The failure to do the 'simple' and transparent work to establish the validity of stats is on the front end, and lies with those who produce the stats. It's not the responsiblity of the general public to prove the validity of stats.
I recall the ridiculous graphs which plotted nothing but international trade as a percentage of GDP, and the conclusions that that proved Smoot-Hawley was some major factor in the Great Depression. In that case, practically any element of GDP could have been plotted against the total with the same result, or same trend line.
This practice of taking one element of GDP, and plotting it against the total over several decades, then claiming a definite valid relationship is nonsense. The relationship might be valid, but not because the statistical method was valid. There are far too many other factors involved.
Listen to yourself: the author of a study needs to anticipate all of the objections to it in advance, no matter how cockamamied they are? You have it backwards. The author presents his "evidence," and his challenger then presents his . . . one needs to follow the "yeah, but" argument with "here's why."
(And the reason you "recall" that graph is because it proves your premise wrong--you've never found anything similar to challenge it). Yeah but, yeah but, yeah but . . . but never proven.
All I'm saying is that, if you want to prove that Smoot-Hawley didn't contribute to the Great Depression, you have to show some numbers supporting the same . . . none of this, "yeah, but we all know the Earth is flat" (or "he was wrong about Mars") stuff.