LAUSR

I n a previous commentary, 1 we discussed the difference between efficacy trials, which ask whether an intervention can work under ideal circumstances, and effectiveness trials, which seek to determine whether an intervention doeswork in the realworld. The article pointed out that many of the design features of 1 were opposite to the features of the other. For example, in efficacy trials, there are very strict criteria regarding who can be a participant, in order to increase the odds that the trial will show an effect; whereas in effectiveness trials, recruitment tries to replicate the characteristics of patients who will actually be seen by clinicians. In this paper, wewill discuss a related topic – the difference between the internal and external validity of a study. Internal and external validity were first discussed in a highly influential book (at least in psychology) by Donald Campbell and Julian Stanley called Experimental and Quasi-Experimental Designs for Research, which has subsequently gone through a number of revisions and name changes. Very briefly, internal validity is concerned with the believability of the results –was the study carried out and analyzed in such a way that you can trust the findings? That is, it deals with the procedures used in implementing the intervention and analyzing the data. On the other hand, external validity relates to the generalizability of the findings to other people, other settings, and other times. Neither internal nor external validity are yes/no concepts. Both exist along a continuum and are subjective assessments by the reader, based on his or her evaluation of the methods used in a study, although some unvalidated checklists do exist. Unlike the case with efficacy and effectiveness, there is not always a trade-off between the 2: increasing the internal validity of a study does not necessarily lead to a decrease in its external validity.