Wednesday, January 15, 2014

Don't Teach Phylogeny This Way (Anatomy of a Failure)

I've posted previously on class activities that work. Here's a post on a class activity that failed. Failed spectacularly. Crashed and burned and took out half the class . It was bad.

Never, never, never teach phylogeny this way.

First, a (mercifully short) description of the bad, bad, so very bad class activity, and then some thoughts for improvement.

Bad Activity:
To teach about phylogenic trees and how they are made, I had students divide into groups of 6-8 and imagine they were each representatives of a separate living species. Using only physical traits that were within their control (that is, the clothes they are wearing, length of hair, glasses, makeup, tattoos, etc.), I asked them to create a tree that grouped subsets of "species" based on shared, derived traits. Using one group's tree as an example, we went through as a group and talked about how to read a phylogeny.

What Went Wrong:
The activity wouldn't have been so bad as a brief introduction, except my class schedule was thrown off this semester for various reasons, so I wasn't able to spend as much time on the concepts of cladistics and systematics and I usually do. So, although the students were exposed to different types of phylogenies in their book, briefly in lectures, etc., the class activity, during which they arbitrarily chose random traits to create subgroupings of "species", became their main exposure to the whole concept of phylogenic trees. On the exam, they were given some fake animals (star and pentagon shapes, with weird squiggles representing traits), and asked to create the most parsimonious tree showing the relationships between the animal species. The vast majority of students did exactly what they'd done in class: picked a random trait and used it to divide the animals, and then kept dividing until they'd created a nice-looking tree. But, in doing so, they created trees where some traits had to evolve independently five different times, or more. The trees looked nice, but said absolutely nothing about the evolutionary relationship between the species. 

What Went Right:
Well, they learned how to draw trees. They just didn't learn much about why one would draw them in the first place, or the purpose of the whole exercise. Ugh.

Ideas for Improvement:
I don't think this is a fatally flawed exercise, it just needs to be done differently. 

1. Explain about parsimony at the beginning of the exercise

2. When students get into their groups, have them write down all of the visible traits present in their group. For example, it may be easiest to have students make trees based on clothing alone, so one set of possible traits could include jeans, chinos and leggings. Another set of traits could be fuzzy (wool, knit, etc.), smooth (leather, silk, etc.), and hard (hell, maybe someone comes to class in armor.) It would probably be best to give them a limit to the number of traits they can consider.

3. Then students could use those specific traits to create the most parsimonious tree using the traits they have written down. During discussion afterward, we could talk about how the traits they chose to consider affected the tree they created, and how adding or taking away traits would change their groupings. This would be particularly helpful since we discuss fossils and the difficulties with fossil species in the same unit.

No comments:

Post a Comment