Evidence guide

Understanding Scientific Uncertainty

Uncertainty is information about the range of plausible answers, not an admission that nothing is known.

Different uncertainties

Sampling error, measurement error, model assumptions, bias and unknown future conditions are distinct and should not be collapsed into one number.

Confidence is conditional

An interval depends on the model and data. It does not cover unmeasured bias simply because it looks narrow.

Decision under uncertainty

Waiting for certainty can itself cause harm. Good decisions weigh plausible benefits, harms, reversibility and the cost of being wrong.

A practical reading checklist

  1. Which uncertainty is quantified?
  2. Which uncertainty is only discussed?
  3. How sensitive is the conclusion to assumptions?
  4. Would the decision change across the plausible range?