Open Source Apothecary At Home / Foundations

Regression to the mean: why you always feel better

The statistical trap that makes almost any remedy look effective, including the ones that do nothing.

guide

People take up a remedy at the worst point of a fluctuating symptom. A symptom that fluctuates is, by definition, unusually bad when you notice it, and unusually bad is statistically followed by less bad whether or not the remedy did anything. This is regression to the mean, and it is large enough that a preparation can acquire an enthusiastic following for a condition it does nothing for.

The practical consequences for a household are three:

• Starting a remedy at the peak guarantees a better week. The peak is exactly when the
• Stopping a treatment at the trough guarantees a worse week. Which is how people
• The honest test is a fixed period with a written tally, started at a stable point, not

A remedy trial that begins on the worst day and reports "much better" has measured the calendar, not the plant.

What the studies found

Cautions and interactions

• Keep a written record before changing anything. A number recorded in advance cannot be revised by hope afterwards.
• A single person's improvement is not evidence about the preparation. It is evidence about that person's week.

Related

• The placebo and nocebo effects, and why a real effect is not the same as a drug effect
• How to read this handbook

Sources

1. Reduction of Chronic Primary Low Back Pain by Spinal Manipulative Therapy is Accompanied by Decreases in Segmental Mechanical Hyperalgesia and Pain Catastrophizing: A Randomized Placebo-controlled Dual-blind Mixed Experimental Trial. The journal of pain 2024; doi:10.1016/j.jpain.2024.02.014 PubMed 38369221
2. Placebo and nocebo effect in pain. Handbook of clinical neurology 2025; doi:10.1001/jamaneurol.2018.2017 PubMed 41161953
3. The Nocebo Effect. Annual review of pharmacology and toxicology 2024; doi:10.1097/j.pain.0000000000003025 PubMed 37585661