In an attempt to provide improved First Time in College (FTIC) enrollment predictions, this study followed exploratory techniques seeking to differentiate yield rates among FTIC applicants based on one or more of the following source characteristics: Geographical Location, Single or Multiple Institution applications, Race/Ethnicity and Sex, High School GPA, and Admissions Test Scores at the University of South Florida (USF), in Tampa, Florida. The sample included 33,692 FTIC applicants to USF between fall 1996 and spring 2000. Based on historical yields of admits to USF among the various possible groupings and interactions of the preceding variables, a set prediction techniques was developed and tested on historical cohorts. Error analyses showed that global prediction techniques were between 1.5 and 5 times more efficient than more detailed approaches involving specific characteristics. Three global techniques were applied, with very small error differences among them for fall or summer semesters. Specifically, a precipitous decline occurs between yields of those who apply only to USF (57%) and those who apply to at least one other SUS institution (32%). Yields continue to decline as the number of applications to different institutions increases. USF experienced higher yield rates of admits among those with lower high school GPA. No test score differences occurred. Although no yield differences occurred between sexes, almost all minority group admits exhibit higher yield rates than white admits. It appears that the use of geographical location may be the most efficient predictive approach due to less variability than GPA groups, which also showed predictive efficiency. (Contains 16 tables and 13 figures.)