Senn Stephen
Perspect Biol Med. 2013 Summer;56(3):452-3. doi: 10.1353/pbm.2013.0024.
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.
本笔记认为，与该期刊中的观点相反，存在无限多个因果因素的可能性并不会使随机化无效。这些因素的影响必须由结果来界定，而且由于推断是基于治疗组间变异与治疗组内变异的比率，随机化仍然是有效的。
