This study tests the effects on student achievement and attitude of two methods of reviewing homework problems in mathematics. One review procedure involved solving and explaining each homework problem the day after an assignment was made. In the second procedure, only the problems that students requested to have reviewed were covered. Two Algebra I classes were involved in this research with one procedure used in each class. Upon the completion of a unit test, the homework review methods were alternated between the two classes. The study found that: (1) there was no significant difference in achievement between treatments for either class; (2) there was no significant difference between classes in achievement when using different review procedures; and (3) students generally preferred to review only homework problems they requested. The research suggests that the preferred method of reviewing should be some combination of the two methods described, with the difficulty of the topics and interests of the students used as the bases for the instructor's decision regarding review style choice. (MP)