This study investigated the efficacy of noncognitive variables and academic background for the prediction of college grade performance and persistence. A total of 7,377 new first-year college students participated in the study, completing demographic questionnaires at the start of the school year. Data were also collected on American College Test (ACT) composite scores, high school rank, cumulative grade point average (GPA) after 1, 2, and 4 years in college, and enrollment status after 2 and 4 years in college. The study found that the combination of ACT composite scores and high school class rank alone explained 26 percent of the variance in college GPA after 1 year and 26.5 percent of the variance after 2 years. This variance effect was more pronounced for white students than for Hispanic, Asian, and black students. The study also found that students with higher financial and social goals exhibited, on average, lower grades and persistence rates than students with lower financial and social goals. The effects of ethnicity, gender, parent educational background, desire for recognition, and achievement expectancies are also considered. (Contains 39 references.) (MDM)