In 1976 the National Institute of Education established the Research and Development Utilization (RDU) Program. This project was intended to apply research and development products to local school problems; to develop a problem-solving and product selection process usable by local schools and to organize a linkage system making national, state, and other external resources available to local school personnel. Seven projects serving over 300 school locations in 20 states were operated under the RDU program from 1976 to 1979. Preliminary findings from a study of the program suggested that engagement in RDU program activities would promote school improvement. Several variables were examined to determine which were important to the impact of the RDU program at the school level. Such variables included characteristics of the research and development products chosen, degree of modification of externally developed products, strategies selected for problem-solving, amount of coordination with outside experts, and techniques used during product implementation. These variables were found to be more powerful predictors of success at implementing innovations at the local school level than was school "readiness" as measured by previous experience with innovative efforts, federally funded programs, or use of externally developed research and development products. (Author/PGD)