A Retention Issue-Predicting The at Risk Student in Community College Web-Based Classes
Date of Award
Doctor of Philosophy (PhD)
Graduate School of Computer and Information Sciences
Steven R. Terrell
Marlyn Kemper Littman
This report describes a quasi-replication of an earlier study. The problem of this study was the need to develop an assessment tool that would assess and predict whether Web based learners at the community college were at-risk of failure in this mode of learning. In this study the instrument was used to identify factors that could then be used to discriminate between 276 successful and non-successful Web-based learners at the community college level. Twenty-eight ordinal-level questions, as used in the previous study, and eight more items related to computer and Web-based skills produced seven factors using factor analysis. The seven factors and seven background variables from the original study were used as input for additional quantitative analyses. Discriminant function analysis produced a significant discriminant function and five variables that contributed significantly to that function: Grade Point Average (GP A), Study Environment, Age, Last College Class, and Background Preparation. The function was used to classify (predict) student membership into successful and non-successful groups and classified two-thirds of the cases correctly.
This study also presented results from a qualitative investigation into dropout of the Web based learner. Twenty-two randomly selected students who dropped their Web-based course were interviewed and each was questioned about their reasons for dropping the course. The reason given most often for dropout was that the student could not obtain, access, or install all the required learning materials in a timely manner, and that he/she dropped the course while a chance to do so was still available. According to the findings of this investigation, students who had a history of academic achievement, were older, had a positive learning space, and believed they were prepared for this learning environment were more likely to be successful than others who had lesser amounts of these qualities. The study also showed that students who could react quickly to logistical demands early in the course were more likely to persist. Recommendations for further research included using more questions to characterize the domains studied in this inquiry, using another statistic to compute the likelihood of success and failure of the Web-based student, and testing students in other populations.
Herbert E. Muse Jr.. 2003. A Retention Issue-Predicting The at Risk Student in Community College Web-Based Classes. Doctoral dissertation. Nova Southeastern University. Retrieved from NSUWorks, Graduate School of Computer and Information Sciences. (742)