Mixed messages (#1238)

Topics/tags: Grinnell, academia, rambly

Because Grinnell has small courses, generally capped at 24 students or so, we often offer multiple sections of the same course in the same semester. For example, in CS, we have three sections of CSC-151, Functional Problem Solving, in the fall, as well as two sections each of CSC-161, Imperative Problem Solving [1], CSC-207, Object-Oriented Problem Solving, Data Structures, and Algorithms, CSC-213, Operating Systems and Parallel Algorithms, and CSC-301, Algorithm Analysis [2]. In Biology, there are four sections of BIO-150, Introduction to Biological Inquiry and three sections of BIO-251, Molecules, Cells, and Organisms.

Historically, we’ve told students that it doesn’t matter which section of a course they take. We set core topics for each course [3]; everyone covers them. The broader learning outcomes are the same even in courses with different themes, such as CSC-151 and BIO-150. And, although different faculty teach different sections, our faculty are generally very good or excellent teachers.

However …

However …

However …

We also score faculty on their teaching and give them different salaries based on those scores. Doesn’t that suggest that some faculty are better than others, in some nebulous sense of better? And wouldn’t we expect students to want to take classes from the better faculty?

Of course, as I tell my students, better is not only nebulous, it’s also inconsistent. We teach differently. Some students preferred taking classes from me to other CS faculty. Some students avoided my classes. I can’t teach like my colleagues, or at least I’ve failed when I’ve tried. They can’t teach like me, or at least they do less well when they try.

Does that make a difference in terms of student learning? I don’t know; we have yet to conduct a formal research study [5]. But I suspect that students learn about the same amount, no matter who they have. Indeed, the research studies on learning styles suggest that students learn the same whether taught in their preferred learning style or another.

On the other hand, we should see some attitudinal effects. If you enjoy having a particular faculty member, you may be inspired to try harder or to persist in the discipline. Some faculty may seem more welcoming, others less so.

I’ve also skipped over the issue of themes or course variants. For courses with different themes, wouldn’t we expect students to find some themes more attractive than others? That’s one of the reasons that CS offers multiple themes; we’d likely attract a broader group of students by choosing different topics. Again, I wonder if there’s an attitudinal difference.

In the end, it’s likely more important that a student takes a particular course than that they take it with a particular faculty member. For example, I have some advisees taking CSC-207 in the fall. Since they’re my advisees, I’d like them to take it from me, and I think they’d like to take it from me. But we’d both rather that they take CSC-207 than not take it; if I weren’t teaching 207 until spring, we certainly wouldn’t have them wait. And I’ll admit that I don’t remember whether my advisees taking 207 are all taking it from me.

Why does this matter? In part, it’s because we’ve switched policies about balancing sections. Grinnell used to balance sections because it helps ensure that students have similar experiences and increases the odds that everyone who wants the course can get it. We no longer do so [7]. I recall a recent discussion in which it was suggested that we go back to balancing. Balancing assumes that all sections are equivalent.

So why balance?

First of all, we’ve seen more cases of students being closed out of multi-courses (e.g., because the only section they can take fills, even though there other sections are open and students in the filled section could have switched [8]).

We also see multi-section courses with large discrepancies, say a situation in which one has 30 students, and one has six. Those in the large section will get less individual attention and a more over-worked faculty member. Those in the small section will hear a less broad range of opinions and ideas from colleagues. The students in both sections would be better off if there were two sections of 18 [9].

I prefer that we balance. But that’s because I want everyone who wants a CS class to be able to get a CS class [10].

I have a vague memory of being told that it may be illegal to balance students to other sections or at least that it may be illegal if the different sections have different textbooks. I believe that’s one of the implications of The Higher Education Opportunity Act of 2008. But most multi-section science courses use the same text [11].

Where was I? Oh, I was addressing the Why does this matter? question. At least, I think I was. This musing is one of those which represents my attempts to think through an issue. The issue my muse raised was, Isn’t it inconsistent for Grinnell to say that the section doesn’t matter at the same time the College scores faculty on their teaching? (or, as the title suggests, Aren’t we sending mixed messages?). I’m not sure that I’ve answered those questions to my satisfaction. Nonetheless, I’ve thought through the issues as much as I care to today. I should remember that in my ideal world, we don’t evaluate faculty for raises, only for improvement.

[1] I thought it was called something like Imperatative Problem Solving and Memory Management or Imperative Problem Solving and Baby Data Structures. I’m wrong. We limit the long title to the third course in the sequence.

[2] I’d prefer that we call it Algorithm and Data Structure Design and Analysis or just Algorithms. Maybe CS7, but few people still know that name.

[3] If we’re following institutional instructions, we indicate learning outcomes [4] rather than topics.

[4] I’m not sure why, but I still struggle with the relationships between learning outcomes, learning objectives, and learning goals.

[5] I suppose we could do such a study. We have specified learning outcomes for our courses. We could compare how well students do on those learning outcomes, provided we had agreed-upon mechanisms for measuring those learning outcomes. And there are some standardized tests. For example, Chemistry may have students take an ACS test at the end of organic. However, I worry that assessing faculty in this way may lead to teaching toward the test rather than what we really want: joy of learning, enthusiasm for the subject, intuition, collaboration, and more [6].

[6] I realize that these are also potential learning outcomes. I’ve yet to see a good measure of any of them.

[7] With a few exceptions.

[8] In my classes, I rely on self-gov to help address this issue. There are students who can’t take the other section. Would any of you be willing to switch? That usually frees up a few slots.

[9] I have yet to encounter this issue in CS. If we do, I plan to rely on self-gov to help address this issue.

[10] Subject to the number of offerings and the class capacities.

[11] CSC-207 is an exception. One section uses CLRS4 [12]. The other uses Stuart Reges’ Building Java Programs: A Back-to-Basics Approach. CLRS4 is $80 used, $141 new, and $100 from Amazon. Reges is $33 for day-one access, whatever that means [14,15].

[12] Cormen, Leiserson, Rivest, Stein. Introduction to Algorithms. 4th Edition.

[14] Day-one access appears to be an electronic book rental. The materials are available on Blackboard. I assume that they are no longer available after the semester.

[15] I’m sure that I’ve been told about day-one access. However, I don’t use Blackboard, and I generally prefer to avoid e-texts, so I probably let it go in one ear and out the other. Nonetheless, I appreciate that the College provides a legal, low-cost alternative.

Version 1.0 of 2023-05-20.