The Placebo Control Group: An Analysis of Its Effectiveness in Educational Research

This article reports a review and meta-analysis of placebo control groups in educational experiments. Studies were clustered into control procedures that included explicit manipulation of attentional elements (Attn + ), explicit descrip tion of the placebo activity plus implied reference to attention (Act + ), and exclusive reference to the activity (Act-Only). The Act + procedure was the practice most con sistent with the model for a placebo control. Compared with the other two proce dures, Act + controls were employed more often in interpersonal rather than aca demic studies in which subjects knew they were participating in an experiment, had an expectancy for performance change, and the placebo activity resembled the treat ment task. Significantly larger effect sizes in Act + studies were found with the ab sence of experimenter controls and child subjects. The results were compared with findings from Hawthorne-labeled control groups in educational studies, and the dif ferences between placebo-labeled control groups in education and those used in psy chotherapy research were considered. ORIGINATING IN MEDICAL RESEARCH, the application of the concepts of placebo effect and placebo controls to non-drug research was first suggested by Rosenthal and Frank (1956; see Goldstein, 1962). The concepts were borrowed because researchers in psychology and education recognized that reactive effects arising in the conduct of drug research were also encountered in research in the behavioral sciences. In the drug context, a placebo is defined as an inert substance with no known pharmacological effect (Shapiro, 1960). Its meaning, when applied to tests of the outcome ef fectiveness of therapy or to classroom tests of new curricula or teaching methods, is less clear. Indeed, the fundamental questions of whether and how the placebo concept applies to behavioral research has never been ade quately addressed.