To date, most of the scholarly and policy debate regarding charter schools has focused on two questions: (a) whether charter schools are using their autonomy to engage in innovative practices and (b) whether students in charter schools, taken as a group, perform better or worse than similar students in noncharter public schools. For the most part, however, scholars have done relatively little to bring these two sets of questions together in order to assess whether some uses of charter school autonomy are more academically productive than others. This study seeks to identify factors that distinguish academically successful charter schools from others. Using a unique data set on Pennsylvania charter schools, the study tests explanations about the correlates of academic success.  Using what might best be termed “pseudo-growth curve analysis,” the study finds that charter schools with higher degrees of perceived accountability produce stronger score growth. Similarly, charter schools with higher degrees of teacher mission commitment and leadership stability produce stronger growth rates in reading and math. Schools with higher degrees of classroom autonomy appear to have lower growth rates, perhaps reflecting recent research on the importance of shared professional culture in teaching and learning. Finally, parent volunteerism appears to be negatively associated with score growth, though it is not clear whether this is simply a proxy for poor governance. While conclusions are limited by the properties of available test score data and a small number of cases, the findings from this study provide a useful early foray into an important policy and school-design question.