How to Evaluate a Biohack and Run a Careful Personal Experiment

Seven questions to ask about any claim

1. Who says so? Check the person or company behind the claim, and look for scientific journal articles that support it and the expertise of the person posting. [1]
2. Who was studied? People, animals or cells? Healthy or ill? Findings from animals or cells are preclinical, not proof for humans; see reading early-stage research.
3. What was measured? A device reading or lab number is not the same as feeling or living better. Accuracy of a device is also a different question from whether using it improves health; see how to read a wearable validation study.
4. How is it being sold? Wording like "clinically proven" or "supports" can mean very little; see how to read health product marketing claims.
5. What is the evidence level? Use the labels in how we label evidence in biohacking.
6. What are the risks, and who should not do it? See evaluating safety claims.
7. What else could explain the result? See avoiding selective interpretation.

What a personal experiment is, and is not

A one-person experiment is often called an "n of 1." Formal N-of-1 trials are planned, repeated, single-patient crossover designs with systematic data collection on treatment effects. [2] A home experiment lacks most of those safeguards; see N-of-1 research: meaning and limits. Researchers note that experiments with serial observations of a single individual differ in scientific validity from randomized clinical trials. [3]

Reasons results mislead:

• Unusually high or low readings tend to be followed by more typical ones. [4] See regression to the mean.


• The placebo effect is a beneficial outcome resulting from a person's anticipation that an intervention will help. [5]


• With many numbers and flexible rules, a pattern can be found almost anywhere. [6]

A careful approach for gentle habits

For a regular bedtime, a daily walk or a food or mood diary, a careful approach is:

1. Ask one clear question. See defining a clear question.
2. Look at your usual pattern first. See establishing a baseline.
3. Change one low-risk thing at a time. See changing one low-risk variable at a time.
4. Note what else is going on such as illness, travel or stress. See recording context and confounders.
5. Judge over weeks, not days, and treat the answer as a hint rather than proof.
6. Stop if it makes you anxious or obsessive.

We did not find evidence for an ideal experiment length, so this page gives no durations.

Tracking can backfire

Clinicians have described patients seeking treatment for self-diagnosed sleep disturbances based on tracker data. [7] In a survey of 386 Australian women, use of diet and fitness apps was linked to disordered eating symptoms. [8] See tracking anxiety and compulsive measurement.

Evidence labels used here

Preclinical means cells or animals only. Established means supported by consistent, good-quality research. We found insufficient evidence that casual personal experiments can prove a habit works; they can show how you felt, which is useful but different.