The purpose of this article is to invite focused discussion and critical debate about the instruments currently used to select students for art colleges in Europe and North America. At this time of significant expansion and diversification in practices of art making, we must ask if current selection instruments still work. What evidence is there to support their continued use? Are they good indicators of success in art college? Who do they advantage, and whose interests do they serve? In what ways do they contribute to, or legitimate class reproduction and class advantage in the cultural sphere? In taking up these questions, this article addresses four topics of particular relevance to the selection and admission debate: reliability, validity, predictability and equality. It reports findings from two national longitudinal research studies that examined the predictive validity of selection instruments in relation to performance in art college in Ireland. While these findings are specific to the Irish higher education context, they have relevance beyond this context given that the selection instruments used by Irish art colleges are the same as those used by the majority of art colleges across Europe and North America. (Contains 3 tables and 7 notes.)