What does current evidence suggest: doing group work in mixed/balanced gender groups or doing group work in single gender groups?


  • College level mathematics/science course
  • Group size approximately 3 or 4 (so for mixed gender groups there will often be a solo-male or solo-female; if there's research on which of these are better/more damaging I would also welcome that).


Looking at this review article from 1999 for example, there are competing theories of how gender balance affect performance in group work. There are some that advocate single-sex groups to minimize discrimination and conflict, or to maximize satisfaction and supportive behavior; there are some that suggest that majority-minority groups will lead to the minority members being ostracized; there are some that suggest precisely the opposite.


When planning group works for groups of sizes 3 or 4 college students in an introductory mathematics course, is it better to have

  • all male, all female groups versus mixed groups;
  • 1 male 2 female versus 2 male 1 female versus 2 male 2 female groups?

Please only answer with evidence-based research, and not theoretical speculation.

  • 3
    $\begingroup$ Especially given the answer, I wonder how one could/would/should go about implementing it without negative side-effects (mainly in the form of misunderstandings about the motivation). I do not want to discuss this here in the comments. The reason I bring it up is that I might end up asking a follow up question, but do not want to front-run you if you might also consider asking a follow up. $\endgroup$
    – quid
    Aug 18, 2016 at 23:29
  • $\begingroup$ @quid: go ahead and ask the followup! (For my class I expect to have the students do a first round of self assembly, and only reshuffle if something is not working out well. I can always just explain it in terms of "observed rates of participation and other course performance" issues.) $\endgroup$ Aug 19, 2016 at 13:41
  • $\begingroup$ Thank you for the reply. I'll have to see how to phrase it to make it a viable question for the site. $\endgroup$
    – quid
    Aug 19, 2016 at 14:28
  • $\begingroup$ May I suggest that "self-assembly" is dangerous---that's too loaded---or at least sub-optimal. There are social/cultural biases that will guide self-assembly in a direction you likely do not want to foster. Random assignments are surely superior to self-assembly. I myself randomly assign with a simple program that I run in full view of the students so they know it is random. It's kinda exciting to see with whom you will be matched. :-) $\endgroup$ Aug 19, 2016 at 21:30
  • $\begingroup$ @JosephO'Rourke: for a bunch of freshmen who just met each other the first day of classes, I'd say that self-assembly is a close proxy to a random assignment. Without me having to actually run the random number generator. I am actually kinda curious given the choice whether the women would stick together or not. $\endgroup$ Aug 19, 2016 at 21:52

2 Answers 2


Here is one article in PNAS. The final sentence quoted below is a summary: "creating small groups with high proportions of women [...] is one way to keep women engaged [...]"

Dasgupta, Nilanjana, Melissa McManus Scircle, and Matthew Hunsinger. "Female peers in small work groups enhance women's motivation, verbal participation, and career aspirations in engineering." Proceedings of the National Academy of Sciences 112.16 (2015): 4988-4993.

Some quotes from the Abstract:

We provide experimental evidence showing that gender composition of small groups in engineering has a substantial impact on undergraduate women’s persistence. Women participate more actively in engineering groups when members are mostly female vs. mostly male or in equal gender proportions. Women feel less anxious in female-majority groups vs. minority groups, especially as first-year students. Gender-parity groups are less effective than female-majority groups in promoting verbal participation. [...]

These data suggest that creating small groups with high proportions of women in otherwise male-dominated fields is one way to keep women engaged and aspiring toward engineering careers.

  • $\begingroup$ Wow, this study answers pretty much exactly the question I want to ask in the setting I need to apply. I guess with it being as new as it is the effect has not yet been confirmed by followup studies? $\endgroup$ Aug 18, 2016 at 19:00
  • $\begingroup$ @WillieWong: The first author, "Buju" Dasgupta, subsequently wrote: Dasgupta, Nilanjana. "Viewpoint: How Stereotypes Impact Women in Physics." Physics 9 (2016): 87. (Web link.) But that is not a followup study. Yes, probably too early for independent studies on the same topic. $\endgroup$ Aug 18, 2016 at 19:05
  • 1
    $\begingroup$ Right, that Physics article is not really about what I am after. I'll leave the question open for a few more days, but if I read the article in your answer correctly, the authors claim to be the first to study the impact of intermediate gender ratios on women in STEM (previous works mostly on the difference between solo woman in an otherwise all male group versus woman in an all female group). Which would mean that you have given essentially the unique answer to this question... $\endgroup$ Aug 18, 2016 at 19:25

Peter Liljedahl has done research showing that visibly random grouping increases student participation. I have been doing this in some math classes, and I like how it is working out.

Citation: Liljedahl, P. (in press). The affordances of using visually random groups in a mathematics classroom. In Y. Li, E. Silver, & S. Li (eds.) Transforming Mathematics Instruction: Multiple Approaches and Practices. New York, NY: Springer.

  • 1
    $\begingroup$ Cool. Can you add a full citation to the paper (title, author list, where published), so we can find it if the link stops working in the future? $\endgroup$
    – D.W.
    Aug 19, 2016 at 6:36
  • 2
    $\begingroup$ That's very nice. I think it bears mentioning that the practice in that paper is intimately connected to re-scrambling the groups on a daily basis (in a high school course). If I'm ever forced to use group work that's how I'll want to do it now. $\endgroup$ Aug 19, 2016 at 7:01
  • $\begingroup$ I've used this method in the previous semesters in two different courses. And it does work well. But the research does not control for gender differences! The model population in this research are bimodal Chinese-immigrant and Caucasian Canadian students with roughly equal representation. Stereotypically both groups are known as "strong in mathematics". This is not the case in male-female breakdown typically. College math courses often exhibit gender disparity in student population, and female students often have anxiety about stereotypical perception as "not being good at math". $\endgroup$ Aug 19, 2016 at 13:52
  • 2
    $\begingroup$ I would be more convinced about the random assignment strategy if the research actually controls for gender, or if it involves minorities who stereotypically "underachieves in mathematics". $\endgroup$ Aug 19, 2016 at 13:54
  • $\begingroup$ For example, research by Sekaquaptewa and Meadows indicated that left to their own devices, students in group work often take up "gender-stereotypical" roles, with man assuming "leader" positions. They observe this in group presentations where the "meat" of the presentation is often left to the male students while the women often end up giving shorter, peripheral discussion. These are the sort of things that necessitates micromanaging group assignments when there are underrepresented.. $\endgroup$ Aug 19, 2016 at 14:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.