In two separate studies, we examined fifth graders' preference for authentic and arranged versions of world music recordings, the relationship of those preference ratings to familiarity, and teachers' ability to predict student preferences. In the first study, intact classes of fifth-grade students were randomly assigned to an authentic or arranged listening condition and indicated their familiarity and preference for the same 19 world music songs on a 6-point Likert scale. Teachers of the students in the study attempted to predict their students' preference and familiarity ratings for those same 19 recordings. In the second study, students heard both versions of each song side by side and were asked to first choose the version they preferred, then indicate strength of preference and familiarity. Results of both studies indicated that familiarity with a world music song is positively related to student preference for that song. Subjects in the first study did not differ significantly in the magnitude of their preference ratings between authentic or arranged conditions, whereas subjects in the second study overwhelmingly preferred arranged versions in head-to-head comparisons. Teachers' predictions of preference and familiarity ratings were significantly related to student ratings, though tended to be significantly higher. July 1, 2004 October 20, 2004.