What is an n-of-1 experiment?
An n-of-1 experiment is a trial with a single participant, where that person acts as their own control by alternating between periods with and without the thing being tested. Because the comparison is one person against themselves rather than against a group average, it answers a different question from a standard trial: not "does this work on average" but "does this work for me". It is an established clinical method — the US Agency for Healthcare Research and Quality publishes a design guide for running n-of-1 trials on individual patients.
Why one person can be enough
A conventional trial randomises hundreds of people so that everything except the treatment averages out between the groups. An n-of-1 trial gets the same protection differently: the two groups are the same person in two time periods, so genetics, history, personality and circumstance are held constant by construction rather than by averaging.
The cost is that the result applies to one person. That is a real limitation for medicine and not a limitation at all for the question "should I keep doing this".
The standard shape
The classic design alternates blocks: baseline, intervention, baseline, intervention. Clinical versions randomise the order of the blocks and blind the patient to which one they are in, usually with a placebo. Both features exist to stop expectation from producing the result.
A version run on yourself cannot blind you. That is the honest weakness, and it means a self-run n-of-1 is best at questions where expectation is a weak driver — sleep duration, resting heart rate, whether you actually finished the work — and worst at questions that are mostly self-report about how good you feel.
What it can and cannot settle
- Can settle: whether a change you can start and stop tracks a measure you can take repeatedly.
- Can settle: that something you were sure about does nothing. This is the most valuable output and the one people never look for.
- Cannot settle: anything with a long lag, a permanent effect, or a one-shot outcome. You cannot alternate blocks on a decision you can only make once.
- Cannot settle: whether the result generalises to anyone else. It was never trying to.
Where Emerge fits
Emerge runs this design as Your Lab — baseline, then intervention, then a verdict based on what the numbers did rather than on how the period felt in retrospect.
Emerge on the App Store →Related questions
Is an n-of-1 trial real science?
Yes. It is a recognised clinical trial design with published methodology and statistical guidance, used when a treatment decision has to be made for one patient and group averages do not settle it.
What is the difference between n-of-1 and just tracking?
Tracking records what happened. An n-of-1 experiment deliberately changes one thing and holds the rest still, so the record can support a claim about cause. Tracking alone cannot separate the effect of the change from everything else that moved that month.
Sources
Last updated 2026-09.