Thus, in order to get the most accurate picture of sex differences, researchers need to measure personality with a higher resolution than that afforded by the Big Five (or other traits at the same hierarchical level). A corollary is that, when investigating sex and personality as predictors of a given outcome (such as health, self-esteem, and so forth), cleaner and more meaningful results are likely to obtain if personality is measured at the level that yields the most clearly sex-differentiated profiles.
http://plosone.org/article/info:doi/10.1371/journal.pone.0029265#article1.body1.sec1.sec4.sec1.p3

If you set up your models based on the level of maximum differences, should you really be surprised when your findings demonstrate large findings? It seems that whatever assumption you make here determines your findings.

I like the paper though, the authors clearly put a lot of thought and effort into it, this just seems like a fatal flaw to me.