Abstract
This study examines survey and transcript data for a stratified random sample of over 4,000 students enrolled in the nine campuses of the Los Angeles Community College District (LACCD) and found that several campuses predict small but statistically significant differences in unit completion ratio (defined as the ratio of units [or credit hours] completed with a grade of A, B, C, or pass to units attempted) after controlling for influences of student characteristics (p < .05). Furthermore, split-file data analyses indicate that some campuses foster higher rates of completion for specific types of student subpopulations—such as first generation or nonnative English speakers—after controlling for other student characteristics. Limitations, implications for practice, and recommendations for further study conclude this article.
Notes
*Percent of cases deleted before regression analysis ranged from 17% to 26% over all nine campuses. Thus, no single campus suffered disproportionate cuts during trimming of the TRUCCS sample, and analysis results should still reflect dynamics of the general population.
*Variables significant (Sig.) at the final step of the model (p < .01) are in bold face.