Abstract:

Teachers have been systematically gathering attendance records, grading assignments, evaluate interaction and activity, etc., and in effect creating a systematic record of educational accomplishment. We propose an approach to extract information from records by analyzing them statistically, and interpret that information into useful knowledge about course outcomes. We provide several case studies in which we use statistical analysis to provide an unbiased understanding of attendance and grades data in relation to educational outcomes, or course final grades. The proposed approach provides insight into which graded educational activities were most indicative and informative in shaping final grades. It is robust to missing data such as late assignments, and scales from a single class towards the national level of education. Case study results indicate that it is possible to analyze attendance to review important lectures and examine generational trends if correlated records are available. Courses with significant in lab activities which can be related to other class assignments are susceptible to finer grained analysis of outcomes.

Keywords:

Statistics, computer science, education, data mining.