Nora Gordon

In this paper we investigate decision-making and the potential for social learning among school administrators in the market for school reform consulting services. Specifically, we estimate whether schools are more likely to choose specific Comprehensive School Reform plans if "peer" public schools have performed unusually well during previous periods of plan implementation. In the absence of concrete data on social networks among administrators, we consider a range of plausible peer group boundaries defined by common governance, geography, or district characteristics. Social learning could be particularly important in this setting, in which there is essentially a vacuum of other data on plan efficacy. We find strong evidence that schools that are located in the same districts tend to contract with the same providers, regardless of past performance. In addition, our point estimates suggests school administrators may use information from their peers to choose the plans they perceive to have performed best in the past. However, despite choosing a market with an unusually comprehensive data source on contracts between public schools and private firms, our statistical power is weak and we cannot reject the absence of such social learning. We conclude by outlining future experimental directions for related research.