Correction: Are Treatments More Effective than Placebos? A Systematic Review and Meta-Analysis

Among studies with continuous outcomes in our systematic review comparing placebo effect sizes with treatment effect sizes we calculated mean differences between to estimate placebo effects, treatment effects, and the difference between placebo and treatment effects. In fact the standardized mean difference should have been used. There was no error in the calculations among studies with binary outcomes. We have corrected this error and revised Figs 3, 4 and 5 from our review here. Correcting the error strengthened our conclusion that there is rarely a statistically significant difference (at P = 0.05 or lower) between the magnitude of placebo effects and the magnitude of treatment effects. In the original review placebo effect sizes and treatment effect sizes did not differ by a statistically significant amount in the three subgroup analyses: all studies with objective outcomes (treatment effects larger), studies of anxiety treatments (treatment effects larger), and studies where all criteria for ruling out bias were met (placebo effects larger).