An analysis of student persistence at Mid-State Technical College
Johnson, Michael A.
University of Wisconsin--Stout
MetadataShow full item record
This study analyzed student persistence at Mid-State Technical College by first measuring the accuracy of the predictions made by a newly acquired software that uses available student data to predict how likely they are to persist. To do this, the software's prediction scores for degree seeking students at the college were captured every two weeks throughout the Spring 2018 semester. These prediction scores were then compared to actual student persistence rates (those that either enrolled in the Fall 2018 semester or graduated). The accuracy of the software's predictions were measured by calculating the correlation coefficient (R2) between the software's prediction scores and actual student persistence at the college. R2 values above 0.95 were considered very strong, with values above 0.90 still considered strong. An R2 value below 0.90 was considered to be a weak correlation. Once the software's accuracy was determined, further analyses compared persistence among different demographic groups and programs at the college to identify areas of opportunity for improved student success.