Model-Based Clustering of Mathematics Achievement Data in Grades 6-11
Authors:Susan Deeb, Reina Galvez, Soeun Park, Nathan Robertson
- David Pagni, Professor, California State University Fullerton
- Sam Behseta, Professor, California State University Fullerton
In this project, we tackle the problem of capturing the effect of ethnicity on mathematical achievement, from K6-11 student data in Orange County, CA. We apply three clustering methods namely, hierarchical clustering, model-based clustering and the method of k-means. The primary goal of this research is to examine the role that socio-ethnic background may play on mathematics achievement. We utilize a data set that is comprised of students with varying levels of academic success, primarily affiliated with three ethnicities: Hispanic, Vietnamese, or other Asian. The measures of mathematical achievement are from standardized tests, California High School Exit Exam (CAHSEE) scores, and math course GPA. The major finding in this work is that regardless of the clustering technique used, the highest achieving cluster is predominantly formed of Vietnamese students. Reversely, the group with the lowest performance is mostly formed by Hispanic students.