The National Reporting System project was designed to establish an outcome-based accountability system for the state-administered, federally funded adult education program. In the pilot test, volunteer states and their local programs compared the feasibility and costs of these three data collection models: (1) local programs collect all the student information and outcome measures, including measures by phone; (2) states conduct the follow-up measure survey by phone and then combine that data with other student data collected by local programs; and (3) state employment and human services agencies and higher education institutions combine their data with that from the state department of education. Findings for Model 1 (eight programs) indicated survey costs ranged from $500 to over $1500; the largest concerns were cost and staff availability; and programs reported general satisfaction with the educational functional level measures. The primary advantage was having the sampling universe at the program rather than state level. Disadvantages were high costs and time consumption. Findings for Model 2 (three states) indicated costs varied from $40.37 to $132.04 (significantly less than the local survey), and it alleviated the burden of data collection from local programs. The biggest drawback was that outcomes were not identified at the program level. Findings for Model 3 (four states) indicated barriers were confidentiality issues and development of measure definitions common across agencies; vastly different costs for developing and maintaining the database system; and a considerable investment in training and technical assistance at the local level was required to ensure valid, reliable data. (YLB)