Two studies examined aspects of the validity of self-report and performance-based measures of emotional intelligence (EI) relevant to their use in personnel selection. In Study 1, structural equation modeling indicated that a two-factor model with separate factors for the two types of EI measures fit better than a one-factor solution. The performance-based EI factor was more related to cognitive ability (R=.38) than personality (R=.26), whereas the self-report EI factor was more related to personality (R=.85) than cognitive ability (R=.09). Although the performance-based EI factor correlated more strongly with job performance (ρ=.24) than did that of the self-report (ρ=.05), it provided little incremental validity beyond cognitive ability and conscientiousness. In Study 2, participants were asked to complete the measures as if applying for a job, and mean scores were then compared with those of Study 1. Results indicated that self-report EI measures were more vulnerable to distortion than were the performance-based measures. Implications for the assessment of EI in personnel selection contexts are discussed.