How to teach college #20: Sabbatical tweaks

(This originally appeared here in my Substack newsletter. It occurred to me a lot more of my fellow educators might see it here.)

I always keep a file on my Google Drive in which I jot down ideas for the upcoming term; just like anybody else, if I don’t write good ideas down, I lose them. Here are the two changes I’m toying with for Fall 2023 back at Bushnell.

First, it’s time for me to think about how my writing assignment sequence works in a ChatGPT world.

(Parenthetically I will note that I wrote this before I read?this article. There are some promising possibilities in it, but I need to think them over.)

In case you don’t know, students in my 200-level classes typically choose a peer-reviewed study from a collection I offer, read it, and write a book report on it. Students in 300-level classes do a literature review in which they follow a single concept or premise through the findings of?multiple?studies, typically three, and explain how the concept/premise is refined or tweaked or clarified. Students in 400-level classes do a work of original research, which usually means collecting and interpreting data, or, in the rhetorical criticism class, carrying out an original criticism on a rhetorical artifact. I grade their?mechanics?(spelling, grammar and punctuation) with a routine I nicknamed the “Triple Jump,” which involves multiple rounds of submission, feedback and revision, along with the risk of a zero on the entire paper if I find a third mechanical error in the final product.

As of right now the only things I know from ChatGPT are what I’ve read. (Although let me pause to say that the next reporter who uses the “Did you like the last paragraph? ChatGPT wrote it!” is getting slapped.) Nevertheless, I need to decide what to do if an online gizmo starts offering students entire paragraphs that are well edited and plausibly coherent in exchange for a carefully worded prompt and a click.

I’m thinking the projects might just become short teaching presentations. I would give students a quick primer on ways to break research findings down and present them to a non-expert audience. Then, I would supply a grading rubric, and invite them to my office to make a presentation to me. For article reports, I might give them ten minutes; for lit reviews, twenty; for research projects, thirty. The wrinkle is that their allotted speaking time would be shorter: for the article report, five minutes, for the lit review, ten, and for the research project, fifteen. The rest would be a question-and-answer session in which I would test their understanding of the work. I would gladly let them come do dress rehearsals in advance, and I would post videos of really good examples just as soon as I collected a few.

What bothers me about this replacement is that I would be guilty of not shouldering my share of the burden to help them learn to write. The trick is to figure out a course-relevant exercise that strengthens that skill, but that ChatGPT can’t spit out. In-class writing is a possibility, but what they need is to learn to?edit.?It’s a problem I haven’t yet figured out how to tackle.

Second, I think I might begin overhauling my upper division classes, starting with just one in Fall 23 and eventually engineering the change across the entire major. Here’s what I’m thinking: one class, and the first might be nonverbal communication, would go from meeting three hours each week to just two. I would want those to be fifty-minute meetings on two days, because I don’t want a six-day break between instructional gatherings; I’m seeing for myself over here why that’s a problem.

The third hour would be a weekly one-on-one meeting in my office to work on a concrete work product, ideally a skill. We would talk for the second and third weeks of the term about their options, about where they see themselves after graduation and what applications of the class they would find most useful in that setting. We would then define the skill and a project they were to complete by the end of week fourteen. I would set up a time and place on ACE Day for them to demonstrate their skill and report on their project.

What kinds of skills? Things like assertiveness, Itzchakovian high-quality listening, conflict coaching, converting technical information into audience-friendly narrative, etc. And in particular I want to keep an eye out for interpersonal skills that are optimized for AI-partnering workplace roles. More and more of what I’m reading says what AI will?really?do to the job market is create a lot of human-AI tandem positions, and I’ve been reading for years that interpersonal communication is the one human activity that AI designers are pretty sure they can’t swipe. ChatGPT might be a very impressive stochastic parrot, and one that might fool me when I’m grading papers, but its ability to recognize someone’s emotional state and craft a person-centered support message is very clunky and ineffective and likely to remain so.

If I do manage to generalize this model across the whole program, then graduating seniors would have a portfolio of skills and work products to show to a prospective employer. I would also want them to use their skill toolbox to guide their choice of internship, and bring me back a finished project from it, and I would encourage them to design a capstone project that involved using one or more of their skills to accomplish something that was a degree of challenge above their class projects.