Providing students with assignments that focus on critical thinking is an important part of their scientific and intellectual development. However, as class sizes increase, so does the grading burden, prohibiting many faculty from incorporating critical thinking assignments in the classroom. In an effort to continue to provide our students with meaningful critical thinking exercises, we implemented a novel group-centered, problem-based testing scheme. We wanted to assess how performing critical thinking problem sets as group work compares to performing the sets as individual work, in terms of student attitudes and learning outcomes. During two semesters of our recombinant DNA course, students had the same lecture material and similar assessments. In the Fall semester, student learning was assessed by two collaborative take-home exams, followed immediately by individual, closed-book in-class exams on the same content, as well as a final cumulative exam. Student teams on the take-home exams were instructorassigned, and each team turned in one collaborative exam. In the Spring semester, the control group of students were required to turn in their own individual take-home exams, followed by the in-class exams and final cumulative exam. For the majority of students, learning outcomes were met, regardless of whether they worked in teams. In addition, collaborative learning was favorably received by students and grading was reduced for instructors. These data suggest that group-centered, problem-based learning is a useful model for achievement of student learning outcomes in courses where it would be infeasible to provide feedback on individual critical thinking assignments due to grading volume.