The effect of collaborative testing on junior high school students' scores was investigated. Twenty-two students in grades 7, 8, and 9 were given curriculum-based tests in their pre-algebra classes. Some tests were taken individually in the traditional manner, and others were taken collaboratively by students in pairs over a 10-week period. Collaborative and individual scores were compared, and collaborative test scores for 1990 were compared with individual test scores for 1989. A questionnaire was administered to the subjects to determine their reactions to collaborative testing. Questionnaire results indicate that all students preferred the collaborative testing. Test scores showed a mean improvement of 9% over the 10 weeks, while the previous year's students, tested traditionally, had dropped 5% during the same time period. These findings indicate that: collaborative testing offers an effective way of improving achievement in certain classes, and student attitudes toward classes can be positively affected by collaborative testing. The study questionnaire and two tables of test results are appended. A 42-item list of references is included. (Author/SLD)