Given the obvious importance of assessing sex in the context of experimental results, why is it that sex differences are not more frequently studied in human genetics? The likely answer is that a number of issues exist when conducting statistical analyses on genetic data, most of which are related to the size and complexity of the datasets generated, as well as the wide variation in selecting, defining, and measuring phenotypes. Because of this size and complexity, researchers often use simple statistical methods to uncover associations between genotype and phenotype. For example, traditional GWAS analyses typically examine >500,000 SNPs across the genome, leading to demanding multiple testing adjustment standards, setting the current standard for genome-wide significance at p < 5 × 10−8 [33]. Such a conservative p value threshold protects against detecting false positives; however, some researchers have suggested that the threshold is too conservative [34] and likely leads to high rates of false negatives, which can stall progress much like publication bias. The inherent issues of power and stringent p value thresholds has created a culture aimed at increasing power at all costs, and one of the casualties is little attention paid to sex.
