Education research: Heterogeneous gender student groups

Patricia Heller and others at the University of Minnesota have general guidelines regarding the optimal formation of student groups for collaborative in-class activities in physics (such as those found in the Lecture-Tutorials for Introductory Astronomy (Adams, Prather, & Slater, 2005)), as reported by Randall Knight at the California Polytechnic State University, San Luis Obispo, CA (Knight, 2004):

...The initial groups [at the University of Minnesota] are formed at random. Once the first test scores are known, mixed ability groups are formed with one student each from the top, middle, and bottom thirds of the class. Students rotate to a different group every three weeks. Groups sizes of three are optimal, two are below critical mass, and five or more generally finds some members not engaged. As for gender balance, groups should be homogeneous or have two women, one man. Groups with a single women tend to not function well, even if the woman is the "top third" member.

Jeffery P. Adams et al. (2002) studied self-formed groups of four students in astronomy (as opposed to assigned groups), and found that:

Specifically, females in mixed-sex groups of unequal numbers (e.g., three females and one male or three males and one female) were less likely to display behavior indicating active engagement than females in all-female groups or in groups with equal numbers of males and females.

Thus the optimal formation of three (or four) students in a heterogeneous group should strive to have two-one (or two-two) females to males. Homogeneous groups are also allowed.

The assignment of student groups in Astronomy 10 (introductory astronomy) at Cuesta College in San Luis Obispo, CA is done alphabetically for the first class, even though this may violate the optimal heterogenous gender rules discussed above. This is mainly done to facilitate a spot-check on enrollment, and to take photos for identification by the instructor, and is only done on the first day of instruction.

Students take conceptual and demographic surveys at the start of the first class, and these results are compiled before the second class. Students are then sorted into new groups using top/middle/bottom thirds of their self-reported confidence in math and science for every subsequent class until after the first midterm. By then students are segregated into top/middle/bottom thirds of exam performance once a week (every other class meeting).

Near the end of the semester, some students have observed that they have never been in the same group as certain other students--these are typically the highest performing students in class, and thus being in the same top third, cannot be in the same group. They told that since they are as "equally smart" as those other students, they can never be in the same group. This is also the same answer given even if the students in question are in the same bottom third of exam performance!

Adams, J. P., Brissenden, G., Lindell, R. S., Slater, T. F., & Wallace, J. 2002, "Observations of Student Behavior in Collaborative Learning Groups," Astronomy Education Review, 1(1), 25.

Adams, J. P., Prather, E., & Slater, T. F., Lecture-Tutorials for Introductory Astronomy, Prentice Hall, 2005.

Heller, P., Hollabaugh, M., "Teaching problem solving through cooperative grouping Part 2: Designing problems and structuring groups," American Journal of Phyics 60, 1992, 637.

Knight, Randall D., Five Easy Lessons: Strategies for Successful Physics Teaching, Addison Wesley, 2004, p. 31.