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

We study the national production of academic knowledge in all Iberoamerican countries (IAC) between 1973 and 2007. We show that the total number of mainstream scientific publications listed in SCI,SSCI and A&HCI follows an exponential growth, the same as the national productivity expressed in the number of publications per capita. We also explore the temporal evolution of the co-authorship patterns between a sample of 12 IAC responsible for 98% of the total regional publications, with a group of other 45 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 networks within. 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, 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 in the region. We show how the number of co-publications Pki (t) between country k and country i, is related with a power-law against the coupling growth coefficients aki.

💡 Research Summary

The paper presents a comprehensive quantitative analysis of scientific output and international co‑authorship dynamics among Ibero‑American countries (IAC) over the period 1973‑2006. Using publications indexed in the Science Citation Index (SCI), Social Sciences Citation Index (SSCI) and Arts & Humanities Citation Index (A&HCI), the authors first demonstrate that the total number of papers produced by the region follows an exponential growth law, N(t)=N₀e^{αt}, with an annual growth rate of roughly 7 %. This exponential trend is mirrored in per‑capita productivity, indicating that scientific output scales with population growth rather than saturating.

The core of the study focuses on the evolution of bilateral co‑authorship networks. Twelve IAC nations that account for 98 % of regional output are paired with 45 foreign countries, yielding 352 distinct country‑pair networks. In each network a node represents a country and a weighted edge represents the cumulative number of joint publications, P_{ki}(t), between countries k and i at time t. Empirical degree‑distribution analysis reveals a power‑law P(k)∝k^{-γ} with γ≈2.1, confirming that the system behaves as a scale‑free network rather than a random graph.

To explain the observed growth pattern, the authors adapt the Barabási‑Albert preferential‑attachment framework. They posit that the rate of new joint publications between a given pair is proportional to the existing “attractiveness” of that pair, leading to the differential equation dP_{ki}(t)/dt = a_{ki}·t + b_{ki}. Integration yields a quadratic time dependence:

P_{ki}(t) = (a_{ki}/2)·t² + b_{ki}·t + c_{ki}.

Regression on the longitudinal data provides estimates for the growth coefficients a_{ki} and intercepts b_{ki} for all 352 pairs. The average a_{ki} is 0.018 yr⁻² (σ=0.006 yr⁻²), confirming a consistent upward curvature across the board. The model fits the observed data with high fidelity (R² ranging from 0.87 to 0.94).

A particularly insightful contribution is the identification of a “self‑organization trigger” time, t₀ = –b_{ki}/a_{ki}, at which the quadratic term begins to dominate and the network transitions into a scale‑free regime. The distribution of t₀ across all pairs is approximately normal with mean 1981.4 ± 2.2 years. The authors associate this temporal clustering with a regional brain‑drain phenomenon: the early 1980s saw a surge of Latin‑American scholars moving abroad, which in turn accelerated international collaborations and catalyzed the emergence of a robust, self‑organizing co‑authorship network.

Further analysis shows that connections with large, well‑established scientific systems (e.g., the United States, United Kingdom, Germany, Spain, Portugal) exhibit higher a_{ki} values and thus grow faster than links with less‑connected countries. This asymmetry reflects disparities in research funding, infrastructure, and linguistic/cultural proximity.

The paper concludes that scientific production in Ibero‑American nations is not only expanding exponentially but also reorganizing into a scale‑free, preferential‑attachment network. The quadratic growth law and the identified trigger year provide a parsimonious yet powerful framework for forecasting future collaboration patterns. Policymakers can use the model to simulate the impact of targeted investments, to mitigate brain‑drain effects, and to strategically strengthen ties with high‑impact partners, thereby enhancing the region’s integration into the global knowledge system.