Found it in the June issue - it was a Cooper .223, not a .22-250 which was screwing up my search.
I'll give the guy credit, he put more work into it than most authors I see who claim 'no difference' between brand x, y or z, or method a, b and c. And I do agree with him to a degree... unless you are really bent on wringing the absolute mostest from your gun and loads, a lot of the twiddly stuff falls into the 'feel-good' category.
From a strictly technical perspective... the sample sizes are a bit small. Then again, time and money being finite, I can't entirely fault him there either. The problem is, that as close as the results are, you'd need a *lot* bigger sample size to say (with mathematical confidence) that there really is or isn't a difference, and that the overlap wasn't simply due to chance. The smaller the difference, the bigger the sample required generally speaking.
FWIW, I did a test (for a final paper in my intro statistics class) last year on case weight vs. case volume vs. muzzle velocity in a .308 Win, and *did* find a correlation between weight and muzzle velocity. Albeit a *slight* one, but still statistically significant nonetheless.
I do want to get back to that project and spend some time working with more aspects of the cases - weight and volume and runout specifically. Something I have wondered about is whether different parameters would have noticeably more or less effect in particular chamberings - say .223 Rem vs. 6mm BR vs. .308 Win vs. 6.5-284, etc. The tests start getting complicated, at that point
I still say that I'm not going to tell Mr. Lewis that he's full of hot air, even if I don't subscribe to quite all that he does to his brass. We need more people willing to go out on a limb and say "This is how I do it" and weather the flak that follows.
Monte