Skip to content

Topic

Biostatistics

Choosing methods that match the design, and reading the output as something other than a verdict.

Guides in this topic

  1. 01Survival analysis: Kaplan-Meier, Cox regression and competing risksTime-to-event methods exist because censoring is information. The commonest errors come from forgetting that, or from ignoring the events that make the event of interest impossible.
  2. 02Missing data: complete-case analysis, imputation and what each assumesEvery method for handling missing data makes an assumption about why it is missing. The assumption is usually untestable, which is exactly why it has to be stated.
  3. 03Sample size and statistical power: what the calculation actually needsA sample size is not a number you find; it is the consequence of decisions you make. Most calculations fail because one of those decisions was never made explicitly.
  4. 04Confounding, mediation and collider bias: what to adjust forWhich variables belong in a model is a question about causal structure, not about which ones reach significance. Adjusting for the wrong one introduces bias that was not there before.
  5. 05Odds ratios, risk ratios and hazard ratios: which to reportThe three are routinely described as interchangeable and are not. Reporting the wrong one, or interpreting it as though it were another, changes what the paper claims.
  6. 06How to choose a statistical testThe test is the last decision, not the first. Almost every wrong test in a medical paper is a symptom of a question that was never stated precisely.

Tell us about your study

Send the scope and we will come back with a written quote. No account needed.

Get a quoteResources