Working in small groups to solve problems is an instructional strategy that allows university students in science, technology, engineering, and mathematics disciplines the opportunity to practice interpersonal and professional skills while gaining and applying discipline-specific content knowledge. Previous research indicates that not all group work prompts result in the same experiences for students. In this study we posed two types of prompts (guided and open) to undergraduate engineering students in a statics course as they participated in group work projects. We measured student discourse, student performance, and perceptions of group work. We found that guided prompts were associated with higher-level discourse and higher performance (project scores) than open prompts. Students engaged in guided prompts were more likely to discuss distribution of labour and design/calculation details of their projects than when students responded to open prompts. We posit that guided prompts, which more clearly articulate expectations of students, help students determine how to divide tasks amongst themselves and, subsequently, jump to higher levels of discourse.