Evolution 2014: Day 2

About five years ago I developed my Evolution course, which is aimed at my non-majors art and design students. I have not taught this course in more than two years, and as it has sat on the shelf I have been able to get that critical distance necessary to make the course better. So coming to Evolution 2014 this summer has been really well-timed, as it has given me the opportunity to seek out new ideas for my own course. Already this meeting has been really valuable to my teaching practice, as Friday’s Experiencing Evolution workshop got me thinking of ways that I can incorporate a lot more lab and lab-like activities into my classroom. This morning’s two-session Education Symposium (“Assessing Undergraduate Student Understanding of Evolutionary Biology“) added a whole new dimension to the way that I am thinking about bringing my curriculum “to the next level”.

Rebecca Price and Tessa Andrews began the session with a talk entitled “Identifying students’ misconceptions about genetic drift and using them to improve instruction”. From what I gathered, Price and Andrews had been part of a NESCent working group that sought to identify areas of undergraduate evolutionary biology education that lack clear conceptual frameworks and tools for assessing student learning. Two of the areas identified — genetic drift and evodevo — were the subject of talks in this symposium; Price and Andrews focused on genetic drift. Drift is a concept that is easy to push to the side, especially in the non-majors classroom. It is a concept that requires students to thinking probabilistically, to imagine large swaths of time, and to let go of their misconceptions that all evolutionary change is “directed” by natural selection. But it is these very challenges that make it important to teach drift: if students understand the concept of genetic drift, then they have the potential to better understand evolutionary change as a whole. Neutral evolution is our null model, and making it at least a potential mental model that our students can employ is valuable.

Price and Andrews discussed the process of developing GeDI, the Genetic Drift Inventory of concepts. They also developed a conceptual model of how students learn the concept of genetic drift; these “models of learning” are clearly something that I have implicitly assumed, but I was struck by the (oft-overlooked) importance of explicitly stating how you expect students to obtain understanding. The GeDI can be used to guide pre- and post-testing efforts designed to assess how well students understand drift as a concept. Price and Andrews shared some examples of the sorts of questions they asked. One cool feature of these questions was that they were framed so that students were asked to portray the thinking of an evolutionary biologist, not their own beliefs; this way of asking the question assures that student beliefs (for example, young earth creationism) are not confounded with true student misunderstanding.

Kathryn Perez followed up with a related talk (“Student misconceptions about evolutionary developmental biology and using the EvoDevo Concept Inventory to document student learning”) that was also inspired by the same NESCent working group. Following a similar process as Price and Andrews, Perez and her collaborators formulated a concept inventory through a combination of student surveys and interviews with both students and experts in the field. What I thought was particularly interesting was that she was able to identify that the biggest gaps in student understanding were related not to evolution itself but to developmental biology. This is not a surprising result to me: if evolutionary biologists are usually the ones who are teaching courses that introduce the idea of evodevo, it is not surprising that the developmental biology concepts are being delivered least effectively; I would be the first to admit that my own training in and knowledge of developmental biology is quite shallow. Perez also showed results of a small-scale study where new teaching modules aimed at teaching critical evodevo concepts were incorporated into half of the sections of an evolution course. These modules significantly improved student comprehension of a number of concepts critical to evodevo.

Ross Nehm is one of the foremost researchers into how students gain understanding of the evolutionary process. His talk, “Assessing students’ mental models of evolutionary change across the tree of life using the ACORNS instrument“, provided an overview of a new tool that can be used to assess how well our students grasp evolutionary concepts. He began by discussing his explanatory model for how students think about evolutionary change. Students can obtain what he calls “normative” ideas in science: the concepts that scientific professionals might use. But they also carry around non-normative “naive” ideas that they have obtained before coming into our classes. Nehm argued that we need to recognize that students employ these two different classes of models in different ways and in response to different challenges. To one challenge a student might bring only normative ideas picked up in scientific training, whereas to a different challenge a student might only bring non-normative ideas. But there’s also the possibility that students might bring ideas from both normative and non-normative sources to a particular challenge, either in some combined form or as separate-but-simultaneous plural ideas.

Nehm emphasized how important it was to recognize this diversity of student approaches to the challenges we provide, and to beware that when we change challenges students may bring a different approach. I found this way of conceptualizing student conceptualization to be really helpful, and a lot more nuanced than the more simplistic “students either get a concept or don’t” assessment. Nehm also showed some really interesting data showing that the tendency of students to bring different amounts of normative scientific conceptualization to bear on different challenges is cross-cultural: students in Germany, Indonesia, China, and Korea showed the same general pattern of “varying responses” as students in the United States. Nehm noted that this was particularly interesting given that Chinese students are not exposed to many of the religious ideas that we often blame for fostering misconceptions in our American students.

Nehm has developed an assessment tool that will allow teachers to assess how well their students incorporate the normative scientific concepts we hope to be teaching into their answers. Called ACORNS for Assessing Contextual Reasoning about Natural Selection, the assessment uses essay-style open-ended questions to force students to not simply find the right answer but to produce their own understanding. A lot like the new essay-style questions that were introduced by SimBio, the ACORNS questions are themselves structured in a way that somewhat standardizes our assessments. We can ask lots of different questions about evolutionary biology using the ACORNS structure, but variation between different sorts of questions is structured so that we can fully understand what we are varying. This seems like a great means of taking a more systematic approach to creating our assessments, but how to properly “grade” or “score” them for student understanding? Nehm and his team have an answer to that problem: EvoGrader. I have not played around with it yet, but apparently you upload student answers and the algorithm tells you how many normative concepts students have correctly employed in their answer.

EvoGrader uses what Nehm calls “machine learning”, but might be more appropriately called “artificial intelligence”. Perhaps saying “machine learning” lowers expectations, but the algorithm had better display sufficient intelligence to be valuable. Although their two tools are not the same and do not seek to accomplish the exact same goals, it is interesting that both the EvoGrader project and SimBio‘s WordBytes questions are trying to take human labor out of assessment. There’s an interesting public-private competition developing here, with Nehm’s lab developing free assessment tools using public grant funding, and SimBio developing fee-for-service assessment tools using at least in part private money. Given my work and general ethic, you can bet who I am rooting for, although I wish both projects success. If either works, they will represent a significant accomplishment. Faculty — especially those at teaching institutions — could be freed up to focus on better teaching rather than spending hours on grading. For Nehm’s project, the computational approach to assessment also takes human variation and error out of large-scale studies of student understanding. Again, whether or not this is good or bad will entirely depend on how smart the algorithm is.

A huge message coming out of Nehm’s talk was that we need to be very mindful of the context in which we test for student understanding of evolutionary concepts. According to his studies, conceptual understanding is not independent of the content with which it is paired: ask a question about natural selection in birds and students may bring different ideas to bear than if you ask an analogous question about natural selection in plants. This is frustrating to face, but also exciting to contend with. I can see a lot of ways in which I might address this problem in my classes, especially as I try to make them more explicitly conceptual.

The final talk of the Education Symposium (“Curricular effectiveness as measured by the Evolutionary Attitudes and Literacy Survey (EALS)“) was given by Patricia Hawley. Hawley is a very interesting person: she is trained as a psychologist, but also has a background in animal behavior. Her unique background allows her to take some unique perspectives on the problems associated with teaching evolutionary concepts. She began her talk by establishing a goal: to create cultural change by increasing the number of people in our society that understand and accept evolutionary biology as a science that explains the existence of the living world. She suggested that secondary school biology teachers may in fact be a conduit for negative attitudes towards evolution, as they are often overly cautious about teaching the subject and don’t have rates of belief in evolution that exceed those of colleagues with other subject expertise. We — the higher education faculty — are responsible for teaching evolution to our biology majors who become teachers; Hawley wants to know what we are doing wrong.

To get at this question in a broad number of areas, Hawley created a survey called the Evolutionary Attitudes and Literacy Survey (EALS). The survey asks students to rate their agreement with statements in sixteen multi-item “subscales”, which can be used to understand the attitudes and perspective of each student taking the EALS. The survey can be administered once to get a sense of where your students are at, and can also be given at the end of a course to see how much that course transformed student ideas. The survey, including all the documentation needed to properly administer it, is available from the EALS site.

Hawley has administered the EALS to a large sample of students and has discovered some interesting  trends. A number of the subscales cluster together in unsurprising ways; for example, students who agree strongly with statements indicative of conservative attitudes are also likely to agree strongly with statements indicative of religious attitudes. But the subscale associated with evolutionary misconceptions does not correlate with any other subscales, including those that measure knowledge of evolutionary biology. As Hawley described it, misconceptions seem to be an independent psychological phenomenon, one that frustratingly does disappear with education.

Hawley also showed data on changes in attitudes and understanding in three courses: her evolutionary psychology course, an introductory biology course, and a political science course (the control). Administering EALS before and after, she found that the evolutionary psychology course did the most to improve both student understanding of and positive attitudes towards evolution. The introductory biology course actually led to some loss of understanding. She presented this as an indictment of the introductory biology course — and it is — but some later questions from savvy attendees brought out the fact that comparing her course to the stereotypical introductory biology course might not be the most informative comparison. I would have liked to see her course compared to a course exclusively focused on evolutionary biology. This comparison would be fairer (because both courses would likely spend comparable amounts of time on evolution) and more interesting (because the major difference between the courses is the taxa and adaptations of interest).

I enjoyed both Nehm and Hawley’s talks, but they seemed to present slightly different ideas of “what we should do”. Nehm made a compelling argument that we need to ask our students to explain evolution across a broad spectrum of taxa and for a variety of traits to assure that students begin to apply evolutionary models in manner that show expertise. Hawley lobbied for the use of a very narrow course focused solely on human behavior as a way of better changing student attitudes and understanding. Maybe both approaches can be effective, but these are certainly very different prescriptions for “what we should do”.

In the afternoon I took it easy, targeting particular talks rather than sitting through whole sessions. I have learned that it is more valuable for me to take the time processing a few talks than to just beat myself into a mental pulp trying to soak in twenty talks a day!

Karin Pfennig presented an interesting talk (“Sexual selection impedes ecological specialization“) that showed that female spadefoot toads can exert sexual selection pressures that reduce the potential for ecological specialization in offspring. It turns out that spadefoot tadpoles have the ability to go from being omnivorous to carnivorous in environments that offer sufficient planktonic prey. This plastic response is irreversible, changing developmental trajectories to produce two distinct morphs. This creates disruptive selection, because intermediate forms do most poorly. What Pfenning found in controlled enclosure experiments was that females who were allowed to mate with males of their own selection produced offspring with reduced fitness when assuming the carnivorous morph. She hypothesized that sexual selection is serving to break down trait integration that is critical to producing a viable carnivorous specialist. While the effect of sexual selection on specialization was not large, there was an effect, reinforcing the fact that sexual selection can oppose ecological selection pressures.

William Harcombe‘s talk (“Conflict increases cooperation between microbial species“) presented data on the role of competition in facilitating the evolution of mutualistic interactions between microbes. Harcombe has evolved his own system for studying microbial mutualism by selecting for mutants of E. coli and engineering a strain of Salmonella such that the two strains depend on each other for critical metabolic products. In these experiments he compared the proportion of Salmonella that contribute to the mutualism to the proportion of wild type Salmonella, which do not contribute to the mutualism. He showed that when a third strain that is mutualistic with E. coli but competitive with Salmonella is introduced, the prevalence of Salmonella cooperative behavior goes down. This finding contradicts the idea that stress always favors cooperative behavior. When a third strain that was purely competitive with both E. coli and Salmonella was introduced, cooperation went up. These contrasting findings suggest that simplistic thinking about the role of competition cannot be used to predict prevalence of mutualism. Instead, the exact configuration of competition will determine whether mutualism will be sustainable. An added value to Harcombe’s study was his use of models to provide a viable explanation of his experimental results; using the COMETS model he showed that the metabolic differences between strains can account for the different interaction dynamics produced by each set of strains.

Michael Wells‘ talk (“Big groups, bad eggs and biogeography: regional and global patterns of brood parasitism’s effect on cooperative breeding“) took on a similar question: does brood parasitism lead to the evolution of cooperative breeding? While it is possible that being targeted by a brood parasite might lead to the evolution of cooperative breeding (because extra parental care can be used to defend the nest from parasitism), it is also possible that brood parasites have evolved to target cooperative breeders because they can provide better care to the “donated” offspring. Wells had a great method for distinguishing between these contrasting hypotheses: he used phylogenetic methods to look at the prevalence of transitions between non-cooperative and cooperative breeding and between being and not being targeted by brood parasites. What he found, particularly in African species, was that the transition from non-cooperative breeding to cooperative breeding has been much more likely to occur first. This suggests that in most parasitized cooperative breeders, cooperative breeding preceded parasitism; the brood parasites appear to be targeting — not selecting for — cooperative breeders.

Deanna Soper presented a talk (“Love “bugs”? Exposure to parasites increases promiscuity in a freshwater snail“) showing that female snails (Potamopyrgus antipodarum, which comes in sexual and asexual varieties) increase their level of mating effort and number of male partners when exposed to a sterilizing parasite. Although Soper’s study established that females change their mating strategy in the presence of the parasite, why is not entirely clear: the females could be simply increasing overall reproductive effort because they face higher risk of “reproductive mortality”, but they also could be maintaining the same reproductive effort while using promiscuity to increase the genetic diversity — and thus the potential for resistance to parasites — in her offspring.

To finish off the day, Mohamed Noor gave the SSE Presidential Address, “Recombination suppression helps hybridizing species persist, and perils of a career in evolutionary biology“. A nice introduction by Richard Lenski painted Noor as a talented and generous polymath who is a talented researcher as well as a dedicated mentor, teacher, and contributor to the activities of the Society for the Study of Evolution. Noor began his talk by highlighting a few of the challenges faced by contemporary evolutionary biologists; I was especially thankful that he acknowledged the hiring crisis in evolutionary biology and academia in general, although he did not mention the culpability of research universities in contributing to this problem. He also discussed a number of things that SSE is doing to address these problems, including the formation of a Careers Committee that he described as being aimed at teaching graduate students how “not to be a clone of your PhD advisor”. I was also excited to learn about the impending launch of an EvoED Digital Library, which I assume will be a complement to the long-standing EcoED DL.

Although I do not have the molecular biology background required to hang with every element of Noor’s mad-cap talk, I can give a pretty basic summary of its messages. Noor is a member of that large army of biologists who study Drosophila, one of the genetic workhorse genera/species (along with unrelated but close friends C. elegans and Arabidopsis). Here he presented work considering the sympatric interaction of two Drosophila species, D. persimilis and D. pseudoobscura. The question Noor wished to tackle was “why don’t these species hybridize and become one species?”. Apparently the answer lines in chromosomal inversions, inversions that might explain how these two species separated in the first place, and inversions that may help maintain the separation. Noor’s talk focused on the latter issue. He showed a great variety of data showing that hybridization is impeded by these chromosomal inversions, although it did not sound like this evidence was conclusive.

The talk was also very funny! Noor has a self-deprecating sense of humor (and body language) that makes him seem very approachable and humble despite being so accomplished. To add some flavor to the talk, he periodically paused to recap not just the major messages of his scientific presentation but also to point out all the stereotypical (and obnoxious) tropes that he had incorporated into his talk. These included some very funny things such as:

• Using the word “we” to describe work that was done entirely by a hardworking graduate student
• Presenting ten year old data
• Blaming collaborators for the omission of data important to the talk
• Pretending as if work done in his lab was the only work on the subject
• Including an image of your kid

There were many more of these tropes, and Noor suggested that they be used as a drinking game: every time you are at a talk where one of these tropes is employed, take a drink. You could get very intoxicated taking Noor’s suggestion at a meeting like this, especially if you go to the marquee talks!