Literacy policy has often developed independently of other social and employment programs. As a consequence, literacy tends to become an end unto itself, and assessment is directed more toward academic, archival ends than toward policy evaluation. Many justifications given for large-scale literacy programs are not based upon empirical data. Foremost among these justifications are the projected skills mismatch between the job market and new entrants to the work force. Additionally, the target audience for literacy training is often greatly inflated. The basis of most recent assessments of adult literacy is functional literacy, but there are concerns that it fails to provide a basis for predicting abilities in new literacy contexts and it lacks a basis for selecting specific assessment tasks. Rather than abandoning functional literacy, an attempt should be made to support it both theoretically and empirically. Questions have also been raised about the effectiveness of different literacy programs and limited experience in measuring literacy program performance. The need for an extensive database to support literacy policy becomes imperative. Perhaps the most important lesson to be learned from national literacy surveys is that they have limitations for policy formation and evaluation. What is needed in most developed countries is more focused information on the groups who need assistance and the effectiveness of programs instituted to assist them. (Contains 41 references.) (YLB)