Community College Student Success in Developmental Mathematics Courses: a Comparison of Four Instructional Methods

The student success rates for three developmental mathematics courses (prealgebra, elementary algebra, and intermediate algebra) taught through four instructional methods (lecture, personalized system of instruction [PSI], hybrid, and online) were examined. The sample consisted of 9,211 students enrolled in a large Texas community college from fall 2009 through spring 2011. Student success was defined as a grade of C or better. Chi-square tests were used to compare the three developmental mathematics courses success rates. Statistically significant differences in student success were found between all four methods of instruction for all three mathematics courses (prealgebra: χ2 [df = 3] = 107.90, p < 0.001; elementary algebra: χ2 [df = 3] = 88.39, p < 0.001; intermediate algebra χ2 [df = 3] = 254.18, p < 0.001). Binary logistic regression modeling was used to determine to what extent age, gender, ethnicity, residency, Pell eligibility and mode of instruction accounted for the community college students’ course success for each of the three developmental mathematics courses. For prealgebra, the independent variables of gender, race, age, residency, and mode of instruction made statistically significant contributions to the model (χ2 [df = 14, n = 1,743] = 159.196, p < .001; Nagelkerke R2 = .119), with greater success among female, White, younger, out of country students taking the course through lecture. For elementary algebra, the independent variables of race, age, residency, and mode of instruction made statistically significant contributions to the logistic regression model (χ2 [df = 14, n = 2,731] = 816.223, p < .001; Nagelkerke R2 = .358), with greater success among , younger, out of country students taking the course through lecture, hybrid or PSI. For intermediate algebra, only race and Pell eligibility made a statistically significant contribution to the logistic regression, with greater success among White, Pell-eligible students, and mode of instruction did not contribute significantly to the model (χ2 [df = 14, n = 3,936] = 53.992, p < .001; Nagelkerke R2 = .019). Recommendations for research and implications for practice are provided