the first is about the music itself. If noise is your problem, silence or steady non-musical sound usually beats music, since you get the masking without the competition.
Choosing what to play, if you play something
The consistent findings point one direction: the less the music demands, the less it costs.
- No lyrics for anything involving words. In a language you do not speak is a real workaround, though not a free one.
- Familiar over new. Novelty pulls attention. A record you have heard two hundred times is nearly furniture.
- Steady over dynamic. Sudden changes in tempo or volume are what break concentration. This is why ambient, minimal, and film scores keep getting recommended.
- Moderate volume. Loud music impairs complex task performance fairly consistently, and the effect grows with task difficulty.
- Steady noise instead, if you only need masking. Rain, fans, brown noise. No melody to follow, no structure to anticipate.
The introvert-extrovert difference here is one of the more robust individual differences in the literature: people higher in extraversion tend to be less impaired by background music on complex tasks. Which is a research-grade way of saying that a friend who genuinely works fine with music on is probably telling the truth about themselves and giving you bad advice.
The test that beats every recommendation
Nothing here overrides your own data, and the effects are small enough that self-report is unreliable. Run it properly for one week.
Same task, same time of day. Three days with your usual music, three days without. Record something you can actually count at the end of each session: pages read, words written, problems finished, or a one-to-five rating of how hard it felt to stay on task.
Most people who run this discover one of two things. Either the difference is negligible, in which case play whatever you like and stop reading articles about it. Or the difference is obvious, in which case you now know, and it is worth the mild inconvenience of changing your habit.
Log it as a note rather than trusting your impression at the end of the week. Retrospective judgements about your own focus are exactly the thing that the illusion of competence shows we are bad at.
Where music fits in a focus system
Music is a small lever. The big ones are the number of things competing for your attention and how often you switch between them, which is why context switching and attention residue matter more than any playlist.
If focus is genuinely the problem, start with structure: time blocking for when you work, the Pomodoro technique for sessions you keep abandoning, and deep work for the conditions that make hard thinking possible at all. Then use music for the job it is actually good at, which is masking noise and getting you into the chair.
More on the mechanics of attention in memory and focus and cognitive load theory.