PIs: Carla van de Sande, Robert Atkinson
Arizona State University

Online homework help forums exist, but there is no good way, right now, for students to know whether they are getting good advice or not. The innovations here are to take advantage of the fact that many students participating in online homework forums want to give each other good advice and almost all want to do well on their homework. Taking advantage of participants’ want to give good advice, these PIs propose to provide scaffolding that will encourage and aid good explanations of the how-tos and whys of worked-out-problem examples students are sharing. Taking advantage of participants’ want to get their own homework right, the PIs propose a simple approach to helping them learn to judge the quality of help they receive. The innovation is quite simple; it includes providing users of a homework site a template for sharing answers and a rating rubric that has them reflect on what made a worked-out answer to a question useful to them. The PIs are carrying out a set of experimental studies in a quasi-authentic homework forum to address several issues: How can we help students judge the quality of solutions that are shared? How can we help sharers present better quality worked examples? There is learning to be had for both kinds of participants. The work builds on research on what makes worked-out examples helpful and on what is known about the value of self-explanations.

One way middle school and high school students get help with homework is through online homework forums. The most popular current online homework forums hold repositories of worked-out examples that those needing homework help can use as analogues as they work on their own homework problems. But students who go to these sites for help cannot know if the worked examples they are looking at are done well or not, and simply solving one’s own homework problem the way someone else solved an analogous problem does not guarantee learning. This project is developing a way to help students who seek help from these forums to know whether a worked example is done well and to promote the best learning possible from worked examples. The worked-example literature reports that structuring worked examples according to the steps in working them out, the goals being addressed in each step, and the means of addressing each goal promotes learning better than structuring worked examples in other ways or note structuring them at all. This project, therefore, is developing a way to encourage and help those who are entering worked examples to structure their examples so that those seeking to learn will get get the most from their answers. The self-explanation literature suggests that such structuring will also allow those offering worked examples to deepen their understanding and capabilities as they work towards presenting their worked examples to others in an organized way. This project is developing a means of helping those who offer worked examples to structure them for promoting learning and a means of helping those who are using those worked examples to know if a worked example is likely to be trustworthy and learn well from the worked example, and it is investigating the benefits to both those offering examples and those using them of using a structured approach. Eventually, such help could be incorporated into the real-life help forums themselves.