Inspired by the theoretical results on network formation by Bala and Goyal (2000) and
experimental evidence by Plott and Callander (2002) and Falk and Kosfeld (2003) we design
network formation experiments in discrete and continuous time. Particularly in the
continuous time experiment, we observe that many groups not only reach a strict Nash
network but also switch to other strict Nash networks during the course of time. Strict
Nash networks, therefore, prove to be a good indicator for predicting network formation.
In explaining our results, we highlight the role of the complexity of network structures,
of inequity aversion, and the players activity levels as most important determinants for
forming strict Nash networks.