The present study was conducted to test the ability of two subscales from the Friendship Quality Questionnaire (FQQ;Parker & Asher, 1993) to discriminate between different target friendships. This was done by testing for mean differences between targets, and by using both exploratory and confirmatory factor analyses to examine the factor structures for these measures when the responses for different targets were considered together. Participants were 293 Italian middle-school students. Both subscales (Validation and Caring, and Conflict and Betrayal) showed mean differences between the best friend and the second and third best friends, but only the validation and caring subscale demonstrated differences between the second and third best friends. Factor analyses showed that for both subscales some items were able to discriminate between friends whereas others were not. Confirmatory factor analyses demonstrated the existence of target-specific factors, as well as a common 'individual differences' factor. Practical and theoretical issues regarding the different sources of item variance are discussed.