School Competition and Geography in New Orleans

• In Event: Roundtable Session 38
  In Roundtable Session: Understanding School Choice as a Geographically Relational Phenomenon Across North America: Using Geographic Information Systems Approaches

Abstract

Objectives: A key goal of introducing school choice is to generate competition between schools (e.g., Wohlstetter et al., 2013). However, studies have operationalized “competition” in different ways, yielding different results and overall mixed findings (e.g., Bettinger, 2005; Bifulco & Ladd, 2006; Hoxby, 2002; Ni, 2009). Many studies use geographic measures to determine the boundaries of the market, or who the competitors are (e.g., the schools in the surrounding 5-mile radius), but research suggests that student mobility and school leaders’ perceptions may also be important determinants of competition (Jabbar, 2014). Furthermore, within districts, there might be subgroups of schools that compete with one another more intensely (Sirer et al., 2015).

Methodology & Data: To explore how these dimensions interact, and whether school leaders in fact perceive competition from student movements or compete with schools nearby, we examine the geography of competition in New Orleans, the city in the U.S. with the largest charter-school market share. We ask: a) How do perceived competitors correspond to school location? b) How do student flows correspond with school location? c) Are there identifiable clusters of perceived competitors, and how do they map onto physical space? We also explore the other characteristics shared by these clusters of competitors. In this paper, we overlay flows of students and school leaders’ perceived networks of competitors to see whether school leaders’ perceived rivals cluster by geography, and how they align with actual student movements between schools. We determine the geographic boundaries of school competitors, identifying schools within a set geographic radius, mapping clusters of schools, and showing the extent to which space matters for competition using various measures of geographic density found in prior literature. To construct networks of perceived competitors, we conducted surveys of 90% of public schools in New Orleans, in which each leader was asked to rate as a competitor every other school in the city. We pair this with data on school location and actual student movements between schools. We also include data from the U.S. Census and the American Community Survey to examine race, poverty, and other factors.

Findings: Our preliminary findings suggest that school leaders do respond to student exits from their school, and appear to be aware of their school’s competitors. Furthermore, rather than competition occurring across the entire district, we find it occurs within subgroups of competitors not confined to geographic areas; in other words, schools perceived rivals do not align with typical measures of competition based in geography.

Significance: Our findings have implications for school choice policy. To the extent that leader’s perceptions of competition do not correspond with commonly shared geographic definitions of educational markets, school choice programs may not achieve their intended results. Our analyses may also aid planners and policymakers to site schools in areas of most need of additional options. Finally, our study contributes to the growing body of educational research that uses geospatial analyses (e.g., Lubienski et al., 2009; Richards, in press) to show how network analyses can complement geospatial analysis.

Authors

• Huriya Jabbar, The University of Texas - Austin

• Spiro Maroulis, Arizona State University

• Joanna Dolores Sanchez, The University of Texas - Austin

• Emily Germain, The University of Texas - Austin