I have heard that :

if you avoid exposure to equal temperament (western compromise tuning that allows 12 keys with equal deviation from the Pythagorean (integer) ratios) , and only listen to "just intonation" - various tunings that adhere to the integer ratios, you get used to the just intonations and find equal temperament "out of tune". I don't get the impression this study tried to use subjects who weren't still hearing "regular" music.