Medical schools must use selection methods that validly measure applicants' noncognitive qualities, but primary-care (PC) schools have a particular need. This study correlated entering students' personality and values scores with their professed interest in PC. 93 medical students completed instruments assessing personality (16PF & PSP), values, and specialty interests. Resulting scores were used to predict students' primary-care interest (PC) or non-interest (NPC) using logistic regression analysis. About 80% were correctly categorized into PC or NPC based on these variables. The strongest noncognitive predictor was Warmth, measured by either the standardized, psychological instrument (16PF) or the non-standardized, self-assessment survey (PSP). Dutifulness and Trusting Nature significantly added to PC prediction, if measured by the 16PF but not the PSP. When personality scores (16PF/PSP) were not entered into the analysis, service and prestige were significant predictors. These noncognitive variables were significant even after accounting for gender, race, and physician parents. Candidates describing themselves as warm and attentive to others during interviews or in their personal statements may be expressing an accurate self-assessment, because Warmth scores from both self-report and standardized measures similarly predicted PC interest. Predictive validity of these noncognitives throughout medical school should be examined. (Contains 1 table.)