The U.S. Department of Energy is sponsoring a research effort through Sandia National Laboratories and the University of Missouri-Rolla to test a correlation for the upper shelf energy (USE) values obtained from the impact testing of subsize Charpy V-notch specimens to those obtained from the testing of full-size samples. The program involves the impact testing of unirradiated and irradiated full-, half-, and third-size Charpy V-notch specimens. A correlation between subsize and full-size testing results has been developed in an earlier effort and tested on HT-9 material utilized in fusion reactor applications. To verify the applicability of the correlation on LWR materials, unirradiated and irradiated full-, half-, and third-size Charpy V-notch specimens of a commercial pressure vessel steel (ASTM A533 Grade B) will be tested. The correlation methodology is based on the partitioning of the USE into crack initiation and crack propagation energies. To accomplish this partition, both precracked and notched-only specimens will be used. Whereas the USE of notched-only specimens is the sum of both crack initiation and crack propagation energies, the USE of precracked specimens reflects only the crack propagation component. The difference in the USE of the two types of specimens represents a measure of the crack initiation energy. Normalizing the values of the crack initiation energy to the fracture volume of the sample produces similar values for the full-, half-, and third-size specimens. In addition, the ratios of the USE and the crack propagation energy are also in agreement for full-, half-, and third-size specimens. These two observations will be used to predict the USE of full-size specimens based on subsize USE data. This paper will provide details of the program and present results obtained from the application of the developed correlation methodology to the impact testing of the unirradiated full-, half-, and third-size A533 Grade B Charpy V-notch specimens.