School choice has gradually become one of the central strategies of urban school reform over the past several decades (Harris, 2014). However, parents face constraints, have limited information, and sometimes display weak preferences for the types of academic outcomes that policymakers might hope for (Bell, 2009; Schneider, Teske, Marschall, & Roche, 1998; Teske, Fitzpatrick, & Kaplan, 2007). We study the topic in New Orleans, a unique context where there are neighborhood schools and where all parents have to make active school choices. Roughly 75 percent of public schools participate in the city’s OneApp system, in which they can rank up to eight schools. We ask two questions:

                (1) What school characteristics are most important to parents when choosing schools?  We answer this question by merging an extensive database of school programs and characteristics to the OneApp parent rankings and using an exploded logit that takes into account every viable option (e.g., all school that include the grade level of the student). We compare results using this approach to another related study of the same population using both self-reports and observed schooling choice (Harris, Larsen, & Zimmerman, 2014), providing important evidence about the validity of various research methods of identifying parental preferences.

                (2) To what degree do parents’ ranking of schools in subsequent years depend on the value-added of teachers? This extends the first analysis by merging in data about students’ teachers, calculating value-added, and including this among the student/school covariates. We hypothesize that students’ experiences in a given school, including how much teachers taught them, will positively influence their subsequent rankings of their initial school. To our knowledge, this is the first study to examine the importance of individual teacher effectiveness in parental school choice relative to more commonly measured school characteristics.