This thesis will investigate the significance that the adoption of Randomized Controlled Trial (RCT) evaluations has for social policymaking and policy expertise. Specifically, I will compare the experience of Mexico and Brazil as they evaluate their respective new poverty alleviation programs to examine the ways that politics, measurement and expertise interact in RCT evaluations. My goal is to trace the institutional, political and epistemic determinants behind the adoption of a RCT evaluation in Mexico, in contrast to a strong rejection to the experimental approach in Brazil. To implement this comparative framework, I will combine historical-institutional research and interviews. First, I will use archival research and secondary sources to outline the institutional arrangements and political struggles within the Mexican and Brazilians states that informed the decisions regarding methods of evaluation and policy legitimation of their new poverty alleviation programs. Second, I will make a topography of the Mexican and Brazilian policy knowledge regimes to examine how the institutional machineries through which ideas and assessments of social policies are produced and disseminated affected the forms and types of evaluations that were seen as the most adequate in the two countries. Finally, I will conduct content analysis of official reports and policy documentation, in conjunction with open-ended interviews with key policy-makers and poverty experts to unpack how evaluation techniques and policy devices used in Mexico and Brazil affected the ways that poverty became conceptualized and shaped. By using this multi-layered research strategy I expect to be in condition of unveiling the implications of RCT evaluations to social policymaking in developing countries.