Evidence

What is established, and what is not

A program that asks a public agency to spend public money should be able to say where its evidence is strong, where it is mixed and where it is genuinely absent. This section does that.

The convention on these pages: each one states what the research question was, who was actually studied, what was found, and what the study design could not capture. Where a finding cuts against a component of this program, it is reported anyway.

Pages in this section

Why we publish the weak parts

Two reasons. The first is that a program built on overstated evidence fails in a way that damages the agency that bought it. The second is that the patients this work concerns have already been on the receiving end of confident claims that turned out to be wrong — in both directions. Restraint is not a marketing posture here; it is the substance of the argument.

The full reference list for the site is at /references/.