Two-stage exams--where students complete part one of an exam closed book and independently and part two is completed open book and independently (two-stage independent, or TS-I) or collaboratively (two-stage collaborative, or TS-C)--provide a means to include collaborative learning in summative assessments. Collaborative learning has been shown to have positive benefits, including increased student engagement and learning. To try to improve student learning, as measured by improvement in exam scores, two sections of introductory geology were taught using two-stage exams. It was hypothesized that class scores would be higher for semesters using two-stage exams--whether part two was TS-C or TS-I--than for semesters using traditional (T) exams. The median attendance rate was approximately 67% for all testing methods and was significantly greater when exams were TS-C (69%) rather than TS-I (53%). The class score was significantly greater during semesters when exams were TS-C (81%) but was not significantly different between T and TS-I semesters. To assess individual student learning over time, part one of the first exam and part one of the comprehensive final exam were compared. Across the F and D grade ranges, improvement on individual exam scores was significantly greater for the TS-C semester than for the TS-I and T semesters. Student learning, as measured by individual exam scores, improved due to the use of TS-C exams. The improvement in class scores due to the collaborative portion of two-stage exams was independent of increased attendance rates, greater for the lower-achieving students, and not observable if part two of the exam was completed as a take-home exam (TS-I).