The resilience of a given impact study is strongly determined by the degree of its internal validity. In general, experimental designs are considered to possess the highest degree of internal validity. Unfortunately, true experiments are not always feasible in evaluation practice and even simple pretest-posttest designs cannot always be implemented. In these cases, evaluators frequently use retrospective pretest methodology (MRP) for assessing impacts. This method is connected with some practical advantages but also with several methodological risks, which are oftentimes not reflected sufficiently. Thus, the objective of this article is to describe how the MRP works, to summarise its strengths and weaknesses, and to provide an overview of the current state of research. Moreover, the article elaborates on implications for evaluation practice with regard to the conditions under which applying the MRP might be an inappropriate approach for assessing treatment effects.

Schlagworte
Retrospective Pretest, Impact Evaluation, Research Design, Response-Shift Bias