Collaboration networks in big science: The ATLAS experiment at CERN

Nowadays big scientific experiments require large organizations and hundreds of researchers who participate from several institutions. An interesting, yet rarely studied aspect of this new kind of scientific enterprise is the internal collaboration between the members of the participating institutions. Here we assess this matter in one of the most well-known examples of big science: the ATLAS experiment at CERN. Applying different network analysis techniques to data from internal CERN databases, we have identified several collaboration patterns in the experiment. We observe, on the one hand, the high level of collaboration between the institutions represented in ATLAS, higher than the average in the field of physics, and we identify the key institutions in the collaboration network. On the other hand, we notice that the collaboration network does not follow a scale-free or power-law model, contrary to what happens in other studied collaboration networks in physics and other areas. Finally, we observe that geographic distance between two institutions does not seem to affect the probability of establishing collaboration relationships, in contrast also to what happens in other kinds of collaboration networks.