The question of the status of emotional intellect as a scientific category of psychology is raised. The existing approaches to its measurement and techniques used in national research practice are analysed. The divergence is emphasised of the data obtained from self-report (self-assessment) of the level of emotional intellect formation and formation of its components with the data of test samples, aimed at assessing the relevance of skills. A critical assessment of the most common and adaptable at the moment methods operating with procedural indicators of emotional intelligence is given. The need to operationalise its components is noted. Based on the author’s model of emotional intellect the analysis of empirical data obtained on a sample of students is made. Using the methods of mathematical statistics it is proved that gender differences in self-reports are not consistent with differences of test samples in boys and girls, and to a large extent are influenced by gender stereotypes. The correlation analysis showed that the self-report and performance test data are not linked. The confirmation of the previously identified data on the autonomy of functioning of the components of processing boys’ emotional information and girls’ simultaneous strategies is received. Presented results exacerbate the conflict between the model of capabilities of and the mixed model of emotional intellect.