Between a rock and a hard concept: teaching multilevel selection


This post is by Leslie Curren.

I recently came across Amy’s post from last October about the ecological concepts people find the most difficult to teach, and knew my own answer immediately: multilevel selection. I was surprised no one else had mentioned it, although perhaps that’s because it is more evolution than ecology. But given the tight links between ecology and evolution, I still expected someone to bring it up—multilevel selection always finds its way into every course I teach, and that includes General Ecology (as well as Animal Behavior, Behavioral Ecology, Intro Bio, and others).

In my biggest course, Animal Behavior, I always begin with the fundamental principles of evolution by natural selection. This allows students to build on that foundation throughout the semester as we discuss the adaptive functions of behaviors. I used to wait to cover social behavior and cooperation until the last quarter of the semester, loosely following the trajectory of the textbook. I found this didn’t work, however—cooperation is so integral to so many facets of animal behavior that during every unit I inevitably heard students list “for the good of the group” and “for the preservation of the species” as frequent explanations for a behavior. After saying, “Not exactly…hold that thought, because we’ll discuss the hypotheses for cooperation later in the semester” several times, I eventually gave in and bumped cooperation to earlier in the syllabus.

My reactions to these explanations from students are always conflicted. On the one hand, based on my (admittedly limited) understanding of multilevel selection, the students aren’t entirely wrong. On the other hand, their answers aren’t coming from a place of high-level understanding of modern multilevel selection theory, but are rather more akin to the simplistic group selection argument of the 1960s. Given that one of my goals in every course is that students leave with a firm understanding of how evolution by NS works at a fundamental level, and considering how difficult these concepts prove to be for many students, I think it’s really important to teach students the traditional group selection fallacy and why it’s incongruous with evolutionary forces acting on the individual.

But I also think it’s irresponsible not to make students aware of multilevel selection theory, especially given the theory’s increasing support. Even if students don’t leave my class with a strong mastery of all the concepts and controversies we address, I hope to at least send them into the world with an awareness of those concepts. (My personal, albeit far-fetched, nightmare: one of my students starting a grad program and being told, “You mean your Animal Behavior instructor didn’t even mention multilevel selection?!? What is that person’s name?”)

These goals are often at odds, particularly at a state university in which the wide range of student abilities/preparation1 can make choosing the appropriate level of difficulty for a class of 80 very challenging. The more in-depth we discuss multilevel selection, the more students I start to lose, and the more I fear I undermine my lesson on what’s wrong with the traditional group selection fallacy. Furthermore, the more in-depth we go, the more we step out of my own academic comfort zone, and I find myself struggling to understand the concept well enough to teach it to keen-minded undergrads who ask probing (and excellent) questions. Pedagogically, I know this latter effect is a good thing, but I worry about leading students astray.

Because of these two hesitations, as well as time constraints, I have historically punted on the subject, relying on every instructor’s trusted friend, the good old “I encourage you to pursue this topic outside of class, but exploring it further here is beyond the scope of this course.” But I am now finding myself unsatisfied with this, especially in my upper-level Behavioral Ecology course.

I’m therefore interested in hearing about other people’s experiences with teaching group and/or multilevel selection (student perspectives would be great, too). I recently found this excellent paper by Daniel O’Brien on teaching multilevel selection, complete with hands-on exercises and follow-up questions, but I’m not sure it reconciles the difficulties I have laid out above. I am, however, excited to try implementing some of the paper’s suggestions this fall, perhaps as a lab activity.

Other ideas welcome! Likewise, if you have experience with other concepts that you think are analogous, I’d love to hear about how you addressed them.

1Yes, Astute Commenter, I know those are different.

2 thoughts on “Between a rock and a hard concept: teaching multilevel selection

  1. As far as I understood the refusals of Williams and Maynard Smith, they criticized naive group selection reasoning, because groups would have needed to be well isolated and for each mutation towards selfish behavior (or a selfish migrant arriving in a group) a whole group would have to go extinct. As group extinction should be less frequent than mutation/migration, the models requiring whole groups to go extinct or thrive and split did not work. In 1975 with David Sloan Wilson published a model in which differential group survival or extinction was replaced by cyclic dissolution of groups in a panmictic pool of individuals and random re-formation of groups from that pool.

    With this admittedly potted history, maybe the discourse can be redressed. Instead of talking about the good of the group/species or individual, you may run the trait altruism through a series of differing population structures (e.g., 1. panmixic, 2. isolated demes, 3. cycles between demes and panmixis) and show how altruism is of no avail to individuals in population structure 1 and 2, but can be so in 3.

    Anyway, something along that line may get you out of the dilemma of teaching multilevel selection theory and setting your students on the wrong path of thinking towards naive group selectionism.

Comments are closed.