This dissertation aims to provide a better understanding of the technology licensing practices of academic research institutions. The study identifies time durations in licensing and incorporates these into a model to evaluate licensing performance. Performance is measured by the efficiency of an institution's technology licensing process and efficiency changes over time, using Association of University Technology Managers annual survey data from 1991 to 2007. Organizational characteristics influencing the licensing performances of 46 U.S. research institutions also are explored. The study resulted in a new approach that integrates the identification of time lags in licensing, analysis of efficiency change, and exploration of the influence of organizational characteristics on efficiency change. A super-efficiency variable returns to scale data envelopment analysis (DEA) model was applied to the time-lag neutralized licensing data, to measure the efficiency of U.S. research institutions' licensing performance over time. The study also includes an innovative approach to resolving issues with the super-efficiency DEA model, including mathematical infeasibility and zero-data issues. The licensing mechanisms included in the study are disclosure, patent applications, patents issued, licenses and options executed, start-ups, and licensing income. The time duration from expenditure to licensing income, including all intermediating licensing processes, ranged from 2 to 27 years. The study identified the organizational characteristics related to licensing practice. Academic prestige and research quality are positively related to disclosure, patents granted, and start-up. The resources of a technology licensing office influences the number of licensing agreements, whereas licensing office experience has a positive relationship with start-ups. Increased licensing resources improve the efficiency of licensing practices, and a research institution with more dedicated licensing staff has improved licensing productivity. Private institutions improved their licensing practice more than public ones during the study period. On the other hand, institutions with a medical school demonstrated low efficiency. This dissertation fills a gap in the understanding of licensing practice and the organizational characteristics related to licensing performance. In addition, the study contributes to research methodology by providing a new approach to identifying time lags and improving the DEA method. The results, grounded in comprehensive observations over multiple time durations, provide an insight into the licensing practices of U.S. research institutions. The dissertation presents recommendations for research institutions based on the relationships identified among academic prestige, research intensity, organizational characteristics of the technology licensing office, and licensing performance.