Quantifying School Neighborhoods and their Relations with Child Outcomes: A Virtual Systematic Social Observation Approach

• In Event: 3-092 - Novel Approaches for Measuring Neighborhoods and their Influence on Child Outcomes

Integrative Statement

Past research has shown robust links between sociodemographic features of children’s residential neighborhoods and their learning and health outcomes (Leventhal & Brooks-Gunn, 2000). Less empirical work, however, has focused on the features of neighborhoods that may explain these links, or whether they may also apply to school communities. We use a new virtual systematic social observation (SSO) protocol in Google Street View to understand how features of young children’s school grounds and surrounding neighborhoods predict their developmental outcomes. We hypothesize that children from school environments characterized by order, safety, and resource availability will demonstrate larger gains in language and literacy skills and self-regulation over one school year than their peers from less optimal school environments.
Data for this study come from the National Center for Research on Early Childhood Education Professional Development Study (NCRECE PDS). We use child outcome data from 767 predominantly low-income, racially and ethnically diverse children (M age = 4.20y; 51% female; 48% Black, 33% Latino) in 151 early childhood centers (“schools”) across nine U.S. cities (e.g., Chicago, Dayton). Building on work in UK residential neighborhoods by Odgers and colleagues (2012), we develop a virtual SSO protocol to capture dimensions of US school settings that prior research suggests may support young children’s development. To score the SSO, coders take a virtual “walk” around the school neighborhood using Google Street View and code for the presence/absence of resources (e.g., playground, open field), while also answering more subjective questions like, “I would send my child to this school” (see Table 1 for items). After double coding all schools, we use classical test theory and confirmatory factor analysis (CFA) in split-half samples to develop subscales characterizing the order, safety, and resources of school grounds (see Table 1 for results).
We find statistically significant, negative correlations between school grounds’ “Perceived Order and Safety” and Census-reported population density, poverty, and proportion of non-White children (r range = -.23 to -.44). Similar associations were found for “Physical Signs of Order and Safety” (r range = -.22 to -.48), although correlations for “Resources for Outdoor Play” were largely non-significant. Table 2 shows results of a multi-level regression analysis predicting children’s end-of-year language and literacy skills and self-regulation skills based on SSO school grounds subscales, controlling for child and neighborhood covariates (including baseline outcome scores). Results suggest that outdoor play resources (e.g., playground, open field) are marginally predictive of gains in child self-regulation. Child outcomes were not associated with either perceptions or physical signs of the order and safety of school grounds.
These preliminary findings suggest that virtual SSO approaches may capture unique, developmentally salient, and potentially malleable features of young children’s schools environments. Analyses for the final presentation will improve the internal and external validity of these preliminary results (e.g., using further covariates; considering variation based on urbanicity, city). We will also expand our SSO to characterize the neighborhoods surrounding schools, their correlations with other neighborhood characteristics (e.g., crime rates, Census data), and their links with child outcomes. Policy implications will also be discussed.

Authors

• Dana McCoy, Harvard Graduate School of Education
  Presenting Author

• Terri Sabol, Northwestern University
  Non-Presenting Author

• Emily Hanno, Harvard University
  Non-Presenting Author