As far as population studies go, this one seems pretty good though. There's a clear pattern attributable to the lifecycle of sperm, very large sample size, and pretty good controls.

I would definitely take this study seriously, and I don't usually pay a huge amount of attention to population studies.