Association for Public Policy Analysis & Management

Historically a large proportion, over 35%, of undergraduates stop-out (Horn and Carroll, 1997). Cross-sectional analysis reveals that factors correlated with persistence also relate to stop-out rates; for example, stop-outs are more likely to be 30 years old or older, be married, and to have children living in their home (Horn and Carroll, 1997; Schatzel, Callahan, Scott, & Davis, 2011). However, it is not necessarily the case that stopping-out negatively impacts persistence. Towards this point, it is ubiquitous for students to dis-enroll from their home institution to study abroad or to participate in special programs at other institutions (McCormick, 2003). Similarly, growth over the last decade in the number of non-White undergraduates and older and returning students (Strohl & Carnevale, 2013; Taniguchi & Kaufman, 2005) may correlate to growth in overall the stop-out rate (since stop-out rates are higher in these populations than in the overall population) (Herzog, 2005; DesJardins, Albergh, & McCall, 2006). While previous work reveals that stopping-out is commonplace, it is dated and does not reflect the current demographic make-up of the undergraduate class. Current descriptive research characterizing the rate and nature of stop-outs is needed. Towards this broader aim, this paper will provide a descriptive portrait of stop-out patterns at the nation’s largest broad-access public four-year university and address the relationship between stop-outs and persistence. Specifically, this paper answers the following questions:

 (1)      What types of stop-out behaviors do first time freshman at broad access four-year universities engage in (is stopping-out common, when does it happen, how long do stop-outs typically last, etc.)?

a.         Do stop-outs behaviors differ, on average, among students of color than in the general population of first-time freshman at broad access four-year universities?

(2)       Is the length of a stop-out associated with the likelihood that student will return?

Data for this study come from administrative data files that comprise a census of undergraduates enrolled at CSU from the 2002 until the 2012 school year. A longitudinal cohort approach is used for analysis. Cohorts are comprised of eight consecutive semesters after participant’s first semester enrolled in classes (Fall 2008); robustness checks are performed to address censorship concerns (Singer and Willett, 2003). In answering research question 1, this paper provides a descriptive portrait of the length, frequency, and timing of initial departure. Frequency distribution graphs and box and whisker plots are used to illustrate distribution and central tendency. Survival analysis—a life table and hazard function—are used in research question 2.

Preliminary findings are that stop-outs are ubiquitous among undergraduates at public four-year universities, however not as common as demonstrated in previous research using national samples. Another key finding is that stop-outs are much more common after the first year. This is interesting given that study abroad is most common in undergraduates’ junior year; the stop-outs observed here may reflect behaviors that do not threaten persistence. More research is needed to understand why some students stop-out, return, and eventually graduate, while others do no return after a stop-out.