Examining the Effect of Homogeneous and Heterogeneous Grouping in Classroom Discourse

• In Event: School-Based Education Research in Diverse Democracies in Three Countries

Abstract

Diverse democracies require a literate, educated public who can engage in thoughtful, informed discourse about complex issues. Unfortunately, global literacy scores indicate that a majority of students are not sufficiently prepared for such discourse (Kelly et al., 2013). Research has shown that classroom discourse can promote basic comprehension, but only particular kinds of discourse are likely to promote the kind of higher-order critical-analytic thinking necessary for effective democratic participation (Murphy et al., 2009). To address this need, we have developed and studied Quality Talk (QT; Murphy, Wilkinson, & Soter, 2011), which is a model of small-group classroom discussion that includes instructional materials on effective discourse, professional development for teachers to implement this model, and assessments that capture changes in both comprehension and critical-analytic thinking. Our ongoing research has shown gains in students’ literacy skills commensurate with involvement in QT (Firetto et al., 2015). In this study we investigated whether these gains differed across different types of student grouping.
We implemented QT in two fourth-grade (n = 28) and two fifth-grade (n = 33) classrooms in a school in the northeastern United States. Oral fluency scores, collected at the beginning of the school year and prior to the QT intervention, were used to classify students as being low, medium, or high performers at that time. Students were matched by performance and then randomly assigned to either homogeneous discussion groups (e.g., all low performers, all medium performers, etc.) or heterogeneous groups (e.g., a mix of low, medium, and high performers).
Two kinds of assessments were used to track students’ literacy performance over the course of QT. At three time points over the course of the academic year, students read a different short text and then completed a multiple-choice comprehension assessment. They also wrote a persuasive essay based upon a prompt about the text. On average, over the course of the intervention students in both grade levels displayed a statistically and practically significant increase on both assessments (i.e., all p-values less than .05, Cohen’s d values ranging from .27 to .83).
Mixed effects ANOVA results showed no statistically significant difference in the growth trend between homogeneously and heterogeneously grouped students, in both grades. We also collected feedback from the teachers on their perceptions of the grouping. While responses were mixed, in general teachers felt that homogeneous grouping was more beneficial overall for reading comprehension, but homogeneous groups were also more challenging to facilitate. On the other hand, some teachers reported there was lower overall discussion quality in the low performing homogeneous group.
Should this proposal be accepted, in the final paper and presentation we will present quantitative findings along with qualitative analyses of videotaped discourse that further illustrate how students with differing oral fluency performance interacted in heterogeneous and homogeneous groups. In sum, we have found evidence that type of grouping does not have a large-scale effect on literacy gains while engaged in QT but particular subgroups of students experience such grouping in differing, important ways.

Authors

• P. Karen Murphy, The Pennsylvania State University

• Jeff A. Greene, University of North Carolina - Chapel Hill

• Carla Marie Firetto, The Pennsylvania State University

• Mengyi Li, The Pennsylvania State University - University Park

• Nikki G. Lobczowski, University of North Carolina - Chapel Hill

• Rachel Miriam Vriend Croninger, The Pennsylvania State University

• Rebekah F Duke, University of North Carolina - Chapel Hill