Abstract: It is estimated that 60% of teams fail to meet their performance objectives, which presents an important and challenging research topic for the academic community. Emotional intelligence is a concept that includes understanding and managing one's own emotions as well as those of others in interpersonal relationships. Research has shown that team emotional intelligence plays an important role in a team's success in achieving goals. However, the extant research mainly examines individuals-based or team as a whole-based team emotional intelligence and there is no research result investigating team emotional intelligence based on the exchange of emotional intelligence behavior among members of the team. To address an important theoretical and empirical gap in the literature, we employed an approach of integrating team-member exchange theory and social network perspective to examine the emotional intelligence within the team at the dyadic level (two individuals in a one-on-one relationship, i.e., actor-target). Specifically, we focused on dyadic emotional intelligence as the smallest unit, which then is aggregated to create a team-level emotional intelligence network structure. We conducted an empirical study of the network composed of the emotional intelligence behavior exchange between individuals in one to one interactions. The research contributes to the literature in three aspects.
First, a concept of dyad-based team emotional intelligence is proposed. We argue that the emotional intelligence behavior exchange at the dyadic level in the team can be summarized as the following four aspects: increasing positive energy, increasing negative energy, decreasing positive energy, and decreasing negative energy. Furthermore, we propose the idea of measuring team emotional intelligence by using the total amount of above emotional intelligence behavior exchanges at the dyadic level in the team (i.e., the density of positive behavior and the density of negative behavior) and the distribution pattern of those exchanges (i.e., the dispersion of positive behavior and the dispersion of negative behavior).
Second, a process model of team emotional intelligence emergence based on the interplay between the positive emotional experience of individuals and the positive emotional atmosphere of the team is proposed. Specifically, on the one hand, we propose that the emotional intelligence ability as a resource input might influence the individual and team's emotional intelligence behavior expressions in the context of the team through the mediation of individual positive emotional experience at the individual-level and the positive emotional atmosphere at the team-level, separately. On the other hand, on the vertical interface, in addition to both the initial state of emotional intelligence ability as a resource and the resulting state of individual emotional intelligence behavior expressions vertically input from the individual to the team, the positive emotional atmosphere of the team activated by team emotional intelligence would top-down affect the positive emotional experience of the individual generated by emotional intelligence ability, and the positive emotional experience of the individual would also bottom-up affect the positive emotional atmosphere of the team. In this way, the initial individual emotional intelligence develops into a dyad-based team emotional intelligence (i.e., the total amount and the distribution pattern of emotional intelligence behavior exchanges at the dyadic level) through the bottom-up process.
Third, the mechanisms through which team emotional intelligence impacts team outcomes (i.e., team task performance and team viability) are proposed. By developing an integrated model, we propose that the dyad-based team emotional intelligence might influence team outcomes by shaping two team processes: one is the team cognitive process (i.e., the ability of team knowledge integration), and the other is the emotional process of the team (i.e., collective affective commitment). Meanwhile, the shaping of the above team processes would be affected by the moderating variable of team task interdependence.

Key words: dyadic level, interpersonal relationship, social network, team emotional intelligence, team-member exchange