This study examined the relationship of multiple-choice and free-response items contained on the College Board's Advanced Placement Computer Science (APCS) examination. Confirmatory factor analysis was used to test the fit of a two-factor model where each item format marked its own factor. Results showed a single-factor solution to fit the data best in each of two random-half samples. This finding might be accounted for by several mechanisms, including overlap in the specific processes assessed by the multiple-choice and free-response items, and the limited opportunity for skill differentiation afforded by the year-long APCS course. (39pp.)