# Grace days and dropped scores: policy text

Version 1.0, August 2026. Lift these into your syllabus and change the numbers.

## Grace days

> **Grace days.** Each student receives 3 grace days that may be applied to any non-exam assignment without penalty and without explanation. Grace days cannot be applied to exams or to the final submission. To use one, submit late and note in your submission comment that you are using a grace day. You do not need to email me in advance and you do not need to tell me why.
>
> After grace days are exhausted, 2% is deducted for each 24 hours past the due date and time. For exams, contact me to reschedule within 15 days. Please tell me directly if you hit an unavoidable conflict.

Four design choices are doing the work, and changing any of them changes what the policy is.

**Bounded.** Three, not unlimited. The number is small enough that students spend them on things that matter and large enough to cover an ordinary bad week.

**Automatic.** No approval step. The moment you have to say yes or no, you are back to adjudicating, which is the thing the policy exists to stop.

**No explanation required.** This is the part people cut, and it is the part that matters most. Requiring a reason means the students who are most fluent at describing hardship to authority get the most accommodation, and the students who will not discuss their circumstances with a professor get none.

**Excluded where it would break something.** Exams and the final deliverable are carved out, because a grace day on those has knock-on effects for grading schedules and for the student's own sequence of work.

## Dropped scores

> **Technical difficulties accommodation.** The lowest 2 lab grades and the lowest quiz grade are dropped. This is automatic and requires no request.

This solves a different problem than grace days. Grace days handle the week that went wrong. Dropped scores handle the single session where the software failed, the laptop died, or the student was in the room but not functional. In a course where in-class work is graded, this is the difference between a policy and a queue of individual appeals.

## Why the policy, stated to students

> **Why these policies.** Rigid deadlines fall unevenly. A student managing a health issue, a family obligation, or a second job absorbs a much larger penalty from the same missed deadline than a student who is not. These policies are bounded so that they do not become an alternative schedule, and automatic so that nobody has to make a case to me about their personal circumstances in order to receive them.

## What you can and cannot claim for this policy

Do not tell students, as an earlier version of my own syllabus did, that research demonstrates flexible policies do not reduce accountability. No randomized trial of this exists.

What does exist is three observational studies. Extension use turns out not to be patterned by gender, race, first-generation status, or employment, and the students with the highest GPAs use extensions most (Ganguly et al. 2025). In a course with 78 percent uptake there was no final-grade difference between users and non-users (Ruesch and Sarvary 2024). And students consume whatever window you give them regardless of assignment type (Conner et al. 2024), which is the finding that argues for bounding the policy rather than opening it.

That refutes the worry that flexibility gets exploited by the students who can least afford it. It does not show the policy helps anyone learn more, and you should not say that it does.

The strongest argument remains the design one. The alternative requires you to evaluate the legitimacy of students' private circumstances, you are not qualified to do that, and the students who fare worst under such a system are not the ones gaming it.

## Cost

Five minutes to paste in. Then roughly zero. The observable change is in your inbox rather than your gradebook: the negotiation thread that used to run all semester stops, because the answer was published in week one.