Factors influencing the success of students in introductory algebra at a community college

Using a sample of 498 students enrolled at a large, urban community college in Texas during the Fall semester of 2001, this study examined the demographic, aptitude, instruction, and environment factors that influence a student's ability to successfully complete Introductory Algebra. Fourteen independent variables representing seven of the nine factors from Walberg's Learning Productivity Model were employed in multiple regression and discriminant analyses. The final grade average in the class served as the dependent variable to determine success.;For the multiple regression analyses, the independent variables entered using the setwise method and the stepwise forward procedure method were gender, ethnicity, age, employment status, ABST score (mathematics achievement), mathematics prerequisite status, cumulative college GPA, level of high school mathematics courses, attitude toward mathematics, amount of time in class, instructor, class location, class size, and course load. The variables that significantly contributed to success in the setwise analysis listed in order of influence were: (1) college GPA, (2) mathematics prerequisite status, (3) ABST score, (4) gender, (5) attitude toward mathematics, (6) ethnicity, and (7) instructor. All the independent variables accounted for 40% of the variance in the final course grade, generating an R 2 = .402 and F = 23.148 (p < .01).;For the discriminant analysis, the independent variables were entered together to examine the independent contribution of each in determining membership into the successful or unsuccessful group. Successful completion of the course was defined as earning a final numerical grade average of 70 or greater in the course. The model produced a Wilks's Λ (df = 14, 483) of .70, which was statistically significant at the p < .001 level. Mathematics prerequisite status, college GPA, AB ST score, and attitude toward mathematics contributed the most influence in determining group membership. The canonical correlation for the discriminant function was .55, representing a moderate association between the groups and the discriminant function. The squared canonical correlation coefficient for the model was .30, reflecting that 30% of the variance between the two groups could be explained by the variables in this model. A classification summary indicated that 76% of the cases were classified correctly.