This study investigated the test-retest and inter-rater reliability and the construct validity of the Social Skills Rating Scales — Teacher (SSRS-T) with an elementary school sample. The SSRS-T is an experimental version of a new series of multi-rater, multi-factor scales for rating the frequency and importance of children's social behavior. The results of this study indicated that the SSRS-T has high (r = .90) test-retest reliability over a 6-week period, is internally consistent (r = .96), and has moderate (r = .65) interrater reliability. The construct validity of the SSRS-T was established by comparisons with the Revised Behavior Problem Checklist and Teacher Ratings of Academic Performance. Data are reported on grade level differences in social skills, teachers' ratings of the importance of behaviors, and the relationships between behavior ratings and direct observations. A discussion of future research and the further developments of the SSRS concludes the article.