Response to Patterns of Contact and Communication in Scientific Research Collaboration

Summary

This paper discussed the relationship between physical proximity and collaborative relationships in research. Specifically this meant, if your office was very close to another researcher, how would this impact your likeliness to work with that research on a paper. To examine this the researchers looked at a "large research lab" with 500 PhD and M.Sc. researchers. The writers had access to data on how close researchers were and who wrote papers with who. They calculated the total number of "pairs" who shared a corridor or a floor and those who worked on a different floor and in different buildings. They found that 10.3% of the pairs who shared a corridor did work together. The problem with this analysis is that people may work with people who have similar interests and specialities, and the results might be explained by this rather then proximity. To adjust for this they considered organizational distance (i.e. being in the same department) did not entirely explain this. They then suggest that proximity does have an important impact on collaboration because communication over large distances is more expensive then local communication. They then speculate on the impact this will have for technology designed to support group work, suggesting that technologies that lower this costs (e.g. ubiquitous video) will be beneficial.

Reflections

This paper can be criticized on two fronts; issues with the analysis and issues with it's recommendations. When the authors try to disentangle the effect of physical distance and organization distance, the best they can do is to analyze how many collaborations are done within a department and without it. This ignores subspecialties, consider that while the Software Engineering professors and (some) Embedded Systems professors work on the same floor, we would not expect to see as many collaborations between them as with other SENG profs. It would have been interesting to see how collaborations were impacted by the self described "research areas" of each professor.

A more serious criticism is about the recommendations made by the authors. Since the paper is so old (published in 1988) we can consider their recommendations in the light of new technological developments. Realtime, low-cost, video connections are now widely available via Skype and other technologies. In spite of this, co-located work is still difficult. Evidence for this is all the difficulties in getting distributed software teams to work well. Missing in this empirical analysis is understanding how work takes place and the social factors involved. Even high quality video does not have the interpersonal aspects of "being there". And users are uncomfortable with always-on video conferencing.