Exploring a pseudo-regression model of transnational cooperation in science

Abstract

This paper reports the results of an empirical study on the impact of three proximity measures: geographical distance, thematic distance and socio-economic distance among the set of 45 scientifically most advanced countries on their cooperation network. In network data, individuals (viz. countries) are linked to one another and the relationships are nested and embedded in groups, with the result that statistical assumptions of independence underlying ordinary least squares regression are systematically violated. Hence, we have used a non-parametric regression procedure, Quadratic Assignment Procedure (QAP), for regressing the matrix of transnational cooperation on the matrices of three proximity measures: geographic proximity, thematic proximity and socio-economic proximity. The results indicate that all the three proximity measures have the expected negative effect on transnational cooperation. Geographic proximity has greater impact than the other proximity measures.