A lot of good points here, and many seem to echo my thoughts. If you're performing flavor analysis, then the results would be most applicable to the target audience that your tasting panel was pulled from. If your tasters are pulled from the typical craft beer fans, then your results can only be generalized to that level of specificity. If your panel are trained tasters (BJCP judges or something along those lines), then your results can only be generalized at that level. I do feel that I have trained my palate beyond the typical beer drinker, so I'd prefer to see tasting panels that use trained tasters. But either way, the results are just as valid.
Something else to consider here is that as brewers, we all tend to have a bias when going through the results of these. It's tough to be objective with these results when we all typically have some prior experience, and/or preconceived ideas regarding the experiment. Not to invalidate anyone's experience, but we all tend to try to find holes when results don't appear to match our own experiences.
The exBEERiments at brulosophy are rarely going to absolutely prove or disprove a hypothesis. That's nearly impossible just due to the subjectivity and variability of tasting, and also given the sample sizes. But they are still great experiments and are excellent discussion points. Several have certainly opened my eyes a bit regarding my own practices. If nothing else, the general lack of consensus on most the the experiments confirms to me that there are many paths to good beer.