Browsing by Subject "Statistical significance"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Connecting students' homework to their participation in a course-based social network
    (2017) Gavrin, A.; Lindell, R.S.; Physics, School of Science
    This paper presents a comparison between students' efforts on homework (problem sets delivered and completed online using WebAssign) and their participation on a course-focused social media site. The social media platform, CourseNetworking (CN), has many features typical of Learning Management Systems (LMSs), but is distinct in several important ways. The interface is far more "student centric" than traditional LMSs, and is designed to increase engagement; most of the CN window is devoted to student-authored content. Also, the site measures and "gamifies" participation, using an algorithm that includes posts, completion of surveys, comments on other students' posts, and other actions. The setting for our efforts was an introductory calculus-based mechanics class enrolling approximately 150 students, most of whom were engineering majors. Course exams, problem sets, and labs followed a traditional model. Social media participation was not required, but it was encouraged and students could earn a small extra-credit bonus. We investigated correlations between social media "micropoints" and three variables associated with the homework: time on task, points earned, and assignments skipped. Our results show small to moderate correlations and statistical significance in all three cases. Pearson's correlation coefficients are r = 0.286, 0.444, and -0.436 for time on task, points earned, and assignments skipped, respectively. The associated p values are 1.2 × 10-3 for time on task, and p < 10-5 for the other two variables. Because the variables we measure are not normally distributed, we verify these results by also calculating Kendall's tau statistic. This analysis confirms both the size and significance of the correlations we observe. We do not suggest a causal connection; rather, our conclusion is that participation in the social network is a form of engagement with the class comparable to traditional measures of engagement such as homework effort and outcome. © American Society for Engineering Education, 2017.