Steps involved in the item analysis and scaling of the 1990 edition of Forms A and B of the Nevada High School Proficiency Examinations (NHSPEs) are described. Pilot tests of Forms A and B of the 47-item reading and 45-item mathematics tests were each administered to random samples of more than 600 eleventh-grade students. A computer program was developed to calculate the classical item statistics and tabulate the proportion of items answered correctly by students in each quintile. These proportions provided information similar to that attainable from item characteristic curves in item response theory (IRT) analysis. Analysis determined that items of Forms A and B of the mathematics test were acceptable. One item on Form A of the reading test was deleted, and nine other items on Forms A and B of the reading test were rewritten to improve their ability to discriminate. The BILOG computer program was then used to confirm these results, scaling the forms using the three-parameter IRT model. These scalings were used to set the cutting scores. Cutting scores were compared to results from the same students from the previous edition of the NHSPEs. There was good agreement between results of the two tests. The Nevada Department of Education is recommending that the passing standard be raised with the expectation that about 16% of students taking the NHSPEs the first time will fail. Two tables present study findings. (SLD)