When published research on teaching doesn't help you, why not use your own classroom as a laboratory?
One of my most frequent head-scratching moments in teaching occurs when I have to divide students into groups for some classroom activity.
I use small groups in class maybe once every week or two; I try to reserve them for those moments when I feel the students are beyond the point of needing my explicit guidance for a task, but not yet ready to complete it on their own.
So, for example, my "Introduction to Literature" course begins with a couple of weeks in which the students analyze poems through a combination of writing, lecture, and class discussion, all led by me. But by the fifth week, they will have to turn in a paper analyzing a poem that we have not discussed in class (and that has been published within the past year, so they can't find any help online). To help prepare them for that assignment, I use a small-group activity in the third or fourth week in which I put them in groups to analyze a poem they have never seen before.
And that's when I run into trouble. I have read plenty of research on the most effective ways to divide students into groups — whether to allow them to self-select, to organize them according to ability, to make the groups homogeneous or heterogeneous, and so on. Unfortunately, I have yet to read anything that strikes me as absolutely conclusive. Almost all of the guidance available offers context-specific solutions: i.e., for students who are working together on X kinds of tasks, in Y kinds of classrooms, use the Z strategy to form groups.
During the semester, I'm too busy with multiple course preparations to think so carefully about the X's and Y's of group formation, so my solution has always been to vary my methods — sometimes homogeneous, sometimes heterogeneous, sometimes random — in the hope that I'll get it right at least once in a while.
Not an ideal solution, of course, but it was one I was prepared to live with until a couple of months ago, when I discovered that a colleague had approached this same issue in a much more thoughtful way. He used his own classroom to conduct research on group formation, and presented his findings in January at the annual meeting of the Allied Social Sciences Association, in San Francisco.
Tom White, an associate professor of economics at Assumption College, in Worcester, Mass., became intrigued by the idea of conducting classroom research in 2006, when he attended a three-day workshop on active learning put on by the Teaching Innovations Program of the American Economic Association. The program invited participants to submit proposals on new approaches to economics education and present them at the ASSA meeting.
Tom took the bait. In the fall of 2008, he re-engineered his corporate-finance course to include six group exercises throughout the semester. He taught two sections of the course, which presented him with the ideal opportunity to test whether a particular group-formation strategy would make a difference in student learning. He focused on the question of group consistency: Would students who worked with the same groups throughout all six exercises fare any differently than students whose groups were reconfigured for each exercise?
To help him set up the experiment, he combed through the research literature on group formation and devised a process for forming groups that involved requiring students to complete an online survey. They were asked for basic demographic information as well as their academic profiles and attitudes toward learning.
In one section of the course, he used the survey results to create groups that were heterogeneous according to gender and academic ability, and remained consistent through all six exercises. Although students in the second section completed the survey as well, they were placed in different, randomly selected groups for each of the six exercises.
Tom designed the experiment so that he had multiple ways to evaluate student performance. For most of the in-class exercises, according to his paper, "students were asked to complete a short case or problem set during the first part of the class. After completing the group exercise, students were given a quiz over the material covered in the exercise. Student grades for the exercise were based on individual performance on the quiz, the group's average quiz grade, and the group's written report for the exercise."
The multiple grading formats were designed to ensure "individual and group accountability," as well as to foster "positive interdependence." But in addition to evaluating the student grades on the six in-class exercises, Tom also looked at their exam scores in the course.
Before I reveal his conclusions, I have to offer the caveat that when articles on educational research include Greek letters and phrases like "dummy variables" and "estimation of regression" (as they probably should), I run up against the limits of my literary-oriented brain. So this is going to be a simplification.
He found that different methods of analyzing the data yielded four sets of results: In two of them, the group formation made no difference; in the other two, the groups that remained consistent throughout the semester outperformed the randomly chosen groups on both quizzes and exams. His modest conclusion: "While the empirical results are far from conclusive, it appears that there is some evidence that using [consistent] groups for cooperative learning exercises may provide benefits over [random] groups."
As his colleagues at the San Francisco meeting suggested, he will undertake further research next semester, attempting to refine and replicate the experiment with a larger sample of students. That is, in part, why he was initially reluctant to have me write about his early findings. His results were not as conclusive as he had hoped.
I explained to him, though, that my point was not to convey to readers the conclusions he had drawn from his specific experiment. Instead I wanted to present the profile of a working faculty member — teaching seven courses a year — who decides to make a change in his teaching, to assess it with an experimental design, and to present it to his disciplinary colleagues.
That seems important to me because I frequently hear faculty members complain that calls for pedagogical innovation and educational research come from people who spend their time doing educational research, and who aren't slogging it out in the trenches with big course loads. I sympathize with that complaint. I also know from my own experience how challenging it can be to take research done in some other discipline and apply it to my own classroom — or even to take abstract pedagogical advice and translate it into practice on Monday morning.
The solution, however, is not to turn our backs on educational research. It's for more of us to undertake the kind of discipline-based research that Tom conducted in his own classroom. Think carefully about a question or problem in your own teaching, come up with a creative solution on your own, and then figure out a way to assess whether it works. Tell your colleagues about it at your annual convention, or in the pages of your discipline's teaching journal.
If you're like me, and you have little or no experience conducting research in the social sciences, you can probably get help designing the experiment from the teaching center on your campus. Will such basic disciplinary research on teaching gain the attention of the world of higher education, and make you a pedagogical expert? Probably not.
But maybe, like Tom, you will find that the opportunity to experiment with your teaching and reshape your courses is rewarding on its own, and can help you break out of your usual pedagogical routines.
James M. Lang is an associate professor of English at Assumption College and author of On Course: A Week-by-Week Guide to Your First Semester of College Teaching (Harvard University Press, 2008). He writes about teaching in higher education, and his Web site is http://www.jamesmlang.com. He welcomes reader mail directed to his attention at careers@chronicle.com. For an archive of his previous columns, see http://chronicle.com/jobs/news/archives/columns/on_course.
