What would happen if 100 people encountered the same unfamiliar song, but only half could see who made it and how many followers the artist had? Would the recording receive similar responses, or would the information around it change the listening experience?
For an independent musician, that question matters. A song usually arrives with a name, an image, a recommendation or a visible measure of attention. Separating the recording from some of that context could help us understand what influences its first reception.
Study status: this article describes a proposed experiment. It does not report results from 100 listeners. The recording, recruitment arrangements and final study materials must be confirmed before the experiment can begin.
One recording, two ways of presenting it
The proposal is to recruit 100 consenting adult volunteers and randomly assign 50 to each of two groups. Every participant would hear the same complete recording using the same player and instructions.
| Group | What the participant would hear | What they would see before rating it |
|---|---|---|
| Blind presentation: 50 assigned listeners | The selected recording in full | A neutral label, such as “Recording 01” |
| Context presentation: 50 assigned listeners | The identical recording in full | The artist’s name, track title and one verified public follower count, with its platform and checking date |
The follower count would be genuine and fixed before the study. There would be no invented streaming totals, fabricated awards or suggestion that previous participants had already approved the song. Both groups would use the same visual layout, without artwork, reviews or additional promotional copy.
This design compares the effect of the information package as a whole. If responses differed, it could not tell us how much of that difference came from the artist’s name, the title or the follower count. Separating those influences would require a different experiment.
Why each person would hear only one version
Playing a song anonymously and then revealing the artist to the same listener sounds simple. It also changes two things at once: the information available and the listener’s familiarity with the recording. The second rating would follow an earlier encounter.
Using separate groups avoids that particular complication. Random assignment would help make the groups comparable before listening, although a sample of this size could still contain chance imbalances.
The audio would remain constant. We would not choose a more exciting excerpt for one group, change the volume between conditions or vary which passage people heard. Participants would receive the same request to listen at a comfortable level and give the recording their attention.
What we would ask listeners
The main question would be deliberately direct: “How much did you enjoy this recording?” Participants would answer on a scale from 1, “Not at all”, to 7, “Very much”.
A separate question would ask whether they would choose to hear it again: yes, no or unsure. This measures an intention. It would not be reported as proof that someone returned to the artist later, became a fan or purchased anything.
Participants could also leave an optional comment about what stood out. Permission to quote that comment publicly would be a separate, optional choice. Declining quotation would not prevent someone from taking part.
After the main ratings, we would ask about previous familiarity with the recording and, after revealing the artist to everyone, familiarity with the artist. Those answers would help describe the sample. We would not quietly remove inconvenient ratings on the basis of those responses.
Who would count as a stranger?
The proposed recruitment would seek adult volunteers from general listening communities, rather than the selected artist’s fan list or personal network. The artist’s identity would stay out of recruitment materials.
That still would not guarantee that every participant had never encountered the music. Recognition could occur, and it would need to be reported. “Strangers” would describe the intended recruitment approach, not an assumption that familiarity had been eliminated perfectly.
We would publish the actual participant numbers. If 100 people were assigned but fewer submitted ratings, the report would distinguish those counts and show completion in each group. The headline could not honestly say that 100 people judged the song unless 100 people actually did.
Consent would come before listening
Participants would receive a short explanation of who is organising the study, what they would do, how their responses would be used and how to withdraw within the stated period. Participation would be voluntary and limited to adults aged 18 or over.
The proposed report would publish group results. Comments would appear only with separate permission and after removing identifying details. The actual survey service, data arrangements and withdrawal deadline would need to be settled before that information could be finalised.
The artist and relevant rights holders would also need to authorise use of the recording in the study and the planned public report. The test should be able to produce an unfavourable or inconclusive result without changing which findings are reported.
What earlier research gives us a reason to investigate
In a 2006 study published in Science, Matthew Salganik, Peter Dodds and Duncan Watts created an experimental music market with 14,341 participants. Making earlier listeners’ choices visible increased inequality in song success and made outcomes less predictable across the experimental groups. [1]
That finding provides a reason to investigate the information surrounding music. It does not predict the outcome of this proposed test: an artist’s follower count is a different cue from a song’s accumulated downloads, and rating one recording differs from choosing among many songs.
How we would decide what the results mean
Before recruitment, the protocol would specify the main outcome, group assignment, stopping rule and analysis. Registering the completed plan in advance would make it easier to distinguish planned analysis from questions prompted by the data. This is the purpose of preregistration described by the Center for Open Science. The proposal is not yet a registered study. [2]
The main comparison would be the difference in average enjoyment ratings between the context and blind groups, accompanied by an uncertainty interval. The report would also show the distribution of ratings, so disagreement would remain visible instead of disappearing into one average.
A higher average in the context group would suggest that this particular presentation increased reported enjoyment under these conditions. A lower average would suggest the opposite. A result close to zero, especially with substantial uncertainty, could leave the effect unresolved. It would not establish that identity and popularity never matter.
The target of 100 participants makes this a practical pilot. It is not a guarantee of detecting a small difference. One song and a volunteer sample could not establish how all genres, audiences or discovery platforms behave.
How this relates to Kallitone
Kallitone’s discovery experience gives listeners a role in comparing unfamiliar recordings. Its standard voting explanation uses randomly selected passages and a minimum of 15 seconds from each side, with votes contributing to rankings. [3]
The proposed experiment would use a different setup: one complete recording, individual enjoyment ratings and controlled presentation conditions. It could investigate a question relevant to blind listening, but its results would not automatically validate Kallitone’s pairings, rankings or rewards.
The useful outcome would be a documented account of how real people responded, including disagreement and uncertainty. Until those people have listened, the result remains an open question.
Sources
- Salganik, M. J., Dodds, P. S. and Watts, D. J. (2006). Experimental Study of Inequality and Unpredictability in an Artificial Cultural Market. Science, 311(5762), 854–856. DOI: 10.1126/science.1121066.
- Center for Open Science. Preregistration.
- Kallitone About and Explore — product descriptions, separate from the proposed research.
Study proposal prepared on 13 October 2026. No listener results, participant quotations or measured effects are reported in this article.