Abstract

Introduction: Scientists have long searched for the best way to evaluate scientific performance and have come up with numerous indices. Probably, the most famous index is the h-index, a metric that has been used widely in science metrics ever since it appeared. In this study, I evaluated the potential of the new metric Research Interest Score created by ResearchGate. Methods: I analyzed the different metric indices for 88 most cited researchers at the University of Sarajevo. In particular, the number of citations and h-indexes were taken from the Web of Science, Scopus, and Google Scholar databases for these researchers and were correlated with their Research Interest Scores. As a measure of correlation, Pearson correlation coefficients was used. In addition, I conducted a regression analysis to examine how scientometric indices from the Web of Science, Scopus, and Google Scholar predict the Research Interest Score. Results: The Research Interest Score was most strongly correlated with the number of Google Scholar Citations, followed by the Web of Science citations. Interestingly, the Research Interest Score was not highly correlated with any of the h-indexes. The regression model was statistically significant and explained 90% of the variance in the Research Interest Scores predicted by the Web of Science, Scopus, and Google Scholar metric indices. The strongest predictor of the Research Interest Score was the number of Google Scholar citations. Conclusions: The Research Interest Score, a new scientometric measure created by ResearchGate, certainly has a potential to be used as a valid measure of scientific impact.