Four factors influencing effectiveness in email communication networks

This paper develops a method to analyse email communication networks in terms of the effectiveness of emails to elicit a reply. Four factors are considered: the email’s sender (sender effect), the email’s recipients (recipient effect), specific properties of each sender-recipient dyad (dyad effect), and specific properties of the email (stimulus effect). A multilevel model is developed and applied to an email communication dataset. The fitted model suggests that dyad and stimulus effects are more important sources of variability in reply rates than sender or recipient effects. Moreover, the correlation between sender and recipient effects is found positive when taken in isolation, but vanishes when accounting for the dyadic effect. The strength of this method lies in the general and systematic way it provides to evaluate organization members and their ties with respect to their influence on others to reply. From a socio-theoretical perspective it operationalizes a co-evolution process that operates not only at the meso-level (nodes and dyads), but also between the meso-level and the micro-level (transactions).