This paper provides school personnel with a practical outline for selecting an appropriate standardized achievement battery by describing the test selection procedures developed by a large metropolitan school system. The system needed a new achievement test battery which would reflect recent curriculum changes and provide for grade level testing. Evaluation concerns included norms, reliability, scaling, content, format, costs, item bias, and accountability to the public. During the first phase of the selection process, 2,000 students in grades 1-8 took each of five standardized reading achievement tests. Reliability, item difficulty, percent of students scoring at the floor or ceiling of each test, and the relationship between raw scores and grade-equivalent scores for each test were examined. After the first phase, two tests remained. The second phase emphasized how well the tests measured what was being taught, especially reading comprehension. The old test and both suggested new tests were administered to 35,000 students in 100 schools. Information on students' reading progress was studied in relation to grade equivalent scores. The final selected test had moderate length and difficulty, good reliability, and low percentage of scores either below the chance level or above 80 percent. (GDC)