`Social and Technical Integration in Data-Pooling in Physics' (draft)
A comparison of different models of collaborative data analysis. Technical integration is maximised whenever two or more laboratories combine their data in a raw state before much local analysis has been done. It is minimised when autonomous laboratories do a great deal of local processing and produce findings for other laboratories to confirm. The in-between level is found where autonomous laboratories process their data sufficiently to produce `event lists' for coincidence analysis by other laboratories. The more technical integration, the more safe any result but the chance of an autonomous first discovery is sacrificed. The decision to go for more rather than less technical integration is analogous to joining a betting syndicate. Social integration it is a matter of signing up to a joint agreement and high when once autonomous laboratories lose their identity and become a socially dense community. The LSC is an example of an organisation with relatively high social integration. Various kinds of scientific collaboration both within and without the field of gravitational wave detection are discussed. The trajectories of the interferometer projects and the resonant-bar collective are analysed in these terms. High or medium technical integration and low social integration can cause trouble.