Our point about sample size is not new. We explained it in our American Journal of Psychiatry paper [16], Uher et al. explained it in two publications [15,17], and Karg et al. also explained it in their meta-analysis [18]. Yet, the meta-analysis protocol does not contain a justification of its choice of N = 300 as a cut off for study inclusion. Why not 500, why not 200? According to PRISMA guidelines for reporting meta-analyses, those that aspire to be authoritative provide a rationale for their decision points, e.g., “Specify study characteristics used as criteria for eligibility, giving rationale” (http://www.prisma-statement.org/2.1.2%20-%20PRISMA%202009%20Checklist.pdf).
