This meta-study of dictator games, based on 129 papers published between 1992 and 2009, summarizes evidence from over 616 treatments. It uses multiple regression to assess the effect of single manipulations, controlling for alternative factors. The study shows how Tobit and hurdle models provide additional insights into dictator behavior. The average dictator gives 28.35% of the pie, with a left-skewed distribution. Dictators are more likely to give little, with 6 out of 616 treatments averaging zero. The study finds that factors such as stakes, social control, and recipient characteristics significantly influence generosity. For example, higher stakes reduce giving, while social cues increase it. The study also finds that students give less than non-students, and that dictators from developing countries or primal societies are more generous. The analysis reveals that the decision to give and the amount given are separate processes, with social proximity and recipient deservingness being key factors. The study concludes that multiple regression provides a more comprehensive understanding of dictator behavior than single regression.This meta-study of dictator games, based on 129 papers published between 1992 and 2009, summarizes evidence from over 616 treatments. It uses multiple regression to assess the effect of single manipulations, controlling for alternative factors. The study shows how Tobit and hurdle models provide additional insights into dictator behavior. The average dictator gives 28.35% of the pie, with a left-skewed distribution. Dictators are more likely to give little, with 6 out of 616 treatments averaging zero. The study finds that factors such as stakes, social control, and recipient characteristics significantly influence generosity. For example, higher stakes reduce giving, while social cues increase it. The study also finds that students give less than non-students, and that dictators from developing countries or primal societies are more generous. The analysis reveals that the decision to give and the amount given are separate processes, with social proximity and recipient deservingness being key factors. The study concludes that multiple regression provides a more comprehensive understanding of dictator behavior than single regression.