ABSTRACT
The designs discussed in the previous chapters share the feature that nesting of individuals within clusters was established prior to randomization to treatment conditions. Examples are trials with pupils within schools, family members within families and patients within clinics. Randomization is done at the cluster level in case of a cluster randomized trial (Chapters 4 and 5) and at the individual level in case of a multisite trial (Chapter 6). With pseudo cluster randomized trials, a two-stage randomization procedure is used as explained in Chapter 7. In either case, it is assumed that randomization does not influence group membership. In other words, subjects are assumed not to change groups due to their assignment to a particular treatment.
