How Altmetric Measures a Paper’s Worth
Like impact factor, Altmetric uses number of citations in its calculation. Unlike impact factor, altmetric goes further, including, for example, the number of times a paper is viewed and bookmarked. I think this is a fine addition to an algorithm. I publish few research papers nowadays, but I read a great many, download a large number, and bookmark a few. The importance to which I assign papers increases in that proportion. Some way of weighting these actions—bookmark>download>view—would reflect the opinion of vastly more scientists than the relatively few that cite the paper in publications.
Is Altmetric a Popularity Contest?
Although I like the inclusion of view/save/bookmark I am leery of some of Altmetric other measures of importance, particularly those that monitor and count the mentions of a paper on the social media platforms such as Facebook, Twitter, and science blogs. Here we enter the territory of “popularized science” and there is a danger that Altmetric can become a mere popularity contest. I’m all in favor of popularizing science, but some types of science don’t lend themselves to popularization, the hard sciences in particular. Evolution is easier to explain (and more interesting) to the general public than quantum mechanics. So, the buzz on social media platforms will tend to be heavily weighted towards certain branches of science, while overlooking others that are equally important.
The Wikipedia Issue
Altmetric also takes into account mentions of articles on Wikipedia pages, which is sure to raise some hackles. I myself cringe whenever I see Wikipedia cited in a publication. (See an earlier blog of mine for more on this.) But so long as Altmetric weights Wikipedia in the same manner as social media citations, I have no objection.
Overall, Altmetric and similar ranking programs are interesting and potentially valuable measures of an article’s perceived importance within a certain discipline. In short, we must compare apples to apples. Papers on quantum mechanics must be ranked along with other papers on the same subject, not with those on psychology or global warming.