In validating tests for course placement in college, the criterion variable is usually defined in terms of the grade earned in a particular standard course. This study considers the issue of interpreting incomplete (I) and withdrawal (W) grades either as unsuccessful outcomes or as missing data. The effects of either type of interpretation on placement indices and optimum cutoff scores were studied using data from four two-year colleges participating in a pilot study of American College Testing (ACT) Course Placement Service. Courses included were mathematics and English/reading courses. Assessment of Student Skills for Entry and Transfer (ASSET) test scores were used as predictor variables. Results show that interpreting I and W grades as unsuccessful outcomes, rather than missing data, generally results in lower conditional probabilities of success, higher optimal cutoff scores, and higher estimated delta accuracy rates. The manner in which I and W grades are interpreted should depend on an institution's policy or philosophy. Ideally, grades of I should be changed to grades of A through F or S/U (satisfactory/unsatisfactory) before being included in analyses of course placement criteria, and only students who receive grades of W because of academic reasons should be classified as unsuccessful. If course placement accuracy indices were determined using these criteria, then the resulting optimum cutoff scores would be more accurate and appropriate. Four tables present analysis results. (Contains 3 references.) (Author/SLD)