A review of current procedures for selecting students for graduate programs in psychology reveals a number of problems due to predictor, criterion, and methodological variables. The idea is advanced that with the multiplicity of procedures used, plus the number of students applying to graduate programs in psychology, the effort on the part of individual selection committees to differentiate students on a number of relevant variables is both time consuming and not sufficiently powerful to be able to make meaningful statements about each applicant's relative potential. A tenative proposal to establish a centralized registry of psychology graduate students is made. A centralized registry might be more efficient and economical and also would establish a research data pool which could provide a more reliable means for identifying and matriculating graduate students who will make significant contributions to psychology. (Author/EVH)