Professors, particularly at research universities, wear many hats. On the one hand, they are instructors, entrusted to pass on knowledge to their students. On the other hand, they are researchers and are expected to add to their field’s body of knowledge. As a way to help professors balance their two roles, “discussion sections” have become a common teaching model in academia.
In a discussion section, a professor shares teaching responsibilities with several “teaching assistants” (usually graduate students). The professor is still the main instructor and delivers his lecture to a large classroom of students. However, there is usually little time for student questions and discussion.
For that reason, the teaching assistants are assigned to lead small group discussions after the lecture to discuss the course material in-depth and, depending on the course, do practice exercises. A lot of the other “grunt work” that comes with teaching, such as grading and providing detailed feedback on assignments, is also often relegated to the teaching assistants.
In the United States, these small discussion sections, sometimes called “tutorials” or “recitations,” are typically led by PhD students. In other countries, however, it is very common for master’s students—and even undergraduates—to lead discussions and assign and grade assignments.
Of course, student instructors are far less credentialed than professors which raises the question: Are students’ learning outcomes hurt by having a student instructor?
An academic paper entitled Students are Almost as Effective as Professors in University Teaching by Jan Feld, Nicolás Salamanca, and Ulf Zölitz sheds light on the question. In the paper, the researchers studied professors and student instructors who both taught tutorials at a Dutch business school.
The Martin Center spoke with Jan Feld of Victoria University of Wellington and the Institute of Labor Economics to discuss the researchers’ findings. The following interview has been edited for clarity and length.
What motivated your research?
[We observed that] students and professors are doing the same job: They both teach tutorials. Given that professors are so much more experienced and qualified than student instructors, we wondered whether that translates into better outcomes for students. We found that the answer is: not really.
When you refer to “student instructors,” what kinds of students fall into that category? Bachelor’s students? Master’s students? What group do PhD students fall into—the student-instructor category or the professor category?
Student instructors are bachelor’s and master’s students. PhD students are separate; we don’t look at them. We compare bachelor’s and master’s students with “senior instructors”—“professors” is a good shorthand for that. Technically, senior instructors are lecturers, post-docs, and professors of all ranks—assistant, associate, and full professors.
In the United States, the majority of people who teach tutorials are PhD students.
Yes, in a previous paper, we ran a survey and I think, even in the United States, you have some undergraduate students teaching tutorials. But the PhD model is much more prevalent.
How did you measure the teaching effectiveness of students in comparison with the teaching effectiveness of professors?
Essentially, we compared the outcomes of students who were taught by professors with the outcomes of students who were taught by student instructors. Students were randomly assigned to their instructors, so any other differences should wash out. We mainly focused on grades, future grades, and course evaluations. And there are some data on students’ labor market outcomes.
What were your findings? Was there anything that surprised you?
We found a small negative effect of being taught by student instructors on students’ grades. But it’s not statistically significant, so that might be a chance finding. We did find a fairly small negative effect on students’ evaluations of the course quality. It seems that students evaluate courses better if they are taught by a professor. But the interesting thing is that poor course evaluations do not translate into any measurable student learning outcomes. It does not translate into any meaningful differences in students’ grades or students’ future grades.
Overall, student instructors are doing surprisingly well in these TA sessions or tutorials.
Did you look at job prospects too?
We did. Of course, you could imagine that professors, since they have more experience, are better at giving students an overview [of the job market] or tips or referring them to jobs. However, we didn’t find any evidence of that. Now, that’s with the big caveat that we have few observations, so there might be a small effect that we didn’t have the statistical power to find. But we didn’t see any differences in terms of students’ earnings, how long it took them to find a job, or job satisfaction.
Do you think your findings reflect poorly on professors’ teaching abilities? Is there something they can do to improve?
I’m also a professor in that category. My interpretation of these results is: should professors really be teaching tutorials? Tutorials are often designed to be simpler, more interactive; often students solve simple exercises or discuss the literature. It’s usually only 16 students and I see our value-added more in terms of lecturing, designing the course, and supervising PhD students. I think that’s where we really shine. I think that’s where our qualifications and experience make a difference.
But doing a job that a student can do as well I think often is not the best use of a professor’s time.
Your research suggests that student instructors don’t teach all courses equally well. Which ones do they teach well and which ones do they teach less effectively? What can account for these differences?
We found that student instructors do as well as professors in first-year courses and in mathematical courses, and less well in second plus-year courses and non-mathematical courses.
These differences are relatively small. But if we speculate what could be driving this, I suspect that with more-advanced courses and less-technical courses, having experience helps. Professors’ broader knowledge and ability to connect different themes or different strands of literature helps—which student instructors have just not come across. We have some evidence for that with our measure of course evaluations. We do see that student instructors are perceived as less able to transfer knowledge from one context to another.
So, you can imagine that if a class is more ideas-based or discussion-based instead of just solving mathematical problems, the professor might be more valuable. But, overall, these differences are fairly small.
Do you see any benefits of bachelor’s students teaching?
I think, depending on the courses, yes. You might have some first-year introductory courses—the courses where you need large numbers of instructors—and say you have some really clever third-year bachelor’s students. I suspect that, in most U.S. institutions, there are some clever bachelor’s students who can teach tutorials just as well.
Would there be some sort of screening process to make sure they’re qualified to teach?
At the Dutch business school [that we studied], and I suspect at other institutions as well, the screening process is looking at past grades. Another thing they look at is students’ command of English. Those are things you can observe fairly easily. Universities have records on past grades. Another good thing is: You can let go of bad student instructors much more easily.
If a PhD student or a professor teach badly, it’s difficult to get rid of them. We see it in our data: Student instructors who did poorly, in terms of their students’ grades, or in terms of their evaluations, didn’t get rehired.
In your research paper, you and your co-authors note that hiring a large quantity of student-instructors, or only student-instructors, to teach tutorials could negatively impact student-learning outcomes. Why might that be the case, given that they seem to teach almost as effectively as professors?
That’s because, as in any research, we can only study the student instructors that are currently teaching in the courses where you usually need student instructors. So, expanding the usage of student instructors might mean hiring additional student instructors, and the additional ones you hire might be of worse quality. Also, the additional courses that you might then ask the student instructors to teach might not be well-suited for them.
That’s why we expect that only using student instructors might have negative effects. Those effects might still be worth it because students are really cheap and professors’ time is really valuable.
Your data comes from a Dutch business school. How do you think your findings relate to institutions of higher education in other countries, including those in the United States?
In a previous paper, we conducted a survey to see how prevalent tutorial teaching is and we saw that many countries use tutorials. And many also use student instructors and professors at the same time. That’s the same kind of setup that we have in the Dutch business school that we studied.
So, we think that our results most speak to those institutions where professors and students teach alongside each other. But it is an empirical question whether in other universities they would find similar effects on student instructors. I hope other studies will follow up on this. I suspect that for these institutions, where professors and students teach alongside each other, using professors is probably not the best use of their time.
If student instructors do actually teach well, you would be hard-pressed to find a reason for using a professor—they’re doing the same job.
Shannon Watkins is senior writer at the James G. Martin Center for Academic Renewal.