The Association Between Intradistrict School Mobility and Academic Performance in Elementary School for Ethnically Diverse Children Living in Poverty

Alex Moffett

Major Professor: Adam Winsler, PhD, Department of Psychology

Committee Members: Timothy Curby, Angela Miller

David King Hall, #1021
November 21, 2019, 01:30 PM to 03:30 PM

Abstract:

The United States has one of the highest school mobility rates of any developed country and is disproportionately experienced by ethnic minorities hailing from lower income families, especially those attending schools in densely populated districts. Disentangling school mobility from preexisting and concurrent factors has proven difficult, with considerable variability in effect size and even directionality. The variability in findings among the school mobility research field is largely reflective of variability in research approaches. Prior research has been inconsistent with inclusion of child characteristics to reduce selection bias, modeling change of the association of school mobility and academic outcomes over time, and accounting for variance at item, child, and school levels. The main goal of this dissertation is to address some of these prior gaps with an applied developmental, ecological systems approach by controlling for preexisting and time-varying child characteristics, and then assessing the association of within-district moves (ever moving, frequency of moving, and timing of first move) with academic outcomes throughout elementary school. This was achieved by using a five cohort-sequential longitudinal dataset of students attending schools in a densely populated school district between first and fifth grade (N= 20,806). Main analyses were conducted with cross-classified multilevel growth models in HLM 6.4 software (Raudenbush & Bryk, 2002) to account for variance at the item level and among the child (row) and school (column) classifications. Controls for annual status of free and reduced lunch, primary exceptionality, and English proficiency, as well as for time-invariant controls of school readiness, gender, and ethnicity were included.