Over the past decade, states and districts have updated teacher evaluation systems in response to federal incentives, such as waivers to No Child Left Behind and grants from Race to the Top. The use of multiple measures, however, raises an important issue for policymakers and practitioners: to what extent should different measures of a teacher's performance align with one another? Chin and Goldhaber explore this issue for two measures commonly used to evaluate performance: classroom observations of teacher practices and value-added. Specifically, Chin and Goldhaber explore levels of correlation between simulated value-added and observation scores when varying two factors: the correlation of each measure of performance to one or more unobserved dimensions of "teacher quality," and the amount of error in value-added or observation scores. By examining the multiple causes for low correlations between the two measures, it is hoped to provide researchers and policymakers a better understanding of the components of new teacher evaluation systems and how they might be expected to relate to one another. Tables and Figures are appended.