Analysis of Variables to Predict First-Year Persistence Using Logistic Regression Analysis at the University of South Florida: Model V 2.0
Herreid, C. H., Miller, T. E., College and University
This is the fourth in a series of articles in College and University describing an attrition prediction and intervention project at the University of South Florida (USF) in Tampa. The project was first described in the 83(2) issue (Miller 2007). The statistical model for predicting attrition was presented in the 83(3) issue (Miller and Herreid 2008). The methods and approaches for intervening with students at highest risk of attrition were presented in the 84(3) (Miller and Tyree 2009).
In this article, the researchers will describe the updated version of the prediction model. The original model was developed from a sample of about 900 First Time in College (FTIC) students enrolling at USF in the summer or fall of 2006. Due to refined practices in survey administration and data collection, the subsequent model was developed with data from about 2,700 FTIC students, generating a model in which the researchers had more confidence.
It is common for colleges and universities to develop programs and activities for the purpose of enhancing student persistence ...
...but, commonly, such efforts are broadly applied, such as assistance programs directed at all freshmen. Some efforts are somewhat more narrowly applied, such as programs for first generation students or those on academic probation or who otherwise exhibit a single risk factor for attrition. However, even those programs typically provide assistance or support to a wide collection of students. Such widely applied programs may be sound for educational reasons; but, as persistence enhancement initiatives, they are often wasteful because they include many participants who would have remained enrolled at the institution without any treatment. Such programs cannot take advantage of the fact that multiple characteristics often predict attrition.
The project described is intended to identify individual students who are at risk based upon a variety of factors. Given the size of the data set, it actually considers several dozen variables from survey data and from institutional data. When individual students are found to be at risk, and there is confidence in the basis for the discovery, appropriate personnel can contact them and begin the process of developing plans to enhance the chance of persistence by the individual student (Glynn, Sauer, and Miller 2005). The efficiency and cost effectiveness of this approach, especially at large institutions, is appealing to involved administrators and managers.
This project also has appeal to practitioners because the model relies entirely on pre-matriculation characteristics of entering new students. That allows for a timely response to the students identified as at risk and gives the institution an opportunity to craft a legitimate early intervention program that can start even before the individual student gives any signal of disconnecting or disengaging from the university, perhaps even before the student has given thought to dropping out.
In this article, we will describe the process of developing the predictive model and then present the specific elements of the model. We will also describe the process of developing interventions on behalf of individual students found to be at risk. Finally, we will discuss the implications for differences between the first model and the later one, and we will present aspects of future research that we will undertake.
Data for the study were obtained from university databases and the College Student Expectations Questionnaire (CSXQ) (Gonyea 2003; Kuh and Pace 1998), which was administered to incoming FTIC freshmen in the fall of 2007. Participants in the study were 4,212 new FTic students who enrolled at the main campus of the University in the fall of 2007 and also new FTic students in the summer 2007 term who returned for the fall. Of these students, 57 percent were women and 37 percent were minority. More than 95 percent …
Questia, a part of Gale, Cengage Learning. www.questia.com
Publication information: Article title: Analysis of Variables to Predict First-Year Persistence Using Logistic Regression Analysis at the University of South Florida: Model V 2.0. Contributors: Herreid, C. H. - Author, Miller, T. E. - Author. Journal title: College and University. Volume: 84. Issue: 4 Publication date: Spring 2009. Page number: 12+. © American Association of Collegiate Registrars and Admissions Officers Fall 1999. Provided by ProQuest LLC. All Rights Reserved.
This material is protected by copyright and, with the exception of fair use, may not be further copied, distributed or transmitted in any form or by any means.