The moderating effect of altmetrics on the correlations between single and multi-faceted university ranking systems: the case of THE and QS vs. Nature Index and Leiden

University ranking systems use various single and multi-faceted methodologies. Despite being efficient and less biased, the former fails to cover all academic performance dimensions, requiring solutions to improve its effectiveness. Previous studies found universities’ ranks to be partly correlated to their social presence and activities via their official accounts. However, altmetrics have a comparatively more diversified and all-inclusive nature. Moreover, altmetrics are assumed to reflect various impact types and therefore represent different academic performance dimensions. This study attempted to discover if the altmetrics aggregated at the university level can bridge the gap between single and multi-faceted rankings. Focusing on Leiden and Nature Index as single-faceted, and Times Higher Education and Quacquarelli Symonds as multi-faceted rankings, it explored a sample of the universities jointly ranked by the systems in 2017 and 2020. Their overall scores in Times Higher Education and Quacquarelli Symonds were regressed against their Leiden crown indicator (PP top 10%), Article-Weighted Fractional Count in Nature Index, Altmetric Attention Score, tweets, and Mendeley readership. According to the results, the universities’ scores in Leiden and Nature Index predicted theirs in Quacquarelli Symonds (33.5% and 21.4%, respectively) and Times Higher Education (63.7% and 33.4%, respectively). Altmetric Attention Score, tweets, and Mendeley readership boosted the predictions, implying their ability to reflect academic performances. However, they differed in their effects’ strengths, importance, and directions, which may be resulted from their differences in the impact realms and values for different social sections, which are not necessarily proportional to the corresponding dimensions’ weights in the rankings.