The purpose of this paper is to discuss a procedure for determining the degree of relationship of the many scores in the Tennessee Career Ladder Program to the final classification based upon this system. The program, now in its sixth year of implementation, places over 40,000 teachers and educational administrators within a career ladder hierarchy. Sources of information pertinent to the evaluation include teacher, principal, student, and peer evaluations. Data from data collection instruments, including classroom observation and interviews, account for 75% of a candidate's total score, while a consensus reached by evaluators accounts for the remaining 25%. The sample for this analysis of the program consists of all general education teachers evaluated for upper level positions on the ladder during the 1988-89 academic year. Sixty-six scores are calculated in this system. It appears that effect size can be used to indicate the degree to which scores in a complex personnel evaluation system differentiate among the levels earned. The use of a modified level in which that data were excluded from the classification scheme reduced the magnitude of the effect sizes while generally keeping different scores in approximately the same relative order. Four data tables and three bar graphs are provided. (TJH)