School Effects Indices: Stability of One- and Two-Level Formulations

Three residual-based methods for indexing school effectiveness were compared using data from elementary schools in Maryland. Two of the methods were regression-based procedures at the school level, and 1 was a hierarchical linear model approach, with Level 1 at the student level and Level 2 at the s...

Full description

Saved in:
Bibliographic Details
Published in:The Journal of experimental education 2000-01, Vol.68 (3), p.239-250
Main Authors: Schafer, William D., Yen, Shu Jing, Rahman, Talsima
Format: Article
Language:English
Subjects:
Achievement tests
Correlations
Educational evaluation
Educational research
Educational sciences
Effective Schools Research
Elementary Education
Elementary Schools
Grade 3
Grade 5
Grade levels
Hierarchical Linear Modeling
Least squares
Maryland
Measurement, Statistics, and Research Design
Methodology
Methods
Modeling
Multilevel models
Observation, experimentation
Reliability
School Effectiveness
School size
Special education
State departments of education
Students
Variables
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Three residual-based methods for indexing school effectiveness were compared using data from elementary schools in Maryland. Two of the methods were regression-based procedures at the school level, and 1 was a hierarchical linear model approach, with Level 1 at the student level and Level 2 at the school level. The 2 methods were applied to a data source in which schools were evaluated separately at Grades 3 and 5 using 3 forms of a statewide performance assessment that were administered to randomly formed groups within each school. The hierarchical linear model approach had higher stability across test forms, but the stabilities of the 1-level formulations were almost as high. The authors concluded that 1-level approaches are viable alternatives when the added data necessary for the hierarchical linear model approach are unavailable.
ISSN:0022-0973
1940-0683
DOI:10.1080/00220970009600094