Background and purpose: Data supporting the predictive validity of the medical school admission interview are mixed. This study tested the hypothesis that the admission interview is predictive of interpersonal interactions between medical students and standardized patients. Method: We determined correlations between admission interview scores and performance on a senior-year Clinical Practice Examination piloting US National Board of Medical Examiners stations. We also completed regression analyses controlling for undergraduate academic performance, gender and ethnicity. Outcome measures included an Interpersonal Skills score and a separate Overall Checklist performance score, completed by standardized patients. Results: The applicant interview had limited but statistically significant correlation with the Interpersonal Skills (r = 0.15; p less than 0.05) score. The applicant interview had a correlation of 0.13 with the Overall Checklist score (p = 0.056). In linear regression models, the applicant interview had limited but statistically significant correlations with the Interpersonal Skills score and the Overall Checklist score. Conclusion: As practiced at this medical school, the admission interview has limited predictive validity for future interactions with standardized patients. More comprehensive assessment of interpersonal skills during the medical school selection process will be needed in order to better select matriculants with desirable interpersonal skills.