In this study, Lutz Bornmann and Robin Haunschild discuss the relationship between altmetrics and the quality of scientific publications (as measured by the post-publication peer-review system of F1000Prime assessments). One intention of the authors is to test whether altmetrics should really be used in research evaluation practice. A necessary requirement for this use is that the metrics are related to scientific quality. However, without this relationship, there is the risk that bad or fraudulent research is rated high, because it simply received attention in society.
The results of the analysis indicate that citation-based metrics and readership counts are significantly more related to quality than tweets and the Altmetric Attention Score. Therefore, these findings question the use of Twitter counts and the Altmetric Attention Score for research evaluation purposes but at the same time indicate the potential use of Mendeley reader counts.