The Placebo in Psychosocial Outcome Evaluations

Abstract

While the study of placebo effects can be of aid in understanding and facilitating therapeutic change, the routine use of placebo control groups in outcome evaluations is of doubtful merit. Placebos are generally unstable, poorly specified, and of questionable ethical viability. Additionally, careful scrutiny of our analytical models suggests that placebo control groups fail to isolate specific treatment effects in the manner commonly assumed. Recommendations include (a) use of aggregate data from metaanalyses to provide normative estimates of expected minimal treatment effects, (b) empirical testing ofputative causal models, and (c) increased utilization of more precise experimental designs. It is suggested that aggregate estimates from existing literature of typical magnitudes of placebo effects provide potentially useful "benchmarks" against which to judge the performance of psychotherapy.