[dllthomas]

Even assuming we knew with perfect certainty (which may be a reasonable approximation but isn't actually the case) that the effect was due to experimental error, given that we yet cannot point to the error we may very well have useful things to learn about experimentation. Any tiny chance of upending well-established laws of physics is just a bonus.

[Steuard]

Can't we learn a lot of wonderfully useful things about experimental technique while studying phenomena that are more likely to be of interest in their own right?

Because in this case, the subject matter of the experiments is explicitly in contradiction with some very well confirmed physical principles, and well within the range of conditions where those prior tests have been done. (This isn't like quantum mechanics supplanting Newtonian physics at molecular scales, this is like someone designing a funny-shaped inclined plane and claiming that Newtonian physics might not apply because nobody has measured accelerations on that shape of ramp before.)

[dllthomas]

"this is like someone designing a funny-shaped inclined plane and claiming that Newtonian physics might not apply because nobody has measured accelerations on that shape of ramp before"

My understanding has it (and admittedly I've not been paying that close attention, precisely because it's not likely there's anything super exciting here) that the difference between that case and this is that it's not just a claim that understood laws "might not apply", but that we actually have experimental results that look funny. That is precisely the case when we're most likely to learn something. (Probably not something tremendously surprising - that's what "tremendously surprising" means :-P)

[NoMoreNicksLeft]

Even this seems unlikely... while it would be great if we could learn to conduct better experiments or learn to recognize the flaws of bad experiments earlier, what's missing are instruments precise enough to determine the exact cause of the results here.

Am I wrong?

[dllthomas]

"[T]hings to learn about experimentation" probably sounded more fundamental than I'd meant it. Improvement of precision of instruments should probably fall in the category, as far as my reasoning above goes. It could certainly still be useful.