Return to search

Developing and Testing an Early Warning System to Improve High School Graduation

The nation has placed a spotlight on improving graduation rates for all students. The current study analyzed retrospective, longitudinal student data from the fifth largest school district in Oregon to create an Early Warning Indicator System (EWS) to identify students on track to graduate and those who are not. This study creates an EWS system using the student demographics and the ABC’s of (a) attendance, (b) behavior, and (c) coursework to identify students who are on track and those who are not.

I employed logistic regression model to build a prediction model using middle school data (N = 2,041) that examined predictors established in sixth through eighth grade with high school graduation. The dependent variable, four-year graduation was coded as graduate or non-graduate. The independent variables were (a) gender, (b) race, (c) ELL status, (d) SPED Status (e) attendance rate, (f) ODR’s, and (g) number of F’s in English Language Arts and Mathematics.

Attendance rate was the strongest predictor of high school graduation. Overall the model predicted graduates with 89.7% accuracy and non-graduates with 33.6% accuracy with the total model predicting 69.5% of graduates and non-graduates.

Identiferoai:union.ndltd.org:uoregon.edu/oai:scholarsbank.uoregon.edu:1794/20440
Date27 October 2016
CreatorsPhinney, Robi
ContributorsBullis, Michael
PublisherUniversity of Oregon
Source SetsUniversity of Oregon
Languageen_US
Detected LanguageEnglish
TypeElectronic Thesis or Dissertation
RightsAll Rights Reserved.

Page generated in 0.0019 seconds