Title

Determining the Relationship of Selected District and Student Demographics With Eighth-Grade Achievement In Alaska

Date of Award

1996

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Educational Studies and Research

First Advisor

Arthur R. Southerland

Advisor Department

Educational Studies and Research

Abstract

The purpose of the present study was to determine whether there was a relationship between the criterion variables of mean student achievement in mathematics and reading on the Iowa Test of Basic Skills Form G and predictor variables of student socioeconomic status, district type, and percentage of total district budget expended on instruction, administration, and operations and maintenance, as found in the state of Alaska. The data were collected from the 54 recognized school districts in the state and compiled by the Alaska Department of Education. Pearson-product moment correlation coefficients and canonical correlational analysis were used to analyze the data. The first canonical root (Rc =.777, $p<.001$) indicated that the variable sets shared 60% of their variance. Inspection of the canonical structure coefficients indicated that student socioeconomic status was the predictor variable most highly related to the criterion variables of reading and math achievement. Instructional expenditures, operations and maintenance expenditures, and administrative expenditures also contributed to the predictor variable composite. Canonical root II (Rc =.384, not statistically significant) indicated that the variable sets shared 15% of their variance. Inspection of canonical structure coefficients indicated that district type and administrative expenditure had a relatively large correlation with mathematics achievement. It was found that although students in rural districts scored lower in general than other students in the state, the scores were within the normal expected distribution. A bootstrap analysis was conducted to estimate replicability of the canonical results. All structure coefficients for the variables included in the bootstrap analysis appeared to be within the 95% confidence level for replicability, suggesting the stability of the findings.