Analyses of Compound Structures of Groups that Produce Intellectual Property

This paper focuses on collaborations between scientists and engineers and investigates their mutual benefits. More concretely, multi-layered networks separated into four scientific/technological areas are investigated. The areas are life sciences (Bio), nanotechnology/materials (Nano), information and telecommunications (IT), and environmental sciences (Env), and they are mentioned in the third science and technology basic plan issued by the Government of Japan. The networks were then analyzed by using p * models to find compound structures. Logistic regression analysis was conducted, and the compound structures were expressed by explanatory variables. In all four areas, joint authorship and joint application tend to overlap. A role interlocking structure is only found in Bio, and itmeans that a gatekeeper exists between scientific knowledge and technical knowledge. A transitivity structuremeans three-person groups emerge such that a central person publishes papers (or patents) with two other people, and the two other people publish the other outcomes, and patents (or papers). It is found that transitivity is generally not reversible. In Bio and Nano, there is no eminent difference in significance of the two different types of transitivity, but in IT and Env, segregations with a joint application expert and joint authorship support emerge more strongly than the other types of segregations.