Although a large number of adults in the United States are assumed to need literacy training, little agreement exists on that number. Adults who attend literacy instruction are assumed to acquire the skills, but little is known about program effectiveness. Another assumption is that adults should define their goals, but this does not ensure that they strive for the levels of literacy they need. The primary reasons for literacy testing, from which data are regularly accessed by policymakers, are as follows: national and state population assessments; individual skill assessments; placement procedures; and program evaluation. Tests of functional abilities favored for population surveys do not assess writing or mathematical ability and are incompatible with national surveys of elementary and secondary literacy abilities. Basic skills tests used for individual skill assessments report scores in grade-level equivalents, which are problematic for adults. Current placement testing methods are lengthy, and entering students' poor test-taking skills may lead to low reliability. Scores from program evaluations have relatively limited use for policymakers for a number of reasons: pretest scores tend to be artificially low; program evaluations often ignore writing and mathematics; and student goals do not always coincide with national standards. Grade-level reporting for test scores needs to be reconsidered. The materials and tasks approach would be easier to relate to home, work, and civic functioning; the cognitive processing approach would be easier to relate to elementary and secondary assessment and instruction. (Contains 16 references.) (YLB)