Ethics statement

This project was approved by the University of Washington Institutional Review Board and conducted under Human Subjects Division Application #44438. All students in the course completed the surveys and in-class activities as part of normal course work. Per the University of Washington Institutional Review Board, students were provided the opportunity to have their data removed from the study after being informed that a research study was being conducted in their class and that their data would be analyzed as part of the study. Students could additionally opt out of the study at any time by filing the request at a central office. Thus consent was informed and implicit if students did not opt to remove their data.

Data sollection

Survey.

To assess student perception of group work, we asked students a series of content- and response-processes-validated questions [31] in conjunction with the ASPECT survey (Assessing Student Perception of Engagement in the Classroom Tool). See [30] for details about the survey development and validation process, and for survey items. The ASPECT survey is specifically designed to gauge student perception of engagement in an active learning classroom during group work and is focused on three areas: value of group activity, personal effort, and instructor contribution. During the item validation process [30], two items related to group dynamics stood out as not aligning closely with any of these three areas. As we were interested in how small groups function, for this analysis we focused on these 2 items as well as a third item related to group composition (Table 1): 1) “Are you friends with at least one person that was in your group?” (binary: yes or no), 2) “One group member dominated discussion during today’s group activity” (6-point Likert scale, ranging from “strongly disagree” to “strongly agree”), and 3) “I felt comfortable with my group” (6-point Likert scale, ranging from “strongly disagree” to “strongly agree”). To limit our analysis to students who worked in groups of two or more, we also asked “how many other students did you work with on the activity?” Table 1 lists survey questions of interest and a summary of student responses to these questions. We implemented the survey three times over the course of the quarter in conjunction with group activities described below.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Questions on the survey and the percentage of responses.

Results are aggregated across all three iterations of the survey, so students’ views are repeated.

https://doi.org/10.1371/journal.pone.0181336.t001

Data analysis

Eliminate dominators and promote comfort in groups

Students who dominate group conversations and group interactions have been documented in STEM classrooms ranging from Physics and Astronomy [21] to Biology [41] to Engineering and Math [42,43], as well as across all sectors of the STEM workforce [13]. All of these studies have a key attribute in common: one group member was more likely to dominate when the assignment was high-stakes, meaning that it was graded, would be evaluated by a boss, or would impact promotion, and thus aligned with personal incentives [13,18,21,43]. In contrast, the assignments implemented in this study were low-stakes and not scored for completion or accuracy (although they were collected). However, at two times during the activity, students were asked 3–4 content-based clicker questions which were scored for accuracy and assigned a very small number of course points. These clicker questions may have contributed to the rates of dominators reported in groups, despite the activity itself not being tied to course points. Either way, our data suggest that dominators can influence group function in a negative way, even for low-stakes tasks.

The finding that Asian-American students and international students report a dominator at higher rates than white students could be related to individual students’ learning preferences: as Eddy et al. caution [29], not all students experience group work the same way. For example, there is some evidence that Western-developed student-centered practices such as cooperative learning have not been equally effective when implemented in Asia due to different cultural values [44]. Second generation Asian American students also showed decreased performance when asked to talk while learning a new task, whereas talking improved performance by white students and students from Western cultures [45,46]. In this way, if one student dominates a conversation, it may be particularly jarring to students from cultural backgrounds that place more emphasis on introspection and thinking on one’s own as opposed to a direct relationship between talking as a way to work through ideas. Similarly, students with low-course grades are more likely to agree that someone dominated their group than students with high-course grades. This tendency may be related to lower-performing students desiring additional time to work through content on their own before joining a group. Alternatively, lower-performing students may lack the confidence to challenge their peer’s ideas, leading to one group member dominating the discussion. Overall, our results suggest that certain students are systematically more likely to underperform in classes where dominators persist in group work.

It is important to note that dominators may self-identify. During think-alouds and focus groups (see [30] for details), it became clear that students did not always perceive dominators as negative. Some students contended that, “I think that the person who knows well should be leading the group activity” and others reported “I feel like I dominated a lot of the discussion.” We did not ask students whether or not they were self-identifying as the dominator, so we cannot make the comparison between students who felt dominated and students who themselves dominated discussion. Exploring this difference in group roles would be an interesting future avenue of research.

Comfort was strongly positively correlated with student performance; in fact, reporting being comfortable had an even larger impact on student performance than reporting a dominator (Table 2). It is interesting to note that the vast majority of students agreed that they feel comfortable in their group (Table 1). The single factor that predicted students’ comfort was whether or not they worked with a friend. However, reporting working with a friend was not directly correlated with performance. Similarly, it was not possible to determine which students worked with a friend because almost all students (78.6%) said they worked with a friend. Taken together, this stresses the importance of instructors working toward increasing students’ comfort in groups. Comfort can also be thought of as “safety” wherein groups are safe places for individuals to share ideas [47,48] and to be themselves [49]. Functionally, this is a different task for instructors than demonstrating the safety of the classroom environment or of the instructor-student relationship. Instructors who can best convince their class that their peers are essentially ‘on their team’ will facilitate more complete learning. Intentional strategies that may increase comfort in group work include 1) establishing group norms [19], 2) having many opportunities to perform group work, ideally with groups that are more stable than single class periods (thus practice the skill of working together while getting to know each group member personally; [13,18,19,47,48]), 3) using approaches to help students get comfortable with each other in class (e.g., allowing for social connections within groups; [19,47,48]), or 4) allowing students to self-select into groups so that they can choose to work with friends if they wish [49,50].

Group work

We wondered if certain group activities could negate some of the negative effects of dominators or increase student comfort in groups, in part because student performance was enhanced on the jigsaw activity [32]. We found that, compared to the single-group activity, a jigsaw activity dramatically decreased (by nearly 70%) the likelihood of students more strongly agreeing that someone dominated their group (Table 2, Fig 2). However, the jigsaw activity had no impact on students’ perceived comfort in groups.

Our findings underscore the importance of constructing group work activities so that all students can succeed–specifically by minimizing the chance that one student dominates and maximizing the chance that students are comfortable. To accomplish these goals, we, like others (e.g., [10,18,29,51–53]), offer suggestions. Improved group function is more likely when the task is appropriately complex and requires input from many individuals [7,13,18,53–56], and when group members experience interdependence as opposed to independence [13,56,57]. Interdependence has the additional benefit that it can help decrease competition between groupmates [9], which is known to disproportionately hinder equity as perceived by women and minority students [58,59]. The jigsaw activity implemented in this study is one example of an activity that is appropriately complex and encourages interdependence.

A common way to increase interdependence between students is by increasing structure in the group work [33] and that structure can change either how groups perform or what groups perform. Two major ways to influence how groups perform are to encourage groups to establish norms of behavior and operation [19,60] or to assign individual roles to group members [18]. For example, in an observational study of how effective teams function at a large technology firm, Duhigg [19] found few consistencies other than that high-performing groups had clear group norms.

Furthermore, one way to change how groups perform is by assigning individual roles to individual students [18]. When the task requires input from each role, this is a natural way to create interdependence between students. We caution that assigning group roles is not the same as assigning groups. Although there is some evidence that assigned groups can benefit some learners [61], there is evidence that self-selected groups are best for students from marginalized and minority groups. For example, Cooper and Brownell [49] found that group work increases the pressure on LGBTQIA students (lesbian, gay, Bisexual, transgender, queer, intersex, and asexual) to explain or defend their sexual identity, decreasing their comfort in class, and potentially affecting their performance. However, these same students with non-normative gender identities felt more comfortable in self-selected groups [49]. Similarly, Freeman et al. [50] found that students self-selected into groups with students who were similar to them, first based on appearance (ethnicity and gender) and then by performance.

A carefully constructed activity can also provide structure to what students are doing. For example, there is evidence suggesting that students explain their reasoning more when instructors include explicit prompts (as simple as “explain your reasoning”) on the assignment [41]. Thus it follows that with explicit prompts geared toward interdependence, such as “take turns,” “defend your answer,” and “probe your groupmates for justification,” students may work together more effectively and equitably, though this should be rigorously tested.

Limitations and future directions

Our data do not provide an explanation for why the jigsaw activity resulted in fewer dominators. Thus, we are left with complementary, albeit distinct, hypotheses to explain the patterns we observed: 1) interdependence and self-accountability are key to promoting group equity, or 2) activity logistics and explicit prompts are the key to promoting group equity. By design, highly-interactive activities, such as jigsaws are thought to be at the apex of student cognitive engagement and student learning (see the ICAP framework, [62]) because these activities promote interactive dialogue, co-constructing knowledge, and interdependence. Jigsaws specifically stimulate interdependence because students first become experts then work collaboratively with other student-experts to complete a task. However, the jigsaw activity tested here also had explicit prompts to guide students through the activity. Perhaps it was these prompts, which promoted turn-taking and guided students to act as facilitators not lecturers which aided group equity. Further research testing the synergy or disparateness of these two hypotheses is warranted as a guide to informing classroom instruction.

Another limitation is that we aggregated data from different demographic groups in order to obtain sufficient statistical power to rigorously test our hypotheses. For example, all African-American, black, Latinx, Pacific Islander, and Native American students were combined into a single group termed “under-represented minority” students. Similarly, we combined all Asian-American students whose heritage span all Asian countries and “international” students who encompass all students who are not citizens of the USA. Similarly, we considered students’ sex instead of gender, which is more fluid than the two definitive categories the registrar’s office currently offers: male and female. We recognize that students come from different cultural and social backgrounds so bring unique experiences that impact how they perceive group work [63]. Future studies could seek to disaggregate these groups and test if the observed patterns hold, and whether particular groups of students are driving the patterns. Additionally, qualitative techniques such as one-on-one student interviews or focus groups could provide insight into the range of ways different assignment structures might impact perceptions of equity.

Finally, we did not identify which students worked together in groups, or if students were self-identifying as the dominator. By adding these layers of information, it could be possible to determine how well dominators perform relative to their groupmates and how aware students are of their own behavior. Additionally, if we knew who was working with whom we would be able to determine if a student’s solo status in a group (i.e. being the only person in their group with a particular racial or gender identity) impacts their perception of inequity. Finally, by asking students if they are self-reporting as the dominator we could clarify how to target interventions to reduce the prevalence of a dominator. Do we first need to make students aware that their actions are not promoting turn-taking, or will simply encouraging more equitable participation be sufficient? Our data suggest that group equity and comfort in a group are correlated with individual performance, but further studies into how exactly groups function are merited.

Conclusions

Maximizing functional groups by promoting comfort and discouraging dominators can improve student learning. Here we report that jigsaw group activities can be one way to accomplish these goals. Providing students with opportunities to work collaboratively in high-functioning groups may have the additional benefit of preparing students for the workforce. Specifically, employers rate teamwork as the highest-rated, most desirable recruit attribute, but they also think that new college graduates have fewer teamwork skills now compared to previous recruits [15]. Thus, by continually practicing group work in high-functioning teams, instructors can achieve higher student outcomes and potentially better-prepare students for the STEM workforce [13,18].