In this article we present an indicator - Probabilistic Partnership Index (PPI) - for use in measuring scientific linkages. This indicator is based on the Monte-Carlo simulation which provides a standard model to each network established in collaboration between two countries. Any relationship that occurs within a (whole) network can be projected to a standard model respectively and thus PPI is useful in examining individual networks within complex exchanges. We investigate inter-sectoral cooperation between France and Japan for the period of 1981-2004, by classifying every research unit appearing in the data set by its sector. We examine international collaborative patterns, domestic collaborative patterns and multilateral relationships established within the French-Japanese cooperation. We also compare PPI with the classic collaborative linkage indexes - Jaccard Index, Salton-Ochiai Index and Probabilistic Affinity Index - in order to describe the specificity of the new indicator. Our hope is that PPI will prove to be a useful and complementary tool for the analysis of international collaboration.