This study sought to predict U.S. public-school principals' risk of turnover using nationally representative data. After screening several personal and school characteristics that might predict the likelihood of principals leaving their jobs, a logistic regression analysis was employed using 15 significant predictors out of 37 variables identified to predict the outcome variable (low or high risk of turnover). The findings revealed that principals were more likely to leave their position when they did not participate in aspiring principal programs, had a doctorate, were male, were a minority, or worked at schools that lacked controlled access to school buildings