I'm now in my third semester teaching my flipped social psychology course. Everything is going well -- I'm now to the pleasant stage of not needing to revise my materials much -- but I did make some changes. I led a Professional Learning Community last spring about metacognition (you can see some info about that here) and one idea I got from that literature is the importance of retrieval in learning. I admit that I used to think of tests as mostly assessments of learning. But research in psychology (where most of the ideas about metacognition come from) clearly demonstrates that learning is strengthened by retrieval. So, this means that tests can (should) be thought of as formative tools that enhance learning. This was the motivation for having so many smaller quizzes that are lower stakes in my flipped course. But if taking the quiz once is good, perhaps taking it a second time would be better. So this is why I changed my class structure.
I used to have the students engage in the jigsaw, take the quiz, and then we'd go over the quiz right away. Now we do the jigsaw and take the quiz, but we don't go over it right away. Instead, the next day they come in and take the exact same quiz a second time, using the same quiz itself with all their notes and comments on it. This means that between taking the quiz the first and second time they can go back to the text or recorded lectures to look up the answers to the questions on the quiz. Or they can talk to other students about them. Either activity would be beneficial in terms of learning the material, I believe. The idea is that thinking of the answers to the questions twice is twice as much retrieval.
Correct answers are worth one point on the first version of the quiz, but only half a point if they get it wrong first and right the second time. So it's in their interest to do as well as possible the first time.
This has lead to higher final grades on the quizzes. Scores on the first six quizzes are on average 10.6% higher than in the previous two semesters (combined). That's a letter grade higher -- not insignificant. So the question is does that reflect a commensurate increase in learning. I don't have data that speaks to that definitively. And I do think the students have pretty quickly figured out a strategy for the second quiz: if they were trying to deiced between two options, pick the best one the first time and the second best the second time. Fair enough, and clever. But they still get more points for the first choice if they get it right, and they still have to narrow it down to two options from four. I wish I had good data about their learning that was not contaminated by this strategy, but I just don't. I just have to trust the research on retrieval, I suppose.
This has introduced some logistical complexity. Each day starts with taking the previous day's quiz a second time, then we go over the quiz, then do the jigsaw on that day's material, and finally take the quiz on that material. So twice as many scantron forms used and twice as much time set aside for quizzes. And the quiz grading is more complex because I have to compare both versions to give credit for getting it right the second time. (A dependable graduate assistant is worth her weight in gold!) Overall I think the increase in learning as result of added retrieval is worth the added complexity.
My courses are always little laboratories. I'll see how this turns out.
