A method for analyzing retrospective pretest/posttest designs: I. Theory.

To control for response-shift bias (RSB) in pretest/posttest designs, G. S. Howard et al (1979, 1980, 1981) proposed the use of retrospective pretests ("thentests"). If an RSB is present, thentest/posttest difference scores will yield better estimates of a treatment effect than will the usual pretest/posttest difference scores. In this article, a method is presented that is aimed at detecting RSB in retrospective pretest/posttest designs. This hierarchical model fitting procedure is easy to apply, prevents capitalization on chance, and makes optimal use of information supplied by the control group.