This paper describes the efforts of the Talent Identification Program (TIP) of North Carolina to relate American College Testing (ACT) Assessment and Scholastic Aptitude Test (SAT) scores of academically talented seventh graders. The study was also designed to interpret the results in the context of "concordance" studies that have been done with other groups of students. The target population for the central study consisted of students who applied to the TIP during the 1986-87 talent search by taking the SAT in December 1986 or January 1987. A stratified random sample of 2,042 applicants was invited to take the ACT Assessment on the April 1987 test date. Results from the 611 students who responded to the invitation indicate that the ACT Composite and SAT Total scores are nearly as strongly related among academically talented seventh graders as they are among college-bound high school students. However, the English Usage/Verbal and Mathematics Usage/Mathematics components of the two test batteries measure different characteristics of academically talented young students, and these characteristics are developed to relatively different extents than they are among high school students. The results illustrate the tendency of conversion tables for non-parallel tests to differ among various subgroups. (TJH)