Stabilizing School Performance Indicators in New Jersey to Reduce the Effect of Random Error. Appendixes. REL 2025-009

Abstract

These are the appendixes for the report, "Stabilizing School Performance Indicators in New Jersey to Reduce the Effect of Random Error." This study applied a stabilization model called Bayesian hierarchical modeling to group-­level data (with groups assigned according to demographic designations) within schools in New Jersey with the aim of improving reliability, particularly for small groups of students. The four appendixes in the document include: (1) Theoretical review; (2) Methods; (3) Supporting analysis; and (4) Other analyses: Implications of the use of group-level data arising from relationships among demographic composition, subgroup size, and scores.