Wind testing in sustained winds would be possible I suppose, but a typical tent is going to vary in its ability to handle wind, depending greatly on wind direction.
As far as a tent blowing away, that's almost certainly going to occur because a stake let loose, which does not reflect the tent's performance.
Wind gust behavior is going to depend on how abruptly the windspeed "ramps-up" which again, would yield an infinite number of test scenarios.
Snow load testing, likewise, could have a million different outcomes. One tent might shed powdery snow very well, but not wet snow, and vice versa. Wind direction again would play a role in how the snow is falling, drifting, accumulating, etc.
IMO rain would be the easiest to simulate, but the results would still be highly subjective.
Sleeping bags assume you're an average human, in still conditions. Very few variables here. A tent has a million different confounding variables that would really yield any such experiment inconclusive, and subject to a lot of subjective interpretation that can be spun for or against a particular product.
Just my thoughts on it, I suppose if you were having one trusted, experienced expert pitch a ton of shelters in a particular location, and then deliver impressions of their performance, it could be somewhat useful, but I still think the results would be subjective and inconclusive.