College teachers often use norm-referenced classroom tests which are too easy for distinguishing adequately among levels of student achievement, yet they are reluctant to adopt more difficult tests. We explored the basis for current practices concerning test difficulty through informal interviews and questionnaires completed by faculty members and by students. Overall, the responses suggest that many faculty and students view the percentage-correct score as an absolute measure, even when tests are clearly norm-referenced. Further, faculty often claimed that students would not tolerate more difficult tests even if grade distributions remained the same. Finally, we found specific support for more difficult tests in only a few of a selection of 19 measurement textbooks we reviewed. In this context we developed a booklet ["College Achievement Testing: A Self-Instructional Interlude," included as appendix A] and a pamphlet ["The Trouble with Headaches," included as appendix B] intended to convince students that norm-referenced tests should be reasonably difficult. Evaluation of these materials, however, showed them to be largely ineffective for this purpose. Apparently, beliefs in this area are not easily changed, which warrants ongoing efforts to inform teachers and students of appropriate practice regarding test difficulty. (Author)