Doctor of Philosophy (PhD)


Social Work

Document Type



May 14, 2013 marked the 10 year anniversary of the creation of the Louisiana Recovery School District (RSD), the turnaround intervention for low performing public schools. Since 2003, the RSD has grown to include over 80 schools across the state. The purpose of this multilevel longitudinal study was to examine the relationship of the Louisiana Recovery School District to student and school outcomes including risk of dropout and standardized test scores. The dissertation measured these influences over time (2007-2010). The data collected were derived from Louisiana’s Educational Assessment Program (LEAP) and quantitative data on dropout status from the Louisiana Department of Education Student Information System (SIS) database. To explore the RSD’s relationship to risk of dropout and student achievement scores over time, two analyses were conducted. First, to examine the relationship between student and school level predictors and the risk of dropout, a multilevel discrete-time survival analysis was conducted. The level 1 analysis included time and student level covariates, while level two included school level covariates. The results of the this analysis indicate that when controlling for student level covariates, RSD students are at a 3.25 times greater risk of dropout than traditional public school students. Next, in the study of the RSD’s relationship to student achievement, a three-level hierarchical linear model was run to account for repeated measures and nested data. The level one analysis examined the influence of time on LEAP/iLEAP scores. The level two analysis examined the influence of student characteristics including race, gender, and free and reduced lunch status on student scores. The third level examined the influence of aggregated school level predictors including school type (RSD versus traditional public school) on student scores. Results indicate that student characteristics of disadvantage including minority status and low socioeconomic status are linked to lower test performance over time. For Sample A, the school level covariate, school type, was significant in each model, suggesting that even when controlling for student and other school level factors, school type accounted for variation in student scores over time. RSD students performed significantly lower than TPS students in both ELA and math over the 4 year period of the study. The school level covariate, school type, was also significant for Sample B indicating that students in RSD schools perform lower on ELA and Math portions of the LEAP/iLEAP/GEE than their TPS counterparts. When controlling for other school level covariates in Sample B, the effect for school type was no longer significant. No other school level covariates were significant in the models suggesting that variation in outcomes was accounted for by student level characteristics and time rather than by school level characteristics



Document Availability at the Time of Submission

Release the entire work immediately for access worldwide.

Committee Chair

Livermore, Michelle

Included in

Social Work Commons