Robert, one of the super interesting things about this particular relationship (height/hand) is that it seems very reliable. Here’s an amazing fact: if I had used the servicewomen’s data instead of the servicemen’s data, I would have gotten the same answer (the expected hand length if he were a woman of his height is within 1 mm of the one computed above, and the women’s residial standard deviation is within 0.05 mm of the men’s. So even with a very different population — women — I would have gotten the same answer. I could break it down by race/ethnicity as well, but that would require more time than I’m willing to spend on it :) The data are available!

Scatterplot of hand length and height with men (light blue circles) and women (light green squares), and respective regression lines added. Note that the regression lines are almost identical.

A critique of the result in the article would thus have to go beyond stating that service-members are different with respect to hand size or height, because any selection on the basis of height or hand size independently would leave the relationship intact. You would have to say that somehow, service-members were selected to have a different relationship between hand size and height. This would be saying that the Army selects people with disproportionately-sized hands (rather than big hands). That’s certainly possible, I guess, but it strikes me as unlikely.