Sunday, October 10, 2010

Collective action among scientists

Recently I got involved in various activities who all have in common to improve collective action among scientists, especially social scientists who work on collective action. This might be a bit awkward, but social scientists are trained to work on individual or small group projects and are not used to share their data, protocols and other research findings such that others can verify the results and build upon the findings of earlier investigators. As a consequence there is a lack of systematic accumulation of knowledge in many areas in the social sciences.
At a keynote address I gave at the World Conference on Social Simulation I addressed the problem that scholars did not document their work well, do not share the source code of their models and as a consequence it is often impossible to verify the results and build upon earlier work. As if the researchers say that you first need to the pain they experienced before you are allowed to extend their research. Within our projects related to http://www.openabm.org/ we develop cyberinfrastructure (such as a model archive) and incentives (such as a annual competition). We hope that this will stimulate some more collective action and sharing of research. The feedback was mixed. In general the audience was in favor of improved collective action, but not everybody perceive there is a problem and like to do their own thing as they did in the past without being bothered by others.
A similar initiative is in preparation for scholars who use experiments in research and teaching on collective action and the commons. Instead of developing our own software in each research group, a group of about 15 research teams like to collaborate on sharing software and protocols and contribute to common development of cyberinfrastructure for the broader community.
On the one hand (social) science is not a place-based activity anymore and on the other hand we increasingly use more comprehensive tools that cannot be developed and maintained by single research groups alone. Therefore we need to experiment with research networks that enable groups to share research findings and collaborate on common infrastructure.