Document type: DU ETD
Collection: Geology Theses  
Title A Study of the Differences Among Freshmen Persisters, Voluntary Withdrawals, and Academic Dismissals at the University of Denver Based on Enrollment Data
Author(s) Landi, Jose Alberto
School/Department Department of Geography
Institution University of Denver
Degree Type Doctoral
Degree Name Ph.D.
Type of Resource text
Degree Date 1981 November
Digital Origin reformatted digital
Rights Statement All Rights Reserved
Reason for Restrictions No restrictions
Type of Restriction No restrictions
Keyword(s) Geography
Genre Dissertations, Academic
Abstract The problem addressed by this study was as follows: Are University of Denver freshmen persisters, voluntary withdrawals, and academic dismissals different on entrance variables? If they are different, can the differences be explained, and is it possible to predict group membership? Of the freshmen who entered the University of Denver in the fall of 1979, 164 persisters, 140 voluntary withdrawals, and 49 academic dismissals were compared on 14 pre-enrollment variables routinely collected in institutional files. The data were analyzed using stepwise discriminant function analysis in order to test group differences, the contribution of variables to those differences, and the predictive accuracy of a model built on pre-enrollment variables. The analyses performed showed that persisters and voluntary withdrawals were so similar on the variables utilized that a more meaningful analysis of the contribution of variables to group discrimination, and a more powerful prediction model, should use the more homogeneous group definition of the combination of (persisters + voluntary withdrawals) versus academic dismissals. The most important findings of this study were: 1. The similarity of freshmen persisters and voluntary withdrawals on entrance characteristics. 2. The interaction effects of "distance away from home," and the extremely different profiles of commuter and non-commuter academic dismissals at the University of Denver. 3. The importance of financial aid, student age, and socioeconomic characteristics to separate academic dismissals at the University of Denver from other freshmen, over academic variables expected to hold much stronger discriminating power, such as high school rank and standardized test scores. This finding challenged the criteria for admissions traditionally used by the university. 4. The possibility of building a fairly strong prediction model for academic dismissal at the University of Denver, based only on pre-enrollment data. This study developed a prediction model that, when separating academic dismissals from other freshmen, yielded overall percentages of correct classifications of 82.8% for the sub-sample of commuters, and 76.7% for non-commuters. These results allow the practical use of the model to build a "high-risk" group in order to apply retention, orientation, and counseling strategies on potential academic dismissals.
Attached Files
Name Description MIMEType Size Downloads
du_mas_1981_Landi.pdf   du_mas_1981_Landi.pdf application/pdf 4.64MB 0

User Comments
Access Statistics: 0 Abstract Views, 0 File Downloads Detailed Statistics
Created: Mon, 15 Nov 2010, 10:12:43 UTC by John Adams . Detailed History