| George Terhanian
NCES
School Policies and Practices, Student Proficiency, and Racial Differences in Proficiency: Evidence and Implications of a Mulitlevel Analysis of the Reading Proficiency of 4th Graders from Pennsylvania and New York
FINAL REPORT:
The object of this study is to apply sophisticated statistical methods, notably, multilevel modeling, on data from the Trial State Assessment (TSA) of the National Assessment of Educational Progress (NAEP) to cast light on the relationship of a variety of school factors and achievement. Specifically, this study links data from the 1994 Trial State Reading Assessment, the Common Core of Data, and the Pennsylvania Educational Policy Studies data base in the interest of understanding whether class size, ability grouping, several dimensions of teacher quality, and per pupil expenditures are associated with two outcome measures: (1) the reading proficiency of public school students in Pennsylvania and New York, and (2) and racial differences in the reading proficiency of public school students in both states. Multilevel analyses simultaneously exploit student and school data to reduce the possibility of aggregation bias - a potential limitation (e.g., see Hanushek, Rivkin, & Taylor, 1996) of recent meta-analytic work (e.g., see Greenwald, Laine, & Hedges, 1996) that has considered the relationship between some of these inputs and student proficiency. Expenditure data are adjusted to account for geographical cost of living differences, and student special needs differences - two potential sources of distortion that may have biased results of earlier research on expenditures and student proficiency. Key findings are as follows:
In Pennsylvania
- Each $1000 increase in per-pupil spending is associated with a five point increase, on average, in school reading proficiency.
- Each five-student decrease in school class size is associated with a 3.2 point increase, on average, in school reading proficiency.
- Schools with highly experienced teachers score eight points higher, on average, than schools with less highly experienced teachers.
In New York:
- Each $1000 increase in per-pupil spending is associated with a four point increase, on average, in school reading proficiency.
- Schools with highly experienced teachers score 7.4 points higher, on average, than schools with less highly experienced teachers.
- Schools with highly certified teachers score 12 points higher, on average, than schools with less highly certified teachers.
- In schools with highly certified teachers, white students outperform non-white students by 8.5 po~nts, on average.
The implications of these findings are considered in the body of the paper.
Back to Funded Research Fellows Page |