The Effects of School Size on Parental Involvement and Social Capital: Evidence from the ELS: 2002
Recent state and federal policies designed to improve American public schools have generally focused on introducing standards (for example, No Child Left Behind) or choice (for example, charter schools and vouchers). However, another increasingly prominent approach to reform has emphasized the possible benefits of creating smaller schools as well as small, focused learning communities within schools, particularly at the high school level.1 The growing national interest in the small-schools movement has been catalyzed largely by private foundations (most notably, the Bill and Melinda Gates Foundation) rather than by explicit state and federal action.2 Regardless of its origin, this reform agenda has brought renewed attention to a long-standing research literature that examines the effects of school size on the organizational character and performance of schools.
This literature focuses on how school size influences both costs and outcomes (for example, test scores and educational attainment). However, it also emphasizes how school size may change the nature of educationally relevant social interactions among students, teachers, and administrators. In particular, the apparent consensus in this literature is that the increased formalization of interactions in larger schools harms school quality by fostering alienation and a loss of organizational focus among students and staff.3 However, there appears to be little corresponding evidence on how school size influences patterns [End Page 77] of parental involvement in schools. This is somewhat surprising in light of the fact that constructive parental engagement with schools is widely seen as an important determinant of school quality.4
Furthermore, the effects of school size on parents may also matter for an important reason that is wholly unrelated to the direct objectives of schools. Public schools are often viewed as vital community institutions that can deepen social networks and promote a variety of welfare-enhancing social norms (for example, trust and reciprocity). The role of public schools in promoting this broad group of outcomes, which researchers currently group under the heading "social capital," has important implications both for the optimal design of schools as well as for the proper division between the public and private sectors.5 The size of a public school, for example, could quite conceivably influence the amount of social capital within a community through its effects on parental interactions.
In this study, we present new empirical evidence on whether the size of public high schools influences measures of parental involvement and social capital. This analysis is based on nationally representative data from the base year of the Education Longitudinal Study of 2002 (ELS:2002). In addition to examining novel measures of outcome based on recent data, our study also engages a substantive methodological concern. Any inferences about the causal effects of school size are likely to be complicated by the fact that the unobservable traits that influence a parent's pattern of civic engagement (for example, the enjoyment a parent derives from interacting with others) may also influence the size of the public school the family chooses. The conventional approach to addressing this concern is to exploit a plausible natural experiment that influences school size.6 However, in the absence of a compelling experiment, we adopt an approach developed in a recent study by Altonji, Elder, and Taber on the effects of Catholic schools.7 Following their lead, we attempt to establish bounds on the causal effects of school size by using the differences in observed traits across parents connected to smaller and larger schools as a guide to the size and direction of their potentially confounding unobserved traits.
The paper is organized as follows. The next section provides brief discussions of the school-size literature and the possible relationships between school size and the engagement of parents. This is followed by a discussion of the ELS:2002 data...
