Description: Peter Salovey and John Mayer (1990; Mayer, Salovey, & Caruso, 2000) suggested that a new kind of intelligence - 'emotional intelligence' (EI) - gives us awareness of our own and other people's feelings. Daniel Goleman (1995) was first to develop valid measures of emotional intelligence and to explore its significance. He scientifically showed that emotional and social factors are important (1995; 1998a). Goleman (1995) first agreed with Salovey and Mayer's (1990) five domains of emotional intelligence, but later his thinking about the dimensions of emotional intelligence, and their accompanying competencies, has evolved and simplified into four domains with eighteen competencies (Goleman, Boyatzis & McKee, 2002, 253-256). The theory as formulated by Salovey and Mayer (1990; Mayer & Salovey, 1997) framed EI within a model of intelligence. Goleman's model formulates EI in terms of a theory of performance (1998b). Goleman argues (2001) that an EI-based theory of performance has direct applicability to the domain of work and organizational effectiveness, particularly predicting excellence in jobs of all kinds, from sales to leadership. Methodology: In this paper we present an 18-item self-rating Likert-scale Leadership Competencies and Characteristics questionnaire (LCCQ) that operationalizes Goleman's et al. (2002) four domains of emotional intelligence. Our goal is to study with an empirical sample the construct validity of the four-domain model of emotional leadership : (1) self-awareness (three items), (2) self-management (six items), (3) social awareness (three items) and (4) relationship management (six items). The sample consists of 312 (52.0% of the sample population) adult employees of a Finnish medium size industrial automation company. Respondents' age mean was 34.6 years (SD = 10.9). The variable structure was examined with Bayesian dependency modeling (Myllymäki, Silander, Tirri & Uronen, 2002) as it allows ordinal indicators and is able to detect both linear and nonlinear dependencies (Congdon, 2001). Conclusions: The results showed that all eighteen competencies were selected for the most probable model. This finding indicates that the theoretical structure of EI is present at least in component level in this empirical domain. Further, the strengths of dependencies between the competencies were found to be equally strong indicating a well-matched contribution to the model. The clustering of eighteen competencies was further studied to see if the theoretical model of four domains (Goleman et al., 2002) is present in this sample. The visual inspection of the Bayesian dependency network showed that all four EI domains were present in the model derived from the empirical sample. However, the model included the fifth domain that was a mixture of self-awareness (self-confidence) and self-management (achievement and initiative) domains.