SCREENING, PLACEMENT, DIAGNOSTIC, AND EVALUATION DECISIONS FREQUENTLY ARE BASED ON THE RESULTS OF ACHIEVEMENT TESTS SUCH AS THOSE EXAMINED HERE. THEREFORE, IT IS IMPORTANT TO ESTABLISH THAT TESTS ARE EQUALLY RELIABLE MEASURES FOR BOTH LEARNING DISABLED AND NONLEARNING-DISABLED CHILDREN. IF THEY ARE NOT, ANY USE OF THEM FOR SCREENING OR ASSESSMENT MUST BE SUSPECT. DATA WERE COLLECTED FROM A LARGER STUDY CONDUCTED BY THE NATIONAL CENTER FOR STATE COURTS AND THE ASSOCIATION FOR CHILDREN WITH LEARNING DISABILITIES. PARTICIPATING YOUNGSTERS WERE FROM ALL SOCIOECONOMIC LEVELS AND A NUMBER OF RACIAL CATEGORIES; THE AVERAGE AGE WAS 14.2 YEARS. OF THE 1,699 PARTICIPANTS, 693 WERE OFFICIALLY ADJUDICATED DELINQUENT. THE EDUCATIONAL TESTING SERVICE WAS RETAINED TO DEVELOP AND CARRY OUT THE LEARNING DISABILITIES DIAGNOSTIC PROCEDURE. DIAGNOSTICIANS DETERMINED THAT 23 PERCENT OF THE TOTAL SAMPLE WAS LEARNING DISABLED. THE KEYMATH TEST, CONSISTING OF 14 SUBJECTS, AND THE WOODCOCK READING MASTERY TEST, CONSISTING OF 5 SUBJECTS, WERE ADMINISTERED TO BOTH GROUPS. THE RELIABILITY OF THE TESTS IS COMPARABLE FOR THE SAMPLE. FOR THE LEARNING DISABLED GROUP, THE RELIABILITIES FOR BOTH THE WOODCOCK AND KEYMATH WERE .88 AND .94, RESPECTIVELY. THE RELIABILITIES WERE SLIGHTLY HIGHER FOR THE NONLEARNING-DISABLED GROUP, OR .92 AND .97, RESPECTIVELY. IT IS CONCLUDED THAT BOTH TESTS CAN BE USED WITH CONFIDENCE FOR THE SCREENING AND DIAGNOSIS OF LEARNING DISABLED CHILDREN. REFERENCES, A FOOTNOTE, AND TABULAR INFORMATION ARE INCLUDED IN THE ANALYSIS. (LWM)