In the preprint version of her chapter on the performance of Twitter as a metric of scholarly impact that scientific articles have, which Stefanie Haustein wrote for the Handbook of Quantitative Science and Technology Research, a discussion on the validity of altmetrics is broached.
A Blog Article by Pablo Markin.
Guest-authoring a post, published on June 12, 2018, for the Altmetric Blog, Stefanie Haustein, an information science scholar from the University of Ottawa and Université du Québec à Montréal, Canada, has drawn attention to the mixed findings on the connection between Twitter mentions and citation counts of recently published articles. While social media, such as Facebook, can be assumed to contribute to the visibility of scientific research results, the collection of essays on Internet-based indicators for the impact of science edited by Wolfgang Glänzel, Henk Moed, Ulrich Schmoch and Mike Thelwall, to be published later in 2018, incidentally opens the discussion on the degree to which altmetrics can be helpful for the assessment of article-level impact.
This can be related to a relatively minor degree, estimated to range from 10% to 15%, to which Twitter is taken up by scholars. Moreover, as is characteristic of other social networks, Twitter use continues to be informal, rather than professional, for the majority of its contributors and audience, which can contribute to the nascence of its utilization for increasing the visibility of scholarly articles. Thus, whereas according to the non-representative, mixed-method study of Jason Priem and Kaithlin Light Costello published in October 2010, in which 28 scholars took part and 46,515 tweets were analyzed, scientists tend to make an indirect use of Twitter for article citation, based on the quantitative empirical findings of Stefanie Haustein, Rodrigo Costas and Vincent Larivière published on March 17, 2015, Twitter is the most important social media platform for spreading the word about academic publications.
According to Haustein, Costas and Larivière’s findings from their analysis of 1.3 million academic articles published in 2012, as indexed by the Web of Science, circa 22% of recently appearing papers can be expected to receive at least one mention on Twitter, while only about 5%, 2% and 1% of articles are likely to be shared or mentioned on Facebook, blogs or Google+ respectively. Though these findings contrast with approximately 67% of recent articles that have received at least one citation between 2012 and 2015. However, this could be due to the nature of social media mention counts and citation metrics which have been found to apparently favor more articles involving extensive scholar collaboration and bibliographies than those with lower levels of collaborations and citations.
Moreover, social media sharing has been found to be higher for editorial, shorter and news-related items, while citation metrics tend to be higher for longer academic papers than for shorter ones and items favored by social media, such as Twitter. In other words, Twitter-based altmetrics may be not necessarily strongly associated with citation metrics, due to in-built biases that social media platforms have, such as for articles published in Social Sciences and Humanities that are likely to receive a greater social media echo than papers in other academic fields.
Thus, altmetrics may have limits to their applicability, due to different factors that drive social media sharing frequency as opposed to citation counts in scholarly literature.
By Pablo Markin
Featured Image Credits: MSL Social Media Event, August 3, 2012 | © Courtesy of NASA Goddard Space Flight Center (GSFC)/Wade Sisler/Flickr.