← Research library
ResearchModerate evidence

The study that finally included the right control, and measured mood with a single slider

Amd (2026) · PLoS ONE · Summary by Sentiosonics research agent
Year
2026
Stimulus
Theta
Design
Two randomized parallel-group studies; the second with pink-noise and silence controls
Sample
n = 219

Summary

For years the fair objection to every study of beats for calm has been that nobody separated the beat from the sound carrying it. Put beats in nature sounds, measure relaxation, and you have measured nature sounds. This study ran that missing comparison: 6 Hz beats inside pink noise against the same pink noise without them, with silence underneath both. The beats came out ahead. The outcome that says so is a nine-point slider.

Two studies sit in the paper. In the first, 101 undergraduates at the University of the South Pacific were randomly assigned to five minutes of 3, 6, 9 or 12 Hz beats on a 250 Hz carrier, rating their calmness, focus and happiness before and after. Only the 6 Hz theta condition raised both calmness and focus, at Hedges’ g above 0.84. The 9 and 12 Hz conditions raised calmness alone, at g above 0.72. In the second, 118 students were randomly assigned to 6 Hz beats, 6 Hz beats embedded in pink noise, pink noise alone, or silence. Both beat conditions raised calmness and focus, at g above 0.74; the two control conditions did not. Asked which they would rather listen to, participants preferred beats in pink noise to beats alone, and both to pink noise or silence.

Everything happened on the participants’ own devices, unsupervised, through a web page that randomized them and generated the tones in the browser. That is the point of the paper, and it is worth taking seriously: this is how people actually use beats, and almost no study tests them that way. The author published the data, the analysis script and the audio application itself.

The reason this does not settle much is the measurement. Calmness here is one slider, answered twice, five minutes apart, by someone who knows they have just taken part in a study about relaxing sounds. There is no cortisol, no heart rate, no skin conductance, no behavioural task, and no validated anxiety instrument. Five minutes is also short enough that the comparison is partly between four different experiences of sitting still with headphones on, and self-reported calm is the measure most exposed to expectancy of anything in this field.

We grade it moderate, which on our rubric means a controlled trial that has not been replicated, and the score behind it is 3.7. The design and the control deserve that; the outcome does not improve it. What this paper changes on our site is a reason, not a grade. Our scorecard has graded everyday stress in healthy people weak on the stated grounds that the studies lack the control condition needed to separate beats from pleasant audio. That is no longer true, so the scorecard now says what is actually missing: objective outcomes, exposures longer than five minutes, and anyone other than this author finding the same thing.

Limitations. Five minutes of exposure, which tells you nothing about a track played for an hour. Every quantitative outcome is a single-item self-report from an unblinded participant. The arms in Study 2 run from 22 to 34 people, so the comparisons are smaller than the headline total of 219 suggests. One author, one institution, one undergraduate population in Fiji, with the two studies drawn the same way, so the internal repeat of the 6 Hz result is not independent evidence. Participants were paid, in Study 2 five Fijian dollars, which is normal and worth knowing. And the paper’s own framing, that these protocols are “effective, accessible, and scalable”, runs ahead of a design that measured how calm people said they felt.

Our analysis

How we graded this study

Each criterion is scored 1–5. Study design and replication count double in the overall score (how we grade). Graded from: full text read.

Study designTwo randomized parallel-group studies. Study 1 assigned participants to 3, 6, 9 or 12 Hz beats. Study 2 assigned them to 6 Hz beats, 6 Hz beats embedded in pink noise, pink noise alone, or silence, which makes it a randomized controlled trial. The paper reports a power calculation, equivalence testing and manipulation checks.
Sample size219 analysed across the two studies (101 of 111 recruited in Study 1, 118 of 130 in Study 2), so the total falls in the rubric's n ≥ 100 band. Scored one point below that band because Study 2's four arms are 34, 34, 28 and 22, so every between-group comparison rests on groups in the 25-49 band. This is the same documented departure as Ahmadi 2021 and it is an open question for the editor rather than a settled rule.
Control conditionThe best-controlled study of beats for everyday calm that we hold. Comparing 6 Hz beats embedded in pink noise against pink noise alone isolates the beat from the sound carrying it, which is the comparison the rest of this literature omits, and a silence arm sits underneath both.
BlindingParticipants were not blind; beats, pink noise and silence are audible as different things, and the primary outcome is the participants' own rating. Scored one above the floor because the researcher was blind to condition allocation until analysis, the two raters coding the free-text answers were blind to condition, and the beats-in-pink-noise arm is hard for a participant to place.
Outcome measureSingle-item nine-point sliders rated before and after ("How CALM or STRESSED do you feel RIGHT NOW?"), plus matching items for focus and happiness. These are ad-hoc self-report rather than a validated anxiety or mood instrument, and there is no physiological or behavioural measure anywhere in either study.
ReplicationOne paper, one author, one university. The 6 Hz result does repeat across its own two studies, but in the same undergraduate population drawn by the same procedures, so it is internal consistency rather than independent replication. No independent replication found.
Conflicts of interestUniversity of the South Pacific research grant (AURC 02/2025/1.1.1) and a research stipend; no competing interests declared. De-identified data, the analysis script and the audio application itself are published on GitHub, which is more openness than anything else in this part of the library.
Peer review & venuePeer-reviewed and PubMed-indexed (PMID 41920802), open access.
Overall grade · 3.7 / 5Moderate evidence

Graded by our research agent on Nov 16, 2026.

Get updates from the field

A monthly letter: new research, graded, and the claims worth checking. No hype.

Monthly, one click to unsubscribe. See our privacy policy.