Changing One Low-Risk Variable at a Time

The idea in one sentence

If you change several things together, you cannot tell which one mattered. Changing one low-risk thing at a time gives a comparison a chance of meaning something. This guide is information only; it describes the reasoning, not a program, and it covers ordinary, reversible habits only.

Why one at a time

Suppose you start a calmer evening routine, change your caffeine habits and take up evening walks in the same week, and you sleep better. Any of the three, a combination, or none of them may explain it. Researchers describe the same problem as confounding: a variable that is linked to both the exposure and the outcome can obscure the "real" effect of the exposure on the outcome. [1] In a personal experiment you are the only subject, so every other change in your life is a possible confounder. See recording context and confounding factors.

A formal single-person design compares periods with an intervention against periods with a comparator, and even then can rest on strong assumptions about how the data arose. [2] Keeping the change simple makes those assumptions easier to defend.

What counts as "low-risk"

A change is a reasonable candidate only if all of these are true:

• It is an ordinary habit that many people do without medical supervision.
• You can stop at any time and go back to how things were.
• It does not involve a medicine, supplement, unapproved compound, injection, implant, device that stimulates the body, long fasting or extreme heat, cold or breath control.
• It is not a substitute for care of a symptom or condition.

If it fails any of these, do not run it as a personal experiment. See when self-experimentation is inappropriate and experimental approaches and self-experimentation.

Example 1: a bedtime routine

The question might be whether a calm wind-down before bed changes how rested you feel in the morning compared with your usual evenings. The one thing you change is the wind-down. You try to leave other things alone as far as your life allows: wake-up time, meals, screens and other habits. The NHS lists relaxing before bed and getting up at the same time every day among habits that can help with insomnia. [3] Because getting up at the same time is itself a change for many people, treat it as a separate experiment rather than adding it to the wind-down.

Example 2: a comfortable walk

The question might be whether a short, comfortable walk changes how you rate your afternoon energy compared with days without one. The NHS describes brisk walking as a moderate-intensity activity, and says one way to judge moderate intensity is whether you can still talk but not sing. [4] Here the point is comfort, not performance. If walking causes chest pain, dizziness or unusual breathlessness, stop and see a clinician. People with heart, lung or joint conditions should check with a clinician before changing activity.

Practical limits

• Not everything can be isolated. Your sleep partner, weather, work stress and illness change on their own. Accept that results will be messy.
• Wait before judging. A change followed by a single good night is weak evidence; see regression to the mean in personal data.
• Beware expectation. Knowing what you are testing can shape how you feel. In studies of perceived sleep, people told they had slept above or below average performed differently on cognitive tests accordingly. [5]
• Do not chase. Stopping an experiment is a fine result. If tracking leads to worry or perfectionism about numbers, a pause may help; see when taking a break from tracking may help.