I use student-level data from Arkansas from 2004-05 to 2010-11 – some 1.4 million student-year observations – to measure principal effectiveness. Following the lead of other researchers, I first use school and principal fixed effects. I then use student and principal fixed effects, while using indicator variables both for student moves and for the type of school (elementary, middle/jr., high school) in an attempt to control for changes in student-principal matches that due to the student having switched schools. Across both specifications, I find that an increase in principal experience has inconsistent effects, and that principals who are especially effective are not seen as often in schools targeted for Title I aid. I also find that students may tend to do better when matched with a principal of the same race, although effects are inconsistent.