In the public sector, performance management systems are designed to create and then deliver performance information to key decision makers to inform decision making and improve organizational performance. Despite the growing popularity of these systems, we have a very limited understanding of how public managers actually use the information performance management systems produce (Moynihan, 2015). A key premise of this project is that existing work has relied too little on experimental methods to understand decisions regarding performance information.

The long-term objective of this project is to apply a behavioral perspective to our understanding of how public managers use performance information and more generally to the study of bureaucratic decision making. More specifically, this dissertation looks within the process of decision making to understand what factors might lead to variation in the interpretation of performance information. Using survey experiments among public managers, it tests two competing frameworks for how people make decisions. According to the "rational" school, people make decisions in a way that seeks to maximize their expected utility (Von Neumann and Morgenstern, 1944). On the other hand, the behavioral framework suggests conditions under which people make non-maximizing decisions because of the limits we face as people and as social animals (Simon, 1947; Cyert and March, 1963; and Kahneman, 2011b).

This dissertation tests for evidence of the assumptions of these schools in several experiments: negativity bias, historical and social comparison biases, and gender and cognitive bias. Results of the survey experiments demonstrate systematic variation in the interpretation of performance information, suggesting evidence in support of using a behavioral perspective in future research on performance information use by public managers.