Having been educated and trained as a scientist, I realize the importance of scientific information, both theoretical and empirical. And while I believe empirical data trumps theory everytime, a single study, or several badly contructed studies, are of little value. And I have seen many, many poorly designed studies in sports science. As someone has already mentioned, a well researched meta-analysis is a great place to start, but those may be difficult to find on a given topic.
As far as anecdotal evidence, it can be good or bad. To point out an individual, a team, or even a series of teams from the same school, may not rise to the level of significant information. But at some point, even anecdotal (non-statistical) information can be used as a BS filter.
When I hear opinions about hypertrophy, I ask myself, do bodybuilders do this? Tens of thousands of bodybuilders have tried everything imaginable (and some things you can’t imagine), so if they don’t use a particular exercise or rep scheme or split routine, I feel pretty confident that it is not optimal. The same with discussions of limit strenth. Do strength athletes do this? If it worked, many of them would be training that way.
I will also say that when it comes to ethnicity and the 100m, no amount of theory or anecdotal evidence can overcome the preponderance of statistical information. That data base is huge and has been building since the late 60’s.