Jennifer Alix-Garcia, Oregon State University

The past decade has witnessed an explosion of research, across many disciplines, relying on remotely sensed data. This explosion reflects an increase in computing technology, along with public access to many satellite data sources. While this clearly represents progress, given the numerous advantages of satellite imagery, researchers unfamiliar with the intricacies of the collection process typically overlook an important feature of the data: non-classical measurement error. Here, we detail the potential sources and nature of these errors, propose a solution for the case of a mis-measured, remotely sensed binary outcome, and validate this solution using a Monte Carlo study. We use our estimator to evaluate a conservation program in Mexico. Our analysis yields practical recommendations for researchers going forward. READ MORE