Learners inevitably enter adult technical training classrooms—indeed, in all classrooms—with different levels of expertise on the subject matter. When the diversity of expertise is wide and the course makes use of small group problem solving, instructors have a choice about how to group learners: they may distribute learners with greater expertise among the groups so they can provide tutoring to their less experienced classmates, or instructors may choose to segregate learners to keep those with greater expertise from dominating problem solving. The literature base on ability grouping in small groups, the bulk of which has been conducted on K-12 learners, provides no one-size-fits-all solution, and no systematic research has been conducted on post-college adult learners in technical training situations nor on learners grouped together to solve relatively ill-structured problems. Therefore, the purpose of this research is to study the relationship between experience grouping and learning gains in small groups of adults learning to solve technical, relatively ill-structured problems.

Forty-six newly hired auditors beginning their careers by taking a mandatory course centered around relatively ill-structured problem solving were assigned to stratified random groups by experience level so that some groups featured similar levels of experience and other groups featured a relatively wide range of experience. No significant differences in learning gains or satisfaction were detected between heterogeneous and homogeneous groups. Analysis of qualitative data, however, indicated that small groups exhibited significantly different behaviors depending on whether they were asked to solve relatively well-structured or ill-structured problems. When groups were asked to solve relatively ill-structured problems, learners were three to five times as likely to exhibit behaviors associated with learning in problem-solving groups. Further, analysis of the qualitative data suggests with marginal significance that during relatively ill-structured problem solving there is a strong parabolic relationship between the spread of experience in the small groups and the percentage of positive collaborative problem-solving strategies: as experience spread rose, groups either sought a larger or smaller percentage of explanations, a finding that reinforces other findings in the literature indicating that the relationship between high variability of expertise in small groups and achievement is complex.