Multiple-choice exams, while widely used, are necessarily imprecise due to the contribution of the final student score due to guessing. This past year at the United States Naval Academy the construction and grading scheme for the department-wide general chemistry multiple-choice exams were revised with the goal of decreasing the contribution of the score due to guessing. A variable number of response choices were used based on question type, and students were encouraged not to guess by giving partial credit for answers left blank. Results of the study are presented.