Can placebo controls reduce the number of nonresponders in clinical trials? A power-analytic perspective.