Abstract
The purpose of this study was to investigate the effect of academic variables (e.g., grade point average, major change, informal meetings with faculty) on six year persistence and attainment among black male students in community colleges. Data was collected from the Beginning Postsecondary Students Longitudinal Study and was analyzed using logistic regression. Findings from this study indicated that select variables reflective of academic integration serve as significant predictors of black male persistence and attainment in the community college. These variables included students’ grade point average, whether a student ever received an incomplete, repeating courses for higher grades, withdrawing from courses after the add or drop deadline, and informal meetings with faculty. Implications for future research are provided.
Notes
1. Persistence refers to students’ continuation in college while attainment refers to meeting certificate, degree, and/or transfer goals.
2. The terms community college and public two-year colleges are used interchangeably.
3. The terms black and African Americans are used interchangeably.
4. High standard errors prevented the researcher from reporting data for Native Americans
5. Income percentile rank is calculated separately for dependent and independent students, and then merged onto a single percentile scale. For dependent students, income rank is determined based upon their parent’s. In contrast, independent students rank is based upon the students own income.
6. Odds ratios should be interpreted with caution, as they can seem to indicate stronger relationships than actually exist. For example, odds ratios of 2.10, and 3.10, could be interpreted as 110% greater odds or 210% greater odds, but also correspond to probabilities of 0.66, and 0.75, respectively. Probabilities can be calculated from odds ratios using the following formula: probability = odds ratio/(1+odds ratio).
7. PowerStats presents a cox-snell maximum likelihood ratio which represents the maximum values of the cox-snell for a logistic regression, had the model been perfect. The maximum value for both models was .743.