The statistical power of family practice research.

Abstract

To determine the statistical power of family practice research, we conducted a survey of all family practice studies published in 3 major family practice journals in 1988. Only 5 of 86 studies calculated statistical power, although 67 (78%) reported statistically nonsignificant results. Calculations of statistical power revealed that more than 80% of family practice studies had sufficient statistical power to detect medium to large effect sizes, but not small effect sizes, which some readers might deem clinically important. We conclude that the authors of family practice research studies do not pay enough attention to the issue of statistical power when reporting statistically nonsignificant results, leaving their readers in doubt as to whether a clinically significant effect could have been detected.