This paper explores the challenges of evaluating social policies directed by states and highlights several lessons learned in this area through the Assessing the New Federalism project: broad-based analyses are hard, limits on research breadth are essential, data needs cannot be compromised, and clear, timely presentation of results is more important than ever. The paper also focuses on a particularly daunting analytic challenge: determining which state reforms are helping low-income citizens and advancing broader social objectives and which are not. The author examines four basic analytic techniques to determine their potential for measuring the contribution of policy reforms to outcomes.