Scaling Peer and Self Assessment
This spring, we collaborated with Coursera to launch the first massive-scale class with self and peer assessment. We leveraged the self-assessment materials (like grading rubrics) that I’ve created and refined for my intro HCI class over the past several years. For peer assessment, we built on a technique called calibrated peer assessment where students learn grading rubrics through training examples and then grade peers. In our first experimental class, http://hci-class.org, students anonymously graded 5 of their peers’ assignments -- and then their own. My experience has been that students get a lot out of assessing their own work. And while there’s a bit of gaming the system (which often usefully attune students), there’s a lot of earnest and frank reflection and assessment that I think it tremendously valuable -- and increases students maturity. Students also do remarkably well.