This study uses Data Envelopment Analysis to evaluate the cost efficiency of 52 Italian public universities. In addition to being one of the first such cost studies of the Italian system, it explicitly takes into account the internal cost structure of institutions' education programs; a task not prevalent in past Data Envelopment Analysis studies on higher education institutions. The findings highlight biases in efficiency measurement arising from not separating the sample into those institutions with medical faculties and those without. Efficiency scores are also shown to vary considerably when education quality measures are input-based or output-based. From a policy perspective, our findings suggest that limiting the enrollment growth of some institutions while expanding enrollments in others could reduce system-wide costs and improve overall efficiency. (Contains 6 tables and 12 notes.)