In yesterday's post on the U of T professor that was prevented of using student assessment methods in his course, I took the position that unions should not be allowed to dictate how an instructor teaches their course. I still believe this to be true, but wanted to elaborate a bit on this specific position.

Steve Joordens, the professor involved in this case, teaches an intro psychology course with 1400 students. That is problem number one, and one that is faced by almost every university. That number of students would overwhelms any conventional teaching scenario. Even setting an acceptable class size at 50 students would require 28 sections, with 28 professors teaching. While that may appear at first to be a solution to the PhD employment problem, the resources and logistics for putting on 28 separate sections (vs. one or a few mega-auditorium classes) is hard. And you would need to pay those 28 professors, who even at typical sessional rates would cost around $140,000 per semester.

That's only about 100$/student, though, so it wouldn't be impossible to do this. If tuition rates could be increased to cover the additional costs of teaching, facilities, and operations. Which they can't, at least not at most universities. So we're stuck with the basic parameters of the problem: one professor teaching a class of 1400 students.

The typical solution involves trying to get 1400 to view the lectures, then using machine-readable assessment methods. Pulling even that off is a lot of work, but it still leaves a lot of be desired as far as student learning goes.

What complicates the matter somewhat is Joordens' motivation for introducing the PeerScholar software. In this blog comment in 2006, he states that his motivation behind the PeerScholar system was pedagogical:

While I did indeed – perhaps stupidly it now seems – point out some of the economic value of the peer-to-peer approach, it was not instituted for economic value at all. It was instituted because it provides a method of teaching thinking and communication skills that is generally regarded as superior, yup superior, to TA grading. That is, by applying critical analysis skills (call that marking if you like) to written pieces that vary from poorly to well written, students gain extremely valuable skills that they can then apply to their own work. Thus, the grading component of the assignments may be where the true learning is occurring.

The "stupid" comment he refers to is the one cited in the courts' judgment, apparently taken from Joordens' web site:

I will be completely honest. The original reason for seriously considering a peer-to-peer evaluation process was financial. We cannot afford to pay a large team of TAs to mark written answers for large classes. Moreover, it would take them so long to do the marking that it also just wouldn't be practical. Peer-to-peer evaluation, when combined with great internet programming, is fast and cheap.

In other words, Joordens wanted to improve the student learning by introducing assessment that was not based on multiple-choice answers. However, the financial restrictions of the situation prevent him from hiring TAs to correct written assignments, so he implements a peer-review solution which provides the same or better student outcomes.

The union, predictably, is concerned only with the impact this has on their members: the loss of jobs of teaching assistants. Note that in this specific case there is no loss of jobs because there was no budget anyway to hire teaching assistants to handle this marking. Joordens wasn't using the peer review system to cut costs, he was trying to improve student outcomes without incurring additional costs.

Still, the union has to fight this on principle: they cannot allow this kind of solution to take root on campus because it has the potential to reduce the need for teaching assistants. This is a real risk, especially with universities looking for ways to reduce costs, and especially if it can be positioned as improving student outcomes.

It doesn't have to be that way, of course. Educational technologies and teaching assistants can co-exist in ways that lead to improved student experiences and outcomes. Students can become more engaged in their learning, teaching assistants, freed of some of the burden of correcting, can focus on interacting with students and teaching.

Now that the union and the Ontario courts have established a precedent that attempts to reduce dependencies on TAs will challenged and most likely prevented, instructors will be wary of making changes to how they deliver and manage their courses. But I hope that people will still try to innovate, to find ways to improve student learning in more efficient and effective ways. The system as it is not sustainable, and we are going to need creative solutions in the coming years if we are to continue in our role as establishments of higher learning. Peer-review assessment methods are not the biggest threat faced by teaching assistants, and it is time that they and their unions faced that fact.