The instructional effectiveness of commercially prepared test preparation materials was studied using Riverside Publishing Company's "Improving Test-Taking Skills" materials. The study further investigated differential effects of test-taking instruction as a result of student characteristics. In the first year of the study (1986-87), performance on standardized achievement tests of 182 fourth graders receiving an average of 12 hours and 212 fifth graders receiving an average of 9.25 hours of instruction with the Riverside materials was matched with that of students receiving test-taking instruction without the Riverside materials (223 fourth-graders, and 215 fifth-graders). In the replication year, students in nine schools received about 10 hours of Riverside instruction. Samples used in the second year included 129 fourth-graders and 92 fifth-graders receiving the Riverside materials, and 148 fourth-graders and 109 fifth-graders not receiving the Riverside materials. In the first year of the study, fourth-grade students receiving formal instruction in test taking did increase their scores, although teacher-made or commercial materials performed comparably well. In the second year, the Riverside method instruction appeared to have resulted in improved test scores only in mathematics for the fourth grade. No beneficial effects of test-taking instruction were found in the fifth grade in either year. No clear-cut patterns of differential effects for sex, socioeconomic status, or ethnicity were apparent for grade 4; results for grade 5 were similar. The policy implications of these findings are discussed. Two tables contain study results. (SLD)