Costs and Benefits of Problem Solving in Small Groups

Principal Investigator: 
Project Overview
Background & Purpose: 

The purpose of this research project is to systematically investigate how group size contributes to gains in learning (by examining collaborative problem solving vs. individual problem solving contexts) and the process and the outcomes of individual and collaborative learning activity. The central goals of this project are: the identification of specific aspects of collaborative contexts that are critical for learning gains; the design and dissemination of effective collaborative activities; and the attempt to generalize across topics and domains. These goals are intended to allow for this project to have broad impact on educational practice.


The higher education classroom is the research setting for this project.

Research Design: 

This study is both cross-sectional and comparative and is designed to generate correlational [quasi-experimental] and causal [experimental, quasi-experimental] evidence. It includes both an intervention [Assigning participants to work alone, in pairs or in triads] and a comparison condition. One comparison for group performance is a nominal group performance measure. All participants also run individually on a transfer problem set.

Original data are collected through assessments of learning or achievement tests observation [videography] and problem sets that require scientific and mathematical reasoning.

Group problem solving performance is analyzed as a function of group composition and size. Individual performance on transfer problems can be directly compared. A supplemental correlational analysis involves examining the aspects of interactions that support the most effective problem solving and transfer. A final experimental intervention attempts to support effective group interactions in all groups, by assigning roles and giving specific guidelines for collaboration.


Two lab experiments revealed that triads are more likely to take advantage of the unique skills/expertise of group members than are dyads. Discourse coding revealed consistent effects across two studies that more evaluative behaviors emerge in the discourse of triads, and this elicits more knowledge sharing. The result is that students who interact with at least one skilled member in a triad perform better on both group problem solving tasks and individual transfer tasks. Dyads do not learn from each other in the same way, nor do they generally engage in mutual evaluation of reasoning. The same overall learning result was replicated in the context of an intact Research Methods course in Psychology. Current studies are attempting to support these effective discourse behaviors in both dyads and triads, with role assignment and explicit instruction to evaluate propositions.

Publications & Presentations: 

Wiley, J. & Jensen, M. (2006). When three heads are better than twoProceedings of the Annual Conference of the Cognitive Science Society.

Wiley, J. & Bailey, J. (2006) Effects of Collaboration and Argumentation on Learning from Web Pages. In A.M. O'Donnell, C. E. Hmelo-Silver & G. Erkens (Eds.) Collaborative Learning, Reasoning, And Technology (pp. 297-321). Hillsdale, NJ: Erlbaum.

Wiley, J. & Jolly, C. (2003) When two heads are better than one expertProceedings of the Twenty-Fifth Annual Conference of the Cognitive Science Society.

Other Products: 

This project expects to generate scripts for effective small group scientific reasoning/problem solving activities.