Thinking Geospatially: How Variable Relationships with Reading Achievement Test Scores in the State of Missouri Vary by Geospatial Location.

摘要

The purpose of this was to examine whether district-level factors had an effect on literacy achievement, and whether the relationships between district composition factors and literacy achievement vary by geographic location, in this case the state of Missouri. The dissertation determined if the variable relationships were significant and if the statistically significant relationships were geographically nonstationary. Stationarity refers to the idea that relationships are stable across a geographic area. In contrast, nonstationarity indicates that relationships vary by geography. This means that relationships may be significant in one geographic area of the state but not significant in another. By including geographic location in the analysis, this dissertation is additive both to the literature on variable effects on literacy achievement as well as policy debates around literacy achievement. This study consisted of secondary data analysis using variables provided by the Missouri Department of Elementary and Secondary Education (MODESE) and representing all public schools in Missouri. The study analyzed these variables using Geographic Weighted Regression. This modeling technique was used to examine relationships and moderating effects of district demographic and socioeconomic composition. The dependent variable was literacy achievement as measured by scale scores on the Communication Arts MAP test for grades 4 and 8 and the English II end-of-course exam for high school students. By using geographic location as a variable and GWR as an analytic technique, it was possible to examine variable relationships between district demographics, socioeconomic composition and literacy achievement measures statewide. This showed whether and how the relationships between variables, including demographics and socioeconomic factors, vary across the state. That is, the study determined whether the relationships are significant and how that significance varies across the state. Overall, geospatial location was found to be an important factor in interpreting variable effects on literacy achievement. Significant relationships were found across grade levels and variables, and nonstationarity was observed.