Very little program-level research on psychology graduate research training environments (RTEs) exists despite the RTE being a program-level (vs. student-level) construct. In the current study, the authors investigated correlates of programs’ RTEs, including characteristics of the students, the faculty, and the programs themselves. Proportional stratified sampling of American Psychological Association-accredited PhD programs in clinical, counseling, and school psychology yielded data from 40 programs. Students and faculty in those programs completed Internet questionnaires, and additional faculty and program characteristics were coded. Hierarchical linear modeling analyses revealed that programs with positive student-rated RTEs had faculty who also viewed the RTE as positive and had satisfying faculty–student relations. These programs also had students with higher levels of research interest and stronger advisory alliances, yet within-program student differences in the RTE explained more variance in these student characteristics than did between-programs differences. This research supports the value of examining the RTE at both the student and program levels, and it suggests that training interventions ought to occur at both levels. (PsycINFO Database Record (c) 2010 APA, all rights reserved)