Not to necessarily defend the study here, but I don't think that's a good reason to dismiss the study out of hand. Words themselves, the actual sounds independent of their technical meaning, still have very strong, deep-rooted associations that are hard to fight even if you've been educated. The subtext and day-to-day use of "intelligence" and "problem solving skill" are different, and color the way you think of them even if you know they're technically synonymous (which, do note, they're not quite). That association operates at the same deep level at which the supposed effect works.
