The aim of this article is to assess and compare three classification statistical techniques--logistic regression, discriminant analysis and classification trees--to identify the personality characteristics associated with the risk of suffering from ischemic cardiovascular acute episodes (ICAE). The sample comprised 313 participants, men and women, aged from 36 to 80. Participants were divided into two groups: a clinical group of patients (n = 143) who were diagnosed as suffering from ICAE, and a control group (n = 170). Both groups were equated in gender, age, socio-economic and educational level. In view of the comparative study of the analytical procedures, we recommend classification trees as the best choice, as it was the most accurate for the individuals in the clinical group, a simple data analysis and a meaningful clinical interpretation. The predictive validity analysis of the MCMI-II allowed the construction of a reduced version made up of 9 personality scales from the 22 scales in the original version. Thus, we could identify the patients with a higher probability of suffering from ICAE, and additionally, generate an empirical model comprising seven and five personality profiles associated, respectively, with the increase and the decrease of the probability of suffering from ICAE.