Type II (beta) errors in the hand literature: the importance of power.

Abstract

When a study concludes that there is no difference between 2 treatments ("negative studies"), it is essential to determine whether the study has sufficient power to find a clinically significant difference. Insufficient power precludes an adequate assessment of therapeutic efficacy and may result in a type II error, an erroneous conclusion that the null hypothesis is correct. In evaluating 39 studies that highlighted negative findings in The Journal of Hand Surgery, we found that 32 (82%) papers had a power of less than .80 to detect a 25% treatment effect and, when the treatment effect was increased to 50%, more than one half of the studies still had a power of 0.80. These "negative studies" frequently have inadequate statistical power to support their conclusions. These findings have important implications for researchers, editors, and readers.