The College Board's SAT[R] Score Choice[TM] policy allows students to choose which set(s) of scores to send to colleges and universities to which they plan to apply. Based on data gathered before the implementation of that policy, the following study evaluated the predictive validity of the various sets of SAT scores. The value of five score sets for predicting first-year college grade-point average (FYGPA) was examined using a sample of 150,377 students from 110 institutions. These sets of scores were from students': (1) first administration, (2) latest administration, (3) section averages, (4) highest single administration, and (5) highest individual sections. The various score sets were found to have nearly equal predictive validity for the total group, by various institutional characteristics, and regardless of how often students tested. (Contains 6 tables and 2 endnotes.)