Historically, calculation of college student retention by colleges, state agencies, and the federal government has been based on tracking enrollment and attrition of an 18- to 24-year old population of first-time full-time students in a fall-to-fall pattern. Use of this "native cohort" assumes that students enroll in college on a full-time basis immediately following high school graduation and proceed on an unbroken progression toward a bachelor's degree. College attendance patterns have changed substantially, however, so that use of this native cohort no longer accurately reflects student persistence, particularly for institutions with a substantial proportion of non-traditional students (e.g., transfers, part-time, or adult-status).
Using a sample of 9,029 students enrolled over a seven-year period, this study examined differences between persistence outcomes for four student cohorts at a public liberal arts university (a) full-time, first-time students ("native" cohort), (b) full-time, transfer students, (c) part-time, first-time students, and (d) part-time, transfer students. Persistence was defined by a ratio algorithm that summed the number of semesters each student was enrolled for at least one credit divided by the total number of semesters of availability. Any student who graduated received a persistence score of 1.0. Across all four enrollment groupings, the persistence ratio averaged .69; full-time, transfer students were most likely to persist ( M ratio = .74) and part-time, first-time students were least likely to persist ( M ratio = .57). These differences remained highly significant after other potential influences on student success (e.g., GPA after first year, SAT scores) were covaried out. These differential outcomes also were similar for men and women and for students who reported having at least one parent who had obtained a college degree versus those whose parents had not completed college. Prediction of persistence from multiple student-input and college-environment variables yielded beta coefficients that were generally similar across the four cohorts.
