This applied dissertation was designed to evaluate the persistence-risk identification processes at a Midwestern technical college, with an eye toward developing a model prototype decision-support tool (PDST) to identify students at risk for not persisting through a degree by using data routinely collected in the prematriculation process. The persistence-risk identification process at the college was identified using a validated survey tool based on a review of the literature and data collected during the college registration process. The validated survey tool was used to identify factors related to persistence at the technical college studied in the dissertation. The writer surveyed subject-matter experts to identify perceptions related to factors associated with student persistence and then compared the perceptions to actual data from a subset of business-division enrollments at the college. The research used both evaluation and research methodologies, with the evaluation methodology used to identify persistence-risk factors and current processes and the research methodology used to develop the PDST. Deidentified data from 400 students on 12 independent variables was analyzed using logistic regression to predict student persistence, the single dependent variable. Half of the data set was used to develop the logistic regression model; the other half was used to assess the predictive efficacy of the regression model. The analysis of the data resulted in a logistic regression model that used 4 of the 12 variables. The model was able to correctly classify 116 of the 200 deidentified students as persisters or nonpersisters. Recommendations were made to collect additional new information on the college registration form to improve the predictive efficacy of the model, based on factors identified in the literature review.
| Date of Award | Jan 1 2008 |
|---|
| Original language | English |
|---|
| Supervisor | Diane E. Bryant (Supervisor), Marian Gibney (Advisor) & Maryellen Maher (Advisor) |
|---|