The use of heuristic judgments is prevalent in organizations and negatively impacts accurate employee assessments. To minimize the negative impact of heuristic judgments (i.e., anchoring and adjustment), we aim to improve rating accuracy by restructuring frame-of-reference (FOR) training. We conducted five studies (N = 1,143) using different samples (three including participants with hiring experience), training environments (onsite and online), and rating contexts (evaluations of sales representatives, teachers, contract negotiation specialists, and retail store managers). Across the five studies, the average improvement in rating accuracy was at least twice as large for restructured FOR (vs. control) training as it was for typical FOR (vs. control) training; the difference in rating accuracy between restructured and typical FOR training was statistically significant. Furthermore, minimizing the anchoring effect rather than increasing opportunities for rating adjustments improved rating accuracy (Study 4). Finally, restructured FOR training achieved higher criterion validity (i.e., a higher strength of the association between ratings regarding a target and the target's objective performance) than did typical FOR training (Studies 3 and 5). We discuss implications for improving the effectiveness of diverse training programs and the accuracy of judgments in organizations.