To compare maxillomandibular relationships for individuals with nonsyndromic complete unilateral cleft lip and palate using the Goslon Yardstick for dental models. Retrospective cohort study. Five cleft palate centers in North America. A total of 169 subjects with repaired complete unilateral cleft lip and palate who were consecutively treated at the five centers. Ethics approval was obtained. A total of 169 dental models of patients between 6 and 12 years old with complete unilateral cleft lip and palate were assessed using the Goslon Yardstick. Weighted kappa statistics were used to assess intrarater and interrater reliabilities; whereas, analysis of variance and Tukey-Kramer analysis was used to compare the Goslon scores. Significance levels were set at p < .05. Intrarater and interrater reliabilities were very good for model ratings. One center that incorporated primary alveolar bone grafting showed especially poor Goslon scores that were significantly poorer than the remaining centers. The surgery protocols used by the other four centers did not include primary alveolar bone grafting but involved a number of different lip and palate closure techniques. Using the Goslon Yardstick assumptions, the center with the best scores would be expected to require end-stage maxillary advancement orthognathic surgery in 20% of its patients; whereas, the center with the worst scores would be likely to require this surgery in 66% of its patients. The Goslon Yardstick proved capable of discriminating among the centers' dental arch relationships. Possible explanations for the differences are discussed.