This work investigated the effects of using a social media web application to assist learning and teaching in classrooms on learners' performance. We examined Facebook activities of 98 students, consisting of the groups of class of 4/1, 4/3 and 5/1 of Sa school located in Nan province, Thailand, during the first semester of 2011. We applied both statistics and data mining techniques for data analysis to determine the effects of Facebook usages for education on students' learning skills. Statistics was used to evaluate the correlation between students' Facebook usage and grade point averages, and a data mining technique was used to learn association rules from data. This project has helped us learn the effects of using social media web applications (i.e., Facebook applications) for education on student-learning performance. We found that examining the ratio of number of Facebook posts and comments for educational purposes to the one for non-educational purposes could help us draw a conclusion that Facebook usage does have an impact on students' learning performances. Specifically, we found that students who spent more time on education-related posting and commenting earned better grades than the ones who did the opposite. Furthermore, as we have to group data to change numerical data to nominal data, we found that data grouping play an important role on the number of rules and their associated confidence values. Too loose and too tight data grouping might be the reason the Association Rule algorithm derives few rules with high confidence values. We hope the results can help learners and teachers realize the advantages, disadvantages and effects of using social media web application for education.
