Does White Noise Help You Focus? It Depends Who You Are, and How Loud

White noise is one of the few study aids where the evidence is genuinely positive - and also one where the evidence says, unusually clearly, that it will not work for everyone.

Two findings carry this article. Volume matters, and it matters non-monotonically. And individual differences are large enough that a substantial minority get worse, not better.

Neither of those appears in the marketing for focus apps.

The volume finding

Awada, Becerik-Gerber, Lucas and Roll (2022) ran 39 neurotypical young adults through a within-subject design with randomised order, comparing white noise at 45 dB, white noise at 65 dB, and an ambient office baseline of about 42.3 dB.

Task	Best condition	Statistic
Sustained attention	45 dB (95.23% vs 93.12% ambient)	F(2,114) = 3.92, p = .02, d = 0.51
Working memory (2-back)	65 dB (66.38% vs 64.26%)	F(2,114) = 3.34, p = .04, d = 0.46
Creativity (remote associates)	45 dB (65.13% vs 50.77%)	F(2,114) = 3.89, p = .02, d = 0.51
Typing errors	45 dB (7.38 errors vs 10.77 at 65 dB)	F(2,114) = 5.13, p = .01, d = 0.59
Selective attention (Stroop)	No difference	F(2,114) = 0.49, p = .61

Read that table as a whole rather than picking a row. The louder condition helped working memory and hurt typing. The quieter condition helped sustained attention and creativity. And selective attention - the thing most people mean by “focus” - showed nothing at all.

So the honest answer to “does white noise help you focus” is: it depends what you mean by focus, and 20 decibels changes the answer.

Note the effect sizes are around d = 0.5, which is respectable, and the sample is 39 people in a within-subject design, which is modest. Treat these as real but imprecise.

The individual differences finding, which is the more important one

Söderlund and colleagues (2024) tested 43 children with a clinical ADHD diagnosis (mean age 14.3) under no noise, auditory white noise and visual pixel noise, within subjects.

The result was not “noise helps children with ADHD.” It was that which symptoms a child had predicted whether noise helped or hurt:

• Higher inattention predicted benefit from noise - auditory β = 1.369, p = .038; visual β = 1.105, p = .025
• Higher hyperactivity and impulsivity predicted harm - auditory β = -1.359, p = .034; visual β = -1.024, p = .033

And the headline the authors themselves flag: about one third of participants performed worse with noise, across both modalities. The effects were consistent within individuals across auditory and visual noise (r = .64, p < .001), which suggests a stable personal characteristic rather than measurement noise.

This is consistent with an inverted-U account - the Moderate Brain Arousal model - in which noise adds arousal, and adding arousal helps if you are under-aroused and hurts if you are already over-aroused. That framework predicts exactly the split observed: inattentive presentations benefit, hyperactive ones do not.

It also means the correct answer to whether white noise will help you is: try it and measure, because the literature cannot tell you in advance.

Why noise is not the same question as music

It is easy to fold this into the broader “background sound while studying” question. The evidence says not to.

Kattner and Meinhardt (2020) tested 50 participants on serial recall with white noise, speech and instrumental music. The disruption was domain-matched: instrumental music disrupted tonal recall more than noise (d_z = 0.45) and more than speech (d_z = 0.24), but for verbal serial recall, only speech was reliably disruptive (d_z = 0.19) and music was equivalent to noise.

In other words, structured sound with pitch content interferes with pitch memory; speech interferes with verbal memory; and unstructured noise interferes with neither in the same way. White noise has no changing structure to grab the mechanism that irrelevant speech grabs - which is the standard explanation for why it behaves differently, and why a colleague’s phone call is worse for reading than a fan.

A separate article on music and concentration covers that literature, where two meta-analyses disagree with each other and lyrics turn out to be the active ingredient. It publishes after this one, so the link will be added once that page is live. This article is about unstructured noise, and the two should not be merged.

The mechanism, and why it predicts the split

The standard explanation is stochastic resonance: in a system with a detection threshold, adding a moderate amount of random noise can push weak signals over that threshold, so a little noise improves detection while a lot degrades it. Applied to attention, the account is that noise adds neural arousal, and arousal has an optimum rather than a direction.

That framework - sometimes called the Moderate Brain Arousal model - makes a prediction that the data match unusually well. If people differ in where they sit on the arousal curve, then the same noise should help those below the optimum and hurt those above it. Söderlund’s split is exactly that shape: inattentive presentations, associated with under-arousal, benefited; hyperactive and impulsive presentations, associated with over-arousal, did not.

It also explains the volume result. Awada’s 45 dB and 65 dB conditions are two points on the same curve, and they produced opposite outcomes on different tasks - which is what you would expect if each task has its own optimum rather than a shared preference for quiet or loud.

Two cautions. This is a framework that fits the data, not a mechanism that has been demonstrated; the underlying neural account has been challenged in the recent literature and we could not retrieve that challenge to report it fairly. And an inverted-U model is very accommodating - it can absorb almost any pattern of results after the fact, which is a reason to hold it loosely.

What it does earn is a practical inference. If the effect is about your arousal rather than about noise having a property, then the right amount for you is not discoverable from a study. It is discoverable by paying attention to whether you are getting more done.

What this does not establish

Nothing about long-term use. Every study here is a single session in a laboratory. Nobody has run a trial of habitual white-noise use over weeks.

Nothing about sleep. White noise for sleep is a separate literature with its own, weaker, evidence base. This article is only about task performance while awake.

Nothing clinical. The ADHD study is a within-subject experiment on symptom dimensions, not a treatment trial, and nothing here is a recommendation for managing a diagnosed condition.

No dose-response curve. Two volumes were tested in one study. The claim that 45 dB is better than 65 dB for attention rests on a single comparison in 39 people, and the mechanism argument predicting an inverted U has not been mapped across the range.

The mechanism is contested. The stochastic-resonance account - that added noise pushes weak signals above a detection threshold - is the standard explanation, but it has been challenged in the recent literature, and we could not retrieve the challenge to report it fairly.

What follows

Three things you can actually act on.

Match the sound to the task. The one thing the volume study found consistently is that the same sound at the same level helped some tasks and hurt others. Louder noise helped working memory and produced more typing errors in the same participants in the same session. If you are doing something requiring accurate motor output, quieter is better.

Treat yourself as the experiment. A third of the ADHD sample got worse, and the predictor was a stable personal characteristic. That is an unusually strong result in favour of testing it on yourself rather than believing a study - including this one.

Do not expect it to fix selective attention. The Stroop task showed nothing, and Stroop is the closest thing in that study to “resisting a distraction.” White noise appears to work on arousal, not on filtering.

And if what you actually want is to stop hearing other people talk, the mechanism literature is clear that it is the speech doing the damage. Masking it with something unstructured is a reasonable response to a well-characterised problem - which is a narrower and better-supported claim than “white noise improves focus.”

Sound is one of several environmental variables people adjust before addressing the ones that matter more - see the screens, attention and focus guide for the wider picture, focus apps that actually block distraction for the software side, and studying for hours and remembering nothing for what the conditions are competing against.