Not a stats expert, but this just seems like a (intentional?) fundamental misconstruing of meta-analysis. My understanding is they correct for random error, not error due to flawed study design.