The Iowa Tests of Basic Skills were administered to over 600 black and white students in grades six through nine, to determine if the test showed bias against minorities. Outliers were identified from test results. Outliers are items which differ from the central core of test items because they fall outside the range expected from a random sampling error. The tests revealed that some items were so difficult for black and white students that the difference in those item scores was essentially random. When these items were removed from the study, there remained 21 items which favored blacks and 16 items which favored whites. However, it is argued that the presence of outliers in distributions of differences in any item statistic does not constitute evidence of item bias, and that the presence of outliers is inevitable whenever groups other than random are compared. A detailed examination of the distributions of differences in item statistics for evidence of clusters which indicate the operation of factors not included in the ability or achievement domain being measured is recommended. In addition, outliers are seen as reflections of educational and cultural influences which should be brought to the attention of teachers and administrators when making educational decisions. (Author/JAC)