Not all limits are the same

Are your students motivated by opportunity or fearful of scarcity?

Photo by Daria Kurennaya on Unsplash

We often write about the importance of limits in alternative grading (a few examples here). If you allow unlimited reassessments or infinite due date extensions, that’s likely to cause a lot of trouble. Infinitely flexible policies usually don’t provide the structure and focus that’s needed for intentional learning… and you’ll be overwhelmed by late work and piles of grading! So putting reasonable limits on your alternative grading system is critical for its success.¹

Today’s post is inspired by a recent discussion about the flip side of this issue. An instructor had limited reassessments by using a token system. Rather than using the tokens, students kept hoarding them, fearful that they would need them later. This made each new assessment feel high stakes as students felt compelled to succeed so that they didn’t have to spend a token.

Even worse, after a while, some students who had been doing this found themselves needing more reassessments than tokens — and eventually, they were mathematically eliminated from earning an A. As you can imagine, this made for a lot of stress among students and unhappiness with the grading system.

The conversation turned on this key problem: Whenever there is a limit, students can run up against that limit. Eventually, students can end up in the sort of hole described here, stopped from demonstrating their learning due to artificial limits. The person I was talking with took this as a fundamental flaw with alternative grading.²

I disagree with that conclusion, but it helped me formulate an important principle: Not all limits are the same. Let’s dig into that and see how different kinds of limits can avoid this type of problem.

Different limits, different incentives

While all limits are limits, not all limits have the same effect on how students act. Some limits encourage a scarcity mindset that encourages holding back and being conservative with resources, while others focus on abundance and encourage taking advantage of those same resources. Students respond differently to each.

For example, consider tokens. These limit the total number of times a student is allowed to do something: If they start the semester with 7 tokens, and they have to spend a token for each reassessment, then they’re fundamentally limited to 7 reassessments per semester. That leads to a sense of artificial scarcity. Students are naturally incentivized to avoid reassessing unnecessarily, and to really make it count when they use a token.

Taken to an extreme, this ends up encouraging harmful behavior. Students avoid reassessing when they really should reassess, out of fear for their future needs. Students can dig themselves into a hole that they could have avoided by… spending tokens. This is what was happening in the situation I described above.

This certainly isn’t the intent, and the artificial scarcity of tokens can be a positive feature that encourages good judgment and taking reassessments seriously. But some students, especially those who are highly risk-averse (perhaps with an avoidance orientation), react poorly to this kind of incentive. The result is that those students feel even more stressed.

Here’s another example of a limit with meaningfully different outcomes. I often alternate new assignments with “revision weeks”, in which students can submit a revision of any one previous graded item. This is still a limit: students can only reassess one item in a revision week, and revision weeks only occur about 7 times per semester – the same as the token example above. But this approach creates a sense of abundance or opportunity. Students know that if they don’t use a revision week, there will be another one in the future. But if they don’t use a revision week, that opportunity expires and they’ve gained nothing by it – so why not use it?

Rather than “I have a scarce resource I must hoard”, this encourages a different kind of thinking: “I have an opportunity to do something that will pass by if I don’t use it.” At the same time, the limits – only one revision every other week – keep workload reasonable and focused both for me and for my students.

Another variation is regularly scheduled new attempts, for example via weekly quizzes. You can make a schedule showing which standards will be assessed each week and let students know this ahead of time. Much like revision weeks, students don’t have to attempt new questions for standards they still need. But if they don’t, that opportunity has passed by. So why not put in the time to make the attempt worthwhile?

There isn’t one right way to do it

I’m not trying to say that tokens are bad and regularly scheduled new attempts are good. Rather, different approaches work for different people, in different contexts, and it’s important to think about which approach makes sense in your context.

In my experience, artificially scarce tokens don't work well when students are used to earning high grades, and demonstrate high anxiety about those grades. That can happen, for example, at “elite” institutions, in pre-med classes, and any other situations that involve highly grade-motivated students. If you’re in that situation, perhaps an abundance approach will work better.³

On the other hand, an abundance approach creates less pressure, which can cause its own problems. Some students might react by ignoring those opportunities or not taking them seriously, which itself can lead to a sudden panic near the end of the semester when few opportunities remain. If this is your case, perhaps the artificial scarcity of tokens makes more sense.

There are lots of variations on these approaches. Here are some examples:

• Reduce artificial scarcity by giving students opportunities to earn additional tokens. Make sure these opportunities are relevant to the class and not busywork.

• If you’re using time-based limits, their frequency can be capped in a variety of ways depending on your needs: Once per week, once every two weeks, once per unit, once per project, etc.

If you don’t like the idea of truly artificial limits (e.g. number or frequency of reassessments), there are also ways to limit reassessments in more natural ways. For example:

• Insist that students fully complete one reassessment to your satisfaction before starting another. This makes the most sense for reassessments that are fairly large (such as revising an entire report, project, or essay). You’ll need to be ready to help students understand when their reassessment is completed successfully and they can move on to another item.

• Have students “unlock” the right to reassess by submitting relevant practice problems. This has the natural effect of encouraging practice where it’s needed most, while also slowing down the pace of reassessments. If a student struggles with the practice, that can become a topic for an office hour discussion.

• Require students to focus on more urgent reassessments before less-urgent ones. For example, if you use a rubric where multiple levels are “good enough” (e.g. Satisfactory and Excellent), insist that students reassess items that are below Satisfactory before they can attempt to improve anything to Excellent.

No matter what, I highly recommend including at least one hard limit on the rate of reassessments, such as one per day. In addition, none of this applies to extreme circumstances — illness, a death in the family — or to required accommodations. In all of those circumstances, you can and should find a solution that works for the student without worrying about other limits.

Each instructor’s choices depend on their and their students’ situations. Think carefully and even experiment with different approaches to see what works best for you!

1. I still use an unlimited due date extension policy in some classes. I always monitor it for students making poor decisions and intervene early — see the end of the post for some more suggestions about how to manage unlimited policies.

2. This problem with running up against limits is baked right into traditional grading, just not in the same ways. When we permanently average a student’s work into their final grade – the opposite of reassessment without penalty – then that artificially limits their final grade by ignoring evidence of eventual learning. We’ve all seen students who are “mathematically eliminated” from earning an A, a B, etc., just by doing poorly on one traditionally graded exam, even if they later learn those concepts thoroughly.

3. “Elite” institutions are also one of the places where I’ve heard about alternative grading leading to grade deflation. That is, if an institution has seen true grade inflation – in which grades increase without representing better learning, typically due to student, family, and/or donor pressures — then alternative grading, by concretely linking grades to demonstrated understanding, can sometimes decrease average grades.