First-year academic success: A prediction combining cognitive and psychosocial variables for Caucasian and African American students

Journal of College Student Development, Nov/Dec 1998 by Ting, Siu-Man Raymond, Robinson, Tracy L

Among the effective predictor variables, the findings revealed two important differences between men and women. First, among all women, high school GPA explained 18.4% of Fall semester GPA, which is more than that predicted for men (15.3%). This finding implies that high school GPA is a stronger predictor for women than for men. Variables that predicted Fall GPA for all males and females are as follows: high school GPA, father's education level, PDI, positive self-concept, and preference for long-term goals.

When the data were disaggregated by gender and ethnicity, the variances of the prediction models for Fall GPA varied between 8% and 29%: 25.2% for Caucasian female students, 18.3% for Caucasian males, 29% for African American males, and 8% for African American female students. The findings revealed a few important differences in the predictors among different student groups. First, among Caucasian women, high school GPA alone explained 19.3% of Fall semester GPA. Second, the housing variable loaded on the prediction model for Caucasian female students whereas it did not for all female students. Third, among Caucasian men, musical ability and age loaded on the prediction model, which also explained their Fall academic performance. Fourth, in the prediction model for African American male students, high school GPA has been removed. The data indicated that once high school GPA was added to the prediction equation, the amount of the variance explained was reduced. For example, high school GPA and SAT Mathematics scores explained 24% of Fall semester GPA. Thus, although it is a statistically significant predictor, high school GPA was not the best predictor of African American males' academic performance during their first semester of college, as it was for Caucasian females.

Fifth, high school GPA and father's education were found to have explained only 8 % for African American females' Fall GPA.

Predicting Spring GPA. Most of the findings were similar to the Fall predictions. Among different student groups, similar prediction models ranged from 16% to 25%. High school GPA was still the strongest predictor; however, the variance explained was smaller (e.g. for all students, 11.8% in spring vs 15.3% in Fall). More noncognitive predictors were found in the Spring semester. For example, among all students, the variance accounted for 5% of the total variance including selfconcept on the NCQ, and SCIMP, father's education level, and PDI on the FYSS.

Among Caucasian students, high school GPA was still the best single predictor. SAT-total scores explained only less than 1% of the total variance. A few noncognitive variables were significant predictors such as self-concept and preference for long-term goals of the NCQ; planning for the highest level of education to be reached, SCISKILL, and planning for work hours on the FYSS, and the demographic variables of age and gender were significant predictors. They made up almost 7% of the total variance. The common indicators for GPAs between the Fall and the Spring predictions for female Caucasian students were high school GPA and father's education level.