Panel Paper: To Play or Not to Play? The Effect of Summer Enrollment on College Outcomes

Thursday, November 3, 2016 : 3:00 PM
Columbia 2 (Washington Hilton)

*Names in bold indicate Presenter

Yuen Ting (Vivian) Liu, Columbia University


While summer learning loss is a well-documented problem in K-12 education, no evidence currently depicts the effect of summer idling or summer school for college students. My calculation from the ELS data showed that four-year institutions lose about 12% of their first time undergraduates after the first two summers. The percentage is double for non-selective school and among students with low socioeconomic status (SES). The college retention research has focused so far on interventions during the academic year, yet it has not considered the role of summer courses in improving student achievement and keeping momentum to return the next fall. Using state administrative data, I employed an instrumental variable approach to examine the effect of summer courses on students’ retention, academic achievement, and degree attainment.

The K-12 evidence showed that summer loss is cumulative and is highest among students with SES. Cooper et al. (1996) found that summer learning loss is equivalent to at least one month of instruction. Furthermore, the summer melt literature reveals that 8-14% of high school graduates decide not to enroll in the college to which they have already gained acceptance (Castleman et al., 2014). Together, the literature indicates that the lack of summer education exacerbates the achievement gap and may affect college students.

Evaluation of K-12 summer interventions is mixed depending on programs’ content and participants’ characteristics. Cooper’s meta-analysis of 93 studies (2000) revealed that on average summer intervention raises achievement scores between one-seventh and one quarter of a standard deviation. Yet, they observe that most studies conduct a simple pre-post comparison with limited internal and external evidence.

I used administrative data containing students entered to the public higher education system of a small state in the fall of 2005, 2006, or 2007. All the students are tracked through summer of 2012. My sample contains approximately 16,000 four-year students. The summary statistics feature a large enrollment gap between spring and the following fall. The enrollment lost is 11% and 5% in the first two summers respectively.  About 20% of students enroll each summer, with 50% of the overall cohort ultimately enrolling in summer courses. These classes tend to be in the field of mathematics, health care, and business. Also, students with lower high school grade point averages tend to enroll in remedial classes in the summer.

Using distance to the closest four-year college as an instrumental variable, I found that each summer credit enrolled yields a 3%increase in degree completion. Though summer courses do not increase GPA nor shorten the time to degree, they have a strong and positive effect on enrollment in the following fall. The results indicate that summer enrollment may keep student momentum and increase retention.

My study has important policy relevance for the college completion agenda. If summer classes help students to stay in school, institutions can increase retention with a minimum cost since classrooms are largely unused in the summer. The Pell grant program should also consider allocating separate funding for summer classes to help low SES students during summer.