ABSTRACT
This chapter explores the phenomenon of regression toward the mean to illustrate a kind of bias, which can be accounted for when there is a concurrent control group in the design. External validity is the validity of inferences as they pertain to the generalizability to future subjects rather than the specific trial participants. To illustrate this concept further, let us consider the lipid-lowering treatment for prevention of coronary heart disease (CHD) as an example. Internal validity is the reasonable representation of the treatment effects within the study population. Internal validity is a prerequisite for external validity. The fundamental idea for supporting internal validity is to design trials that include comparable groups such that the outcomes, whether different or the same, measure only the effect of the treatment, and nothing else. The basic philosophy for repeatability of clinical trials is that scientific experiments should be verified by replication.
