DETERMINING CRITICAL COURSES THROUGH STUDENT PERFORMANCE
Vanda Johnson1, Gregery Buzzard2, Benjamin Wiles2.
1Savannah State University, Savannah, GA, 2Purdue University, West Lafayette, IN.
Every institution of higher learning has at least one funding model (formula) imposed on it by the state and federal government. The most common funding model is the enrollment model, in which institutions of higher learning receive funding based on how many full-time students are enrolled each academic year. In recent years an increasing number of states have adopted a new funding model called the performance-based model, in which funding is based on criteria such as the number of degrees awarded, time to degree completion, etc. In this research, we analyze patterns in student registration data to identify what we call critical courses: courses in which poor performance is predictive of delayed graduation. We analyze anonymized course data with statistical analysis and data mining methods, using both original grades and grades adjusted according to a regression approach that estimates course difficulty. Our procedure allows us to identify a relatively small set of courses in a given major which are highly correlated with delayed graduation. This information could be used to provide extra student support and/or better placement in these classes as a first step toward higher on-time graduation rates.