External Validity

External validity refers to the ability to generalize the findings of an experiment to other settings, populations, and times. For educational psychology, this ability is a vital one, because it indicates the extent to which various programs shown to improve learning in one (perhaps artificial) setting can be confidently expected to show similar improvements in other (real) classroom settings. Without high external validity, the program may not have its intended or presumed effect in the various populations of interest. This entry will describe external validity in detail, offer suggestions for increasing it, describe threats to it, and discuss the trade-off between external validity and internal validity.

When a new educational technique or program is being evaluated in an experimental trial, of necessity the evaluation involves, for example, a specific school or district or set of classrooms. In addition, it takes place with specific children, teachers, materials, and at a particular time. However, researchers and educators clearly do not wish to restrict their conclusions to those particular schools, teachers, children, and so on; they prefer to generalize, that is, assume that their conclusions about the program extends to other (or even any) schools, teachers, children, and the like. The issue of external validity concerns whether such a generalization is defensible and true.

Logically, such a generalization is hardly ever completely justifiable. Even if the experiment is later replicated in many classrooms or schools, and the same result is found, it remains logically possible that the next school will find different results. Although such an outcome always remains a logical possibility, its plausibility is related to the external validity features of the experiment.