The second model focused on the dependent variable of self-reported grade point average (GPA) of students. Of the questions asked in the survey the second model showed eight variables that were statistically significant predictors of student GPA, b = -.397, t(110.066) = -5.767, p < .001. These eight survey questions explain a significant portion of the variance for students’ self-reported grade point averages, R² = .359,F (1, 178) = 33.261 , p < .001. As shown in table 4, some of the questions that were identified as self-predictors of GPA scores were tied to prior learning experiences (e.g., questions 1, 2, & 8). Other questions were related to the students’ personal circumstances such as family and time managements (e.g., questions 4, 5, 6, & 8).

Question three, which is about the need for face-to-face interaction when new material is being taught appeared as a significant predictor of both self-reported GPA and success in online courses. The presence of question 3 as a self-reported predictor of success in online courses and also as a predictor of GPA may indicate the importance of interactive structures within the online learning management system. These findings echo those of recent studies on predictors of student success in online education and the how the absence of coaching and support lead to high dropout rates (Allen & Seaman, 2013; Allen & Seaman, 2017; Hung et al., 2010; Online Learning Consortium, 2018). One notable difference between the two models is that in the first model, the 11 question accounted for 68% of the predictability of self-reported success in online courses. For the second the eight questions accounted for 59.9% of the predictability of self-reported GPA scores.