(445i) Correlations and Reproducibility in Reading Participation and Auto-Graded Homework Completion When Using an Interactive Textbook for Material and Energy Balances

Combining interactive reading participation with online homework generate large, real-time data that can help quantify student engagement and learning. A fully interactive textbook for a material and energy balances course measures over 1,400 reading interactions per student as well as the fraction correct and number of attempts on up to 700 auto-graded homework questions during a semester. In total, seven cohorts representing over 600 students and over 700,000 reading interactions exhibit interesting correlations between reading participation and final course grades. The interactive content applies some best practices from learning sciences, including using visuals in animations, chunking content into organized sections and subsections, immediate feedback when using multiple choice, matching, and true/false learning questions, and varying the reading activities. Since reading participation is an effort-based grade, the correlation between reading participation and final course grades may be considered surprising. Also, when allowing unlimited attempts before homework is due, students earned a median correct of 91% or higher across six cohorts on auto-graded problems. The randomized questions are also used by students as a means of practicing problem solving with thousands of attempts after correct being logged in recent cohorts. Students final grades more strongly correlated with their scores on auto-graded problems than reading participation. Statistically significantly lower competition of auto-graded problems was measured for each subsequently lower grade, i.e., B students answer statistically more auto-graded problems than C students. Overall, these correlations provide data to weigh the importance of formative and summative assessments in a material and energy balances course.