The agreement between the Educational Testing Service (ETS) and the Golden Rule Insurance Company of Illinois is interpreted as setting the general principles on which items must be selected to be included in a licensure test. These principles put a limit to the difficulty level of any item, and they also limit the size of the difference in difficulty between two groups for any item. Both of these principles are based on sound psychometric theory and practices. The application of these principles may reduce the true differences in ability in order to eliminate bias; in the present state of psychological measurement it is not possible to eliminate all bias in the test and detect all true differences in ability between two groups if the test is biased. The internal consistency of the test is positively related to the amount of bias in the test. Predictive validity, the most important consideration in any licensure test, increases with the elimination of items with aberrant differences in difficulty between groups. Content validity is more than just content relatedness and it is not affected to a greater extent by the agreement than by any other method of item selection. (Author)