Date of Award


Document Type


Degree Name

Doctor of Education (EdD)


School of Education

First Advisor

Dane C. Joseph, PhD

Second Advisor

Susanna M. Thornhill, PhD

Third Advisor

Linda Samek, EdD


Each year a significant percentage of high school students in the United States do not graduate. School practitioners need accurate indicators for identifying potential dropouts in order to focus scarce intervention resources on students most in need. While the process of dropping out is complex, indicators measured at the end of students’ ninth-grade year provide information regarding their future graduation outcome. The current study used receiver operating characteristic (ROC) empirical curves to assess the accuracy of three ninth-grade risk factors, GPA, credits earned, and over-age status, in predicting the likelihood that students sampled for the National Center for Education Statistics High School Longitudinal Survey of 2009, dropped out of high school. The results showed that all three gave better than chance predictions. GPA had a 74 percent probability of correctly distinguishing between dropouts and graduates. The cut point of GPA less than 1.7 identified 48 percent of the dropouts, 88 percent of the graduates, and had a false positive rate of 12 percent. The three indicators provide quantitative data for identifying students at the end of ninth grade who may benefit from strategies designed to keep them on track for graduation. School practitioners may want to conduct a similar analysis using their district data to assess the accuracy of the risk factors for their specific student population.

Included in

Education Commons