A Discussion of “Class Meeting Frequency, Start Times, and Academic Performance”
Anyone who went to college…remember that feeling when the only class that was open was an 8 AM?
As a night owl, I dreaded it. And usually avoided it.
If you are a morning person, maybe you do not know what I am talking about. But you would appear to be in the minority as Cotti, Gordanier, and Ozturk (2018) found, among other things, that grades increase as start time is later.
They were also curious if the weekly schedule made a difference. Typically, colleges schedule classes in the US follow this schedule:
- 3 times a week (MWF) 50 minutes per class
- 2 times a week (MW or TR) 75 minutes per class
- 1 time a week (night or day) 150 minutes per class.
HBU, the college I teach at, follows this convention. If you notice, each schedule results in 150 minutes of class time per week. The only thing that changes is how many class days it is spread across.
The authors mention the student government in the University of Texas system introduced a resolution to increase the number of MW classes and decrease the number of MWF offerings. Such a scheduling change would make it more likely a student could group their classes into MW and TR classes, leaving Fridays open.
Reasons for such a scheduling change include (p. 12–13)
- Fridays are open for jobs for both internship experience as well as earning income to pay for college expenses and hopefully accumulate less debt
- Likewise professors would enjoy a block of time to focus on research and university committee obligations
- Greater scheduling flexibility could increase graduation rates
- Colleges could save money on energy and staffing costs by closing buildings an extra day
- Lowering the number of days of commuting and number of cars on the road
Reasons against such a scheduling change include (p. 12)
- Administrators face space constraints because the shorter 3 day a week classes means there are more available blocks available to schedule. (You could fit in 9 class times MWF but only 6 on MW between 8:00–5:00).
- Political pressures to not look like space is not being used or that staff and faculty are not working on Fridays
- The authors even mention a phenomena called “Thirst-day” where students start the weekend early and go drinking on Thursday nights. Some think Friday classes lower the number of students participating in Thirst-day.
Quite the debate then!
One thing seems to be missing — in the list of pros and cons where is the mention of how it impacts the students’ education?
The essential question then is, would fewer days a week in class lessen the learning experience, as measured by grades? That is, would we be short-changing students by meeting less often?
The Data
The authors gather data from “a large public university in the Midwest” for six semesters from 2013–2015 of 17,244 unique students with grades being recorded on a 4.0 score, and the data they collect allows them to control for student, professor, and class specific factors. (p. 13)
If you are concerned about the validity of using grades as a measure for the outcome, I wrote an earlier article exploring GPA as a metric. The authors of that paper found it a noisy, but useful, metric of student performance.
GPA: What Does This Mysterious Metric Teach Us?
Results
They find that if you control for instructor, student and course fixed effects, (p. 14)
- There is no statistically significant difference in performance from 2-day a week and 3-day a week classes
- There is a small, negative effect associated with 1-day a week classes
- Later start times are associated with higher grades (0.01 higher grade point per hour)
- And class size is not associated with outcomes
One issue they wanted to address in this paper is instructor sorting. Professors are not usually randomly assigned a schedule. Those who are full time, as opposed to adjuncts, are often given a choice of the available sections. So this could open up a few issues that would confuse the results.
More experienced professors may prefer the 2 day a week schedule and their students could perform better because of the teachers’ experience. Or perform worse because they grade harder. Or, adjuncts may end up with the 3 day a week schedule and as a group, other studies have shown adjuncts tend to give higher grades.
The point is, since the professors are not randomly assigned to the different class schedules, there could be some bias in the results.
That was why they wanted to do this analysis where they can account for instructor sorting and see if there is still any difference between 2-day a week classes and 3-day a week classes.
It would seem that since they found no difference between the 2 options, while previous studies had, any difference is due more to instructor sorting than the number of days spent in class.
Thus, the authors worry that instructor sorting could be an issue that likely hits students who have less scheduling flexibility because they need to work, if the stronger professors are selecting times that more often conflict with student work schedules.
Another source of bias in the data could come from student sorting. That is, students choosing a particular schedule based on what they know works best for them. Essentially, “students with more flexibility could be positively sorting into certain schedules, which could bias estimates.” (p. 14)
To account for all of this, the authors note that unlike many previous studies, their data allows them to control for these professor, student, and class factors.
They also run the data a second time limiting the observations to courses where there was only one time offered so there could be no student sorting. All coefficients were of the same sign and similar in magnitude, thus it does not seem student sorting is a significant issue.
They do find evidence, however, of instructor sorting away from 3 day a week classes. (p. 15) They suggest that previous studies that did not account for instructor specific differences and found that 3 day a week classes resulted in higher grades may instead have been picking up issues around instructor sorting.
And they issue a final note of concern about the lower income students who have to work and who have less flexibility in selecting class times may be unable to access the best professors due to instructor sorting.
With that in mind, I think they are implying that a MW or TR set of offerings would likely reduce this impact. Students have all day Friday to work, and professors seeking 2 day a week schedules would be more open to MW than MWF.
And as their results show, fewer class days do not result in lower student outcomes.
References:
Cotti, Chad, John Gordanier, and Orgul Ozturk (2018). “Class Meeting Frequency, Start Times, and Academic Performance.” Economics of Education Review, 62: 12–15
By Ellen Clardy, PhD on .
Exported from Medium on December 15, 2022.
