How would you implement micro rewards for wrong steps in a classroom? The idea would be to either have students get savvy enough that they didn’t take the reward automatically or else learn from the experience of receiving a reward that was short-term good and long-term bad. The easy answer is candy. Like, “I’ll give you a piece of candy if you eliminate 13x instead of 2x” when 2x is the right thing to do. The extension and real lesson here is that sometimes people are trying to get you to do something that’s better for them than you by offering you a short term reward. This article is really about game design, but it has instructional implications.
It’s easy to see the implications of rewarding correct steps. This is shaping. But even that is questionable according to the paper in the linked article. The paper concludes that when someone is rewarded for correct steps, their learning is not transferable to novel problems, only to identical or very similar problems.