Network Embeddedness and Solidarity at Work – Andreas

1. Short summary of results of the original research (extended summary see below) At this moment (1 September 2001), about two third of the research period are completed. In accordance with the work plan, the focus has been on theoretical model building in the first two years and will shift to empirical testing and publication of research results in the third year. Articles based on the research have already been accepted for publication in international journals or are still under review (see bijlage B). Further articles are under preparation. The numbers in the following refer to the list of publications in bijlage B. The substantive part of the research resulted in further theoretical elaboration of the counter intuitive and innovative hypothesis of a “weakness of strong ties” (see original proposal). New predictions have been derived about the conditions under which a cohesive social network may fail to foster or even undermine group solidarity. Article [6] addresses the condition of task uncertainty. The model developed in the article predicts that under increasing task uncertainty, group members will no longer enforce norms of solidary behaviour, in order to protect their personal relationships from disruption. This is a new form of the weakness of strong ties hypothesis. A follow up article [10] deals with the effect of the content of social relationships. The theoretical analysis predicts that the previously derived negative effect of task uncertainty may disappear when social relationships mainly comprise monitoring (mutual observation) rather than social rewards (e.g. friendships). The methodological part of the research compares various microfoundations of solidary behaviour. In addition to its substantive contribution, article [6] also contains the first methodological result that leads beyond previous work. The weakness of strong ties can not only be derived from a backward-looking micromodel (adaptive learning) but also with a forward-looking approach (rational anticipation). Article [2] then focuses on solidary behaviour in long term interactions between backward-looking actors. The paper shows that the mechanism of “reinforcement learning” can explain solidary behaviour and the study identifies actors’ aspiration level as decisive condition for this. Article [7] integrates and compares two specifications of reinforcement learning that are prominent in the literature and highlights previously unnoticed differences between the models. Further research in [1] addresses the microassumption of individual risk preferences. The innovative prediction is derived that risk aversion may have non-linear effects on the level of solidarity in a social network. Beyond this, the methodological research has generated three further articles about the general theoretical framework ([3]) and computer simulations with cellular automata ([4] and [5]).