The collection and compilation of occupational supply data from formal, organized training programs were investigated to identify instances of multiple counting and undercounting of labor supply. The study combined a literature review; interviews with 274 individuals in 10 states involved in data production and use at the state or local level; interviews with agency and institutional representatives at local, state, and federal levels; and observations in 60 local training institutions or agencies. Multiple counting was found to occur as a result of duplicate counting within an institution and between agencies. Undercounting resulted from failure to secure data on the output of proprietary schools, and from lack of reports on public schools' vocational programs not in the state plan and nonvocational programs producing trained individuals available for employment. (The report further describes the state reporting systems and the flow of data from the local institution to the federal level. Other findings are presented that deal with the human factors involved in the generation, collection, analysis, and use of occupational supply information. Recommendations are made for the improvement of supply data and the system. Appendixes include the literature review and an overview of federal laws and regulations pertaining to data collection and reporting.) (YLB)