Rudman and colleagues (1980) deplored the paucity of descriptive information relative to teachers' test use patterns. The present study addresses the abundant prescriptive, and lack of descriptive information concerning teacher testing. A mailed survey procedure gathered testing practice information from elementary and secondary South Dakota school teachers (n=336) regarding: (1) testing context, (2) test construction, (3) test administration, (4) test analyses, and (5) test results. The survey indicated that teachers use a variety of testing techniques, but only teacher-made objective tests play a major evaluative role across all grade levels and curricular areas. There appear to be three important factors which influence teacher practice: time, expertise, and tools available for teachers' use. Nearly 20 percent of in-class time is devoted to test-related activities. This substantial time investment is a strong argument for skill in the practice of testing; however, most teachers have limited preparation in the area. Improved practices require changing the habits of teachers and educating them to overcome their lack of knowledge of sophisticated tools (e.g., calculators, microcomputers). Perhaps the most clear need is for a return to development of measurement techniques that will be appropriately used in the classroom. (Author/PN)