作者: Mike Thelwall , Stefanie Haustein , Vincent Larivière , Cassidy R. Sugimoto
DOI: 10.1371/JOURNAL.PONE.0064841
关键词:
摘要: Altmetric measurements derived from the social web are increasingly advocated and used as early indicators of article impact usefulness. Nevertheless, there is a lack systematic scientific evidence that altmetrics valid proxies either or utility although few case studies have reported medium correlations between specific citation rates for individual journals fields. To fill this gap, study compares 11 with Web Science citations 76 to 208,739 PubMed articles at least one altmetric mention in each up 1,891 per metric. It also introduces simple sign test overcome biases caused by different usage windows. Statistically significant associations were found higher metric scores positive all cases sufficient (Twitter, Facebook wall posts, research highlights, blogs, mainstream media forums) except perhaps Google+ posts. Evidence was insufficient LinkedIn, Pinterest, question answer sites, Reddit, no conclusions should be drawn about zero strength any correlation citations. comparisons values published times, even within same year, can remove reverse association so publishers scientometricians consider effect time when using rank articles. Finally, coverage Twitter seems low it not clear if they prevalent enough useful practice.