As political scientists, we spend much time teaching and doing scholarly research, and more time than we may wish to remember on university committees. However, just as many of us believe that teaching and research are not fundamentally different activities, we also need not use fundamentally different standards of inference when studying government, policy, and politics than when participating in the governance of departments and universities. In this article, we describe our attempts to bring somewhat more systematic methods to the process and policies of graduate admissions.We had a role in the graduate admissions process at the Department of Government at Harvard University at different times over the past half-decade. We conducted a study of the admissions committee's policies and attempted to bring some of the modern methods of statistical inference, common in political science research, to the task of choosing among applicants to our graduate program. We report here our experience, our statistical studies, and our improvements to the process, as well as a variety of information that may be of use to scholars and administrators at other universities in similar circumstances.As political scientists, we spend much time teaching and doing scholarly research, and more time than we may wish to remember on university committees. However, just as many of us believe that teaching and research are not fundamentally different activities, we also need not use fundamentally different standards of inference when studying government, policy, and politics than when participating in the governance of departments and universities. In this article, we describe our attempts to bring somewhat more systematic methods to the process and policies of graduate admissions.We had a role in the graduate admissions process at the Department of Government at Harvard University at different times over the past half-decade. We conducted a study of the admissions committee's policies and attempted to bring some of the modern methods of statistical inference, common in political science research, to the task of choosing among applicants to our graduate program. We report here our experience, our statistical studies, and our improvements to the process, as well as a variety of information that may be of use to scholars and administrators at other universities in similar circumstances.Admissions committee decisions represent an interesting combination of judgments based on quantitative and qualitative information. Until our changes, virtually all decisions were made using only qualitative (or "clinical") methods, even though some of the data on applicants were quantitative, such as grades and standardized test scores. We speculated that this pointed to an inefficiency in our admissions process since "a search of the literature fails to reveal any studies in which clinical judgment has been shown to be superior to statistical prediction when both are based on the same codable input variables" (Dawes 1982, 394). Because some of the information...