Publish Late, Publish Rarely!: Network Density and Group Performance in Scientific Communication

Abstract

It is almost a tautology that scientists should try to be as “connected” as they can. Yet studies have found abundant communication links in a network of inquirers to be detrimental to group performance. This chapter hypothesizes that these results are partly due to the effect of network “spamming”: as more links are added, a network becomes more sensitive to low-quality information. This study shows how network density can actually be positively correlated with group performance if inquirers agree (i) to communicate only information they judge to be highly certain and (ii) not to recycle information in the absence of new supporting evidence. The results raise doubt about the recommendation to “publish early, publish often.” As the scientific community becomes increasingly connected, scholars should be increasingly concerned with publishing only material they take to be practically certain and avoid recycling material for which there is no new supporting evidence.