This paper compares descriptive information based on analyses of the pilot and live administrations of the Alabama High School Graduation Examination (AHSGE). The AHSGE, a product of decisions made in 1977 and 1984 by the Alabama State Board of Education, is composed of subject tests in reading, mathematics, and language. The pass score for each subject test was established using empirical and judgmental methods based on data from the initial pilot administration. Therefore, it is essential that the live administration data be consistent with the initial pilot data. The purpose of this study was to validate the test development procedure by comparing the difficulty level of common test items measuring competencies on the initial item pilot and on the live administration. The data were collected during 2 phases of pilot testing and 16 live administrations of the AHSGE. The initial item pilot tested 40 schools, with 560 students statewide for each of the 4 test forms. In 1983 form pilot tests, there were about 1,200 11th-grade students. Over 48,000 students participated in the first live administration. The stability of the difficulty level of common items was demonstrated, as was the consistency of the difficulty level of the parallel forms of the AHSGE from initial pilot through live administrations. Results support the test development procedures of the AHSGE. Fourteen tables present summary data, and an eight-item list of references is included. (SLD)