This study investigated infant listening preferences for two versions of an unfamiliar Chinese children's song: unaccompanied (i.e., voice only) and accompanied (i.e., voice and instrumental accompaniment). Three groups of 5-, 8- and 11-month-old infants were tested using the Headturn Preference Procedure. A general linear model analysis of variance was carried out with gender and age as the between-subjects variables and listening time to the two renditions (unaccompanied, accompanied) as the within-subjects variable. Results indicated a clear preference for the unaccompanied version of the song in all age groups. (Contains 1 figure and 1 table.)