The general emphasis of Project LONGSTEP was on the identification of changes in student achievement that occur as a result of exposure to intensive educational innovation. During the course of the project, questions were raised concerning (1) the relevance of the standardized achievement test utilized in the project to the stated objectives of the educational treatments included in the participating educational programs and (2) the extent to which the findings may have been influenced by the particular instrument chosen to measure cognitive achievement. To provide at least partial answers to these questions, a study of the arithmetic items contained in the test and their relevance to the educational objectives of the arithmetic treatments encompassed by the project was conducted. Specifically, two research questions were addressed: (1) Were the particular skills necessary to answer correctly the arithmetic items on the Comprehensive Tests of Basic Skills (CTBS) actually incorporated into the curriculum objectives of arithmetic treatments? and (2) To the extent that arithmetic treatments differed from one another in the relevance of the CTBS test items to their curriculum objectives, what effect did this difference have on test performance? "Data Collection Instruments and Guidelines" developed for Project LONGSTEP referenced in Vol. I, Chapter II, Section C, will be accessioned TM 005 987 in RIEMAY77. (RC)