Study design and data analysis The study was both quantitative and cross-sectional, with an associative and comparative strategy design [36]. Two types of data analysis were carried out. For the variables ‘Level of national and international scientific collaboration’ ,
‘Efforts made by the host country to achieve collaboration’ and ‘Efforts made by national scientific institutions to pro- mote collaboration’, an inferential analysis was carried out with the aim of comparing the average ranks of formed groups according to four different areas of knowledge: social and legal sciences and humanities, sciences, health sciences, and engineering and architecture. This inferential analysis was made with a non-parametric statistic (Kruskal–Wallis test) due to the failure of the normality assumption or because of the ordinal nature of the data. In the case of the variables ‘Types of collab- oration maintained with scientific institutions in Spain’ and ‘Profile of Spanish researchers abroad’ a descriptive analysis was used, including graphics, frequencies and percentage analyses. The analyses were conducted using SSPS 17.0 software.