Twenty-eight protocols of the Stanford-Binet Fourth Edition (SB:IV) obtained from graduate students were examined for scoring and clerical errors that contributed to the inaccuracy of test scores. Scoring of individual items was identified as the most error prone process, as evidenced by the fact that 96% of the protocols contained scoring errors. The most frequent scoring errors by subtest occurred on Comprehension, Vocabulary, Copying, Absurdities, and Verbal Relations. A relatively high occurrence of basal or ceiling errors (61%) was also found, with the Copying subtest being the most problematic for the establishment of both basal and ceiling levels. Clerical errors involving computation and coding were found in 32% of the protocols. Although the overall impact of examiner errors was fairly small in magnitude on the composite scores, 7% of the protocols produced a discrepancy of 6 points of more, sufficient to affect classification and placement decisions. (Contains 4 tables and 16 references.) (Author/SLD)