Learning communities and first -generation college students: A mixed method study of student retention, peer learning, and faculty engagement

For decades colleges and universities have been perplexed by the problem of retaining students within systems of higher education. With the advent of more diverse student populations and the increasing demands for more innovative pedagogical approaches, many universities have implemented learning community programs. To study how learning communities impact first generation college students, this study was undertaken to formulate a better understanding of student retention, peer learning, and faculty engagement. This was accomplished by drawing upon both qualitative and quantitative research methods to explore the data concurrently to derive a comprehensive picture of the phenomena under scope. Using a grounded theory approach for the qualitative data analysis, 24 interviews were conducted, an exhaustive document review transpired, and the researcher engaged in 32 hours of observation for a 16-week period. From this analysis, four primary themes emerged: psychosocial integration, intellectual integration, familial integration, and ecological integration. In addition, stages of peer learning were developed from the observations of the students in the learning community setting, as well as positions of faculty engagement in the learning communities program. From the grounded theory analysis, a socialization model was built to explain first year student retention. The study also employed a logistic regression analysis in the quantitative component of the research investigation, to determine how well the following variables could predict first year student retention: high school grade point average (GPA), first semester GPA, first year academic standing, gender, ethnicity, admission status, major decidedness, ESL status, and the number of developmental courses needed upon enrollment in college. A total of 900 archival student records were examined using a forward logistic regression and a 3-predictor model of student retention was yielded with a classification accuracy of 80.1%. The three significant variables in the model included: first semester GPA, first year academic standing, and ethnicity. When this model was applied to a cross validation sample (n = 685), the classification accuracy was found to be 82.0%. Overall, the findings generated from this study help to shed light on factors that might be predictive of student retention among first generation college students in a learning communities program