It is argued that publishers of achievement tests, especially those who publish tests intended for use in many parts of the United States, are for the most part not in a position to obtain any decent evidence about the consequences of the uses that are made of their tests. What responsibilities and actions publishers can reasonably be expected to take, with respect to the consequences of test use, is explored. The uses of tests vary by teacher, school, district, state, and over time, especially the time between norming and test use, and no direct mechanism exists for obtaining evidence of the many consequences of test use. Publishers should undertake to study the matter for each of their tests to the extent possible and they should try to persuade academic researchers to study the matter objectively. Because there are so many tests and so many uses, it will take a large-scale cooperative effort to produce any generalizable evidence about the consequences of using nationally normed tests whatever their formats. This leadership cannot come from test publishers alone, but they should play a substantial role in the undertaking as they work with professional organizations to bring about many studies of test use. (Contains 12 references.) (SLD)