Network Density and Group Competence in Scientific Communication

It is almost a tautology to propose that communication is a good thing and that one should, as a scientist, try to be as “connected” as one can. And yet studies in various disciplines, from social psychology to economics, persistently undermine this notion. These studies have found a lot of communication links in a network of inquirers to be detrimental to group competence in a wide range of cases. We hypothesize that these results are partly due to the effect of network “spamming”: as more links are added so that the density of the network increases, it becomes more sensitive to low quality information. We show, by means of computer simulations in a Bayesian framework, that network density can be positively correlated with group competence if inquirers agree to post only high quality (relevant) information in the network, but also that these positive results are sensitive to what we mean by “group competence” more precisely.