Towards the “cognitive management” of a research institute

Abstract

The purpose of this article is to describe a method for classifying the activity of large general-interest research centres. The system of classification should permit an evaluation of the interface between any given research project and any given economic goal. It should thus become a management tool.Working from a description of research activity in terms of elementary tasks, the method - called co-word analysis - consists in aggregating key words describing these tasks into larger “themes” of about 10 words. It is then possible to compare the research efforts of particular laboratories within the research centre in terms of a short, “natural”, variable list of themes. This classification can be compared with the current administrative structure, and can be used as a basis for the negotiation of change.It is then suggested that the themes attached to a specific subfile of research operations can be displayed on a two-dimensional diagram, with one axis being the “centrality” (the weight of external links) and the other being the “density” (the weight of internal links) of each theme. This diagram structures the file into core research activity (central and developed), general research context (central but not developed) and subfields (peripheral, developed or undeveloped). Being a representation of implicit strategies, this diagram can be used at many levels - from that of the choice of key words to the determination of research goals. It can be used for the negotiation of policy both within and outside of the research centre.