If you can replicate the handedness-immune disorder effect, the more power to you

Abstract

Bryden, McManus, and Bulman-Fleming's (1994) meta-analysis of studies examining the Geschwind-Behan-Galaburda hypothesis indicates that the original data set supports the hypothesis significantly more than the body of replication attempts. We present data on 2256 subjects that clearly fail to support the hypothesis and describe practical and statistical reasons suggesting why the original data cannot be easily replicated. A power analysis suggests that to find the effect in a general sample, one should plan to have 17,000 to 40,000 subjects.