Jesus Christ but the level of statistical incompetence in vast areas of the scientific community fucking astounds me. In what possible world does it make more sense to divide a population distributed along a continuous variable (i.e. age, weight) into arbitrary categories - along what mark? 5 year? 10 year? with what biological significance, exactly? - rather than use regression including quadratics? Holy fucking hell. And when a paper is about a new phenomenon involving a particular statistic which is not a previous variable and not intended to be about that variable...is it really wise to ask about it? My rule when reviewing an article is that if I'm too dumb to post anything meaningful, I don't. Is that really so hard? I'm tempted to start posting my reviews here, except that I'm sure it would be a violation of some other arbitrary fucking rule instead. Peer review has laughably failed. The idiot ratio of those reviewing me is now, incredibly, greater than 50%. I shit you not. The explanation is simple: there are few guidelines to the selection of effective and competent reviewers. Experts in the biology or physiology of a phenomenon are frequently not experts in the statistical analysis of that phenomenon. Just for a laugh, in one of the last inspections I was doing of a lab for a certain department here I threw out the comment "well, as you're using different strains, you'll probably be fitting a random or mixed model for the analysis". Then I sat back and cruelly smiled, inwardly, as they blanked completely. The PI's response? "I don't even know what that means." I was tempted - so very tempted - to say "yeah, I didn't think you would", but I didn't. Why bother? None of these fuckheads know anything about the construction of what they're doing. "Oh, well I know this protein conforms into the teensy little hole on this one because of some fucking blotto test I performed with an antibody" - really? How many runs, dickwad? Did you block it? RCBD? Nobody does that anyway - except plant people, who are a longitudinal scourge of unacknowledged plot heterogeneity - but for fuck's sake, if I mention it could you please know enough not to blank the fuck out? If I use statistical terminology, maybe it would be best not to smirk and - again, I shit you not - tell me that I made it up. FUCK. FUCK FUCK FUCK FUCK FUCK. IDIOTS!
