A three year study of the relationship of prematriculation variables on student success at a private, Southern university
University of Alabama at Birmingham, Birmingham, AL, Doctoral Dissertation, (1999)Аннотация
This study investigates the prematriculation variables gathered from the application for admission at a private, Southern university and determines if any significant and predictive relationship exists between the variables and freshman grade point average. The research population was the 1995–1997 freshman classes (N = 2,045) of the university with 43 independent variables studied. Multiple regression was the statistical procedure used to determine if gender, race, religious affiliation, ACT Composite score, type of high school attended, high school grade point average, core high school grade point average, rank in class, living accommodations in college, state of permanent residence, high school leadership, family income, employment, legacy or college major individually or in combination with another, were related to the freshman grade point average of the student. The correlation produced 25 statistically significant variables, with 14 having the greatest individual relationship (r \textbackslashtextgreater .10) to be used in the prediction model. The forward and backward stepwise regression procedure was used to select the model the best predictive model of freshman grade point average. The model selected included the independent variables of ACT Composite Score, core grade point average, housing status, high school grade point average, rank in class and church leadership. The model was cross validated with the 1998 freshman class actual data to determine accuracy for prediction within the critical grade point averages for freshmen.