By Kevin Cheung, Associate Professor, School of Mathematics and Statistics
As I start to become more familiar with cuLearn, I realize that there is a lot of course data that can be useful for improving the learning environment for my students.
Without a doubt, the ability to perform complex queries in cuLearn is limited. For example, obtaining a list of students who made no more than two quiz attempts within a specific time period simply cannot be done without external scripts at the moment. However, I have found the online quiz statistics per quiz question to be tremendously helpful in identifying potential weaknesses of my students or deficiencies of the course content.
In the past two years, I have encountered a few surprises from online quiz results. What I thought were straightforward questions turned out to be challenging for a substantial number of students. For my online course Linear Algebra I (MATH 1107), I have on two occasions added new sections to the lecture notes as a result of what I saw in the quiz statistics. Since then, questions from students on these topics have reduced dramatically.
I have also discovered that requiring students to submit their written assignments electronically has been beneficial in many ways. For instance, the risk of losing assignment submissions is practically nonexistent. In addition, not only do I have a record of the breakdown of marks for each submission, I also have a record of the actual work along with any comments from the markers. I have at times gone back to assignments from previous terms to see the types of mistakes students made and adjusted my teaching accordingly.
Now that big data in education seems to be in fashion, it may be prudent to take a step back and see if such a heavy-handed approach is really necessary. My experience suggests that just looking at course data is already quite helpful.