A brief note regarding randomization.
- Competence Center for Methodology and Statistics
This note argues, contrary to claims in this journal, that the possible existence of indefinitely many causal factors does not invalidate randomization. The effect of such factors has to be bounded by outcome, and since inference is based on a ratio of between-treatment-group to within-treatment-group variation, randomization remains valid.
2013 Jul. Perspect Biol Med.56(3):452-3.