The long-term dynamics of co-authorship scientific networks, Iberoamerican Countries (1973-2006)

We study the national production of academic knowledge in all Iberoamerican and Caribbean countries between 1973 and 2007. We show that the total number of mainstream scientific publications listed in the Science Citation Index (SCI), the Social Science Citation Index (SSCI) and Arts and Humanities Citation Index (A&HCI) follows an exponential growth, the same as the national productivity expressed in the number of publications per capita. During the last 35 years, Portugal shows the highest rate growth in both indicators. We also explore the temporal evolution of the co‐authorship patterns between a sample of 12 Iberoamerican countries responsible for 97.95% of the total regional publications between 1973 and 2006, with a group of other 45 different nations. We show that the scientific co‐authorship among countries follows a power‐law and behaves as a self‐organizing scale‐free network, where each country appears as a node and each co‐publication as a link. We develop a mathematical model to study the temporal evolution of co‐authorship networks, based on a preferential attachment strategy and we show that the number of co‐publications among countries growths quadraticly against time. We empirically determine the quadratic growth constants for 352 different co‐authorship networks within the period 1973‐2006. We corroborate that the connectivity of regional countries with larger scientific networks is growing faster than with other less connected countries. We determine the dates, t 0 , at which the co‐authorship connectivities trigger the self‐organizing scale‐free network for each of the 352 cases. We find that the last follows a normal distribution around year 1981.4 ±2.2 and we connect the last effect with a brain‐drainage process generated during the previous decade. We show how the number of co‐publications (t) between country k and country i, is related with a power‐law against the coupling growth coefficients. We develop a methodology to use the empirically determined growth constants for each co‐authorship network to predict changes in the relative intensity of cooperation among countries. We finally discuss the consequences of our findings for the science and technology cooperation policies.