Research on collaborative testing has shown that it decreases test anxiety, increases learning and critical thinking skills, and allows students to practice collaboration and teamwork. However, it has most often been used as a second test following traditional individual testing. This quasi-experimental study compared two models of collaborative testing in an introductory statistics course. Students who experienced collaborative testing as a pretest (n = 35) and as a posttest (n = 33) were compared with a control group (n = 61) that experienced traditional individual testing only. Grade book data and a survey administered at the end of the semester provided information on student attitudes and perceptions toward learning statistics and use of collaborative testing, and the impact of each model on student learning. The same instructor taught all students and administered the same tests. The authors found that group-first testing was favored by students and provided the greatest benefit for reducing anxiety while still supporting student learning.