In Fall 2024, I experimented with a new lecture attendance policy in CS 3110, which I co-instructed with Anshuman Mohan. Attendance increased compared to recent semesters and student satisfaction was high. The key idea is to offer each student their own choice of whether attendance is graded as optional or mandatory.

The Inspiration

After an unexpected course-policy disaster in Spring 2024 in which attendance plummeted, I read around to find ideas on how to improve attendance in college courses. I found an article by Gerald and Brady titled “Time to Make Your Mandatory-Attendance Policy Optional?” in the Chronicle of Higher Education (January 13, 2019). They write,

Multiple surveys have shown that undergraduates tend to dislike a mandatory-attendance policy but admit they would miss class more frequently without one. Similarly, most students think that regularly attending class improves their grades but also believe that it is ultimately their decision whether or not to attend.

To address that dichotomy, they propose offering each student a choice of an optional or a mandatory attendance policy at the beginning of the semester. Under the optional policy, a student would receive a little extra credit in their final grade if they attended consistently. Under the mandatory policy, a student would receive even more extra credit in their final grade but instead would be penalized (i.e., the “extra credit” would be negative) if they missed too many classes. Students are not allowed to change their choice later in the semester.

Gerald and Brady reported positive experiences with that attendance model over five semesters of instruction in a large, required, upper-level course for public-health majors. It seemed worthwhile to try it in our large, required, upper-level course for CS majors.

How We Instantiated It in CS 3110 Fall 2024

We gave students this choice:

• Optional Attendance: Attendance is nice but optional. A student’s lecture attendance would result in a bonus of 0 to 2 points to their final grade, which was out of 100 points. This was a typical policy I had used in the course since Fall 2021.
• Mandatory Attendance: Attendance is a value to which you commit. A student’s lecture attendance would result in a bonus (or possibly penalty) of –1 to 4 points to their final grade.

Students were required to make their choice by the end of the third week of classes. After that, they could not change their choice. Students who failed to register a choice were assigned to Optional Attendance.

We stated up front that every student would have three absences dropped. We chose that number because it was equivalent to about 10% of the lectures in the semester. The New York State Education Department defines “chronic absenteeism” as missing 10% of school days for any reason.

iClicker hardware remotes were used to measure attendance. We did not permit the software app, as there had been indications of large-scale cheating with it in Spring 2024.

We did not determine in advance how attendance scores would be mapped to bonuses. At the end of the semester we settled on the following scheme:

• With Optional Attendance, every undropped absence resulted in a deduction of 0.1 points from the max bonus of 2 points, until the floor of zero points was reached.
• With Mandatory Attendance, the deduction was 0.5 points, it started at a max bonus of 4 points, and bottomed out at –1 point.

As a consequence, students who committed to Mandatory Attendance but missed too many lectures (which given the numbers above turns out to be three weeks’ worth) received lower attendance scores than students who chose Optional Attendance and attended the same number of lectures.

Outcomes

Attendance was high. At the end of the semester there were 318 students enrolled in the course. Of those, about 86% chose Mandatory Attendance, and the mean attendance for students in that group was about 90% — versus about 55% for students who chose Optional Attendance. Compared to recent semesters, Fall 2024 attendance was higher:

Percentages are calculated based on final enrollment in the course, which is why some early data points are greater than 100%. The increased participation in Fall 2024 compared to Spring 2024 is statistically significant (p < 0.05) but compared to Fall 2022 is not (p = 0.07). Spring 2024 used a fully mandatory attendance policy in which attendance was 2 points out of 100 in the final grade. Fall 2022 used Fall 2024’s Optional Attendance policy for everyone.

Choosing Optional Attendance is associated with lower grades. Of the students who chose Optional Attendance, about 27% got a final grade of C+ or lower. But of the students who chose Mandatory Attendance, only about 9% got a final grade of C+ or lower. Moreover, students who chose Optional Attendance performed lower on exams than students who chose Mandatory Attendance; that result is statistically significant (p < 0.05) but the effect size is small (4 points out of 100 on average). Percent attendance does not seem to be an accurate predictor of performance in the course — at least not in a linear regression model — but there is a moderately positive correlation (ρ = 0.4) between attendance and final grades.

Students were satisfied. In a post-semester anonymous survey (response rate 57%), 94% of students said that they liked having a choice of attendance options, 89% said that they are satisfied with the choice they made, 14% suggested eliminating the choice and making attendance optional for everyone, and 17% suggested eliminating the choice and make attendance mandatory for everyone.

We observed what we were led to expect. Our quantitative findings largely replicate those of Gerald and Brady. The biggest difference is that they found attendance among the Optional Attendance group was 71%, whereas ours was merely 55%. They took attendance by having students sign in with their TA in lecture, whereas we used iClicker. Could that additional human interaction be a nudge toward attendance?

Conclusions

In the last three semesters of teaching CS 3110 I’ve tried an optional-with-bonus lecture attendance policy (Fall 2022), a purely mandatory policy (Spring 2024), and the student-choice policy described above (Fall 2024). The student-choice policy was the most pleasant from my perspective. (Other than a little bookkeeping.) It gave students agency in how they wanted to allocate their time while also incentivizing them to commit to attending lecture.

I have no reason to recommend making attendance purely mandatory. The act of choosing Mandatory Attendance does not necessarily cause a difference in scores or grades. Rather, choosing Mandatory Attendance is merely associated with having better scores and grades. The underlying cause could be the attitude of the student, their academic maturity, external factors in their life, etc. On the other hand, that association could be the basis for different intervention strategies — for example, students who choose Optional Attendance and have a low first exam score perhaps would benefit from a more robust alert to their academic advising team.

Coda: A Short Story

One day when my co-instructor was lecturing I took the opportunity to come in a little bit late and sit in the back row. I could see the laptop and phone screens of students around me, most of whom did not realize I was there. Few were using their devices in a way that was consistent with paying attention to the lecture. Most students were multitasking: doing work in other courses, applying to jobs, shopping online, and so forth.

Partway through lecture my co-instructor asked a question and there was appropriate reason for me to call out an answer. That alerted students as to where I was in the classroom, especially those nearby me — who looked at me in some surprise. But as the lecture resumed, none of those students changed their behavior. They knew I was there, and presumably they were aware their screens were visible, but apparently they felt comfortable multitasking while the lecture was proceeding and their instructor could view their activities.

Is this a problem?