Interactions in Online Courses and Student Academic Success

Abstract

The Sloan Foundation reported a phenomenal 17% annual growth in online course enrollment from 2002 through 2012, while the overall enrollment in higher education has shown only a 2.5% annual growth. Despite the growth in enrollment, McFadden and others have reported that drop out rates were as much as seven times higher for online courses. To address these concerns, this study investigated how three types of interactions (student-student, student-content and student-instructor) in the first two weeks of fully asynchronous online courses were associated with student academic success. This investigation analyzed archived tracking data from a learning management system for 1,703 students in 200 semester-long, fully online community college courses. Multinomial logistic regression analysis was used to consider the relationship between the three types of online interactions and academic success. The three types of independent interaction variables were measured as follows. Student-student interaction was the number of student posts to the discussion forums. Student-content interaction was the total number of pages the students accessed. Student-instructor interaction was a measure of the number of instructor posts and the number of instructor emails. The outcome measure, academic success, fell into three groups: successful completers (students completing the course with A, B, or C grades), low score completers (students completing the course with D or F grades), and non-completers (students who did not complete the course). Odds ratios derived from the regression analysis were used to determine the percentage of change in academic success attributable to each type of interaction, holding all other factors constant. The multinomial logistic regression was statistically significant (chi square = 461.96 p<. 001), indicating that the predictors reliably distinguished between the three outcome groups. The findings suggest that increasing the number of times a student posts by one unit (5 posts) would increase the individual's odds for success by 74% for the non-completers group and by 71% for those in the low score completers group. The findings also suggest that if an individual would increase interaction with the content by one unit (49 online pages) the individual's likelihood for success would increase by 57% for non-completers and 39% for low score completers. The number of instructor posts had no effect on the outcome for low score completers, however increasing the number of instructor discussion posts by one unit (15 posts) increased the likelihood that the students would complete the course by 34%. An increase in one unit (151 messages) of instructor email was associated with a 45% decrease in the odds for success for non-completers and a 28% decrease in the odds for success for the low score completers. This study provides support that student-student interaction and student-content interaction in the first two weeks of online courses contribute to student academic success with student-student interaction being most influential. More instructor postings in the discussion areas increased the likelihood that students would complete the course. The finding that the more instructors interacted with students by email the lower the academic success seems counterintuitive at first glance. This may be because instructor emails are often in response to increased requests for clarification by students and could be a reflection of poor course organization or insufficient course support materials.