This paper has the purpose of establishing the variables that explain the behavior of ResearchGate for the Top100 Latin American universities positioned in Webometrics database for January 2017. For this purpose, a search was carried out to get information about postgraduate courses and professors at the institutional websites and social networks, obtaining documents registered in Google Scholar. For the data analysis, the econometric technique of ordinary least squares was applied, a cross-sectional study for the year 2017 was conducted, and the individuals studied were the first 100 Latin American universities, obtaining a coefficient of determination of 73.82%. The results show that the most significant variables are the number of programs, the number of teacher's profiles registered in Google Scholar, the number of subscribers to the institutional YouTube channel, and the GDP per capita of the university origin country. Variables such as (i) number of undergraduate programs, (ii) number of scientific journals; (iii) number of documents found under the university domain; (iv) H-index of the 1st profile of researcher at the university; (vi) number of members of the institution; (v) SIR Scimago ranking of Higher Education Institutions; (vi) number of tweets published in the institutional account; (vii) number of followers in the Twitter institutional account; (vii) number of "likes" given to the institutional count, were not significant.