Scientific collaboration has become a major issue in science policy. T
he tremendous growth of collaboration among nations and research insti
tutions witnessed during the last twenty years is a function of the in
ternal dynamics of science as well as science policy initiatives. The
need to survey and follow up the collaboration issue calls for statist
ical indicators sensitive enough to reveal the structure and change of
collaborative networks. In this context, bibliometric analysis of co-
authored scientific articles is one promising approach. This paper dis
cusses the relationship between collaboration and co-authorship, the n
ature of bibliometric data, and exemplifies how they can be refined an
d used to analyse various aspects of collaboration.