We demonstrate that estimates of effectiveness can be substantially improved by controlling for biases along dimensions that are observable and testing the sensitivity of estimates of potential hidden biases. We used matching methods to evaluate the impact on deforestation of Ecuador's tropical Andean forest protected-area system between 1990 and 2008. We found that protection reduced deforestation in approximately 6% of the protected forests. These would have been deforested had they not been protected. Conventional approaches to estimate conservation impact, which fail to control for observable covariates correlated with both protection and deforestation, substantially overestimate avoided deforestation. Our conclusions are robust to potential hidden bias, as well as to changes in modeling assumptions. In addition, it is assumed that this research will help decision-making in the framework of international climate change mitigation policies, such as REDD+. (C) 2015 Elsevier Ltd. All rights reserved.