CommonSense Blog

What Altmetrics Can Do, Where Else Might It Lead?

By Robert Newcomer | Dec 14, 2015

As described in a previous post, Altmetrics is an emerging measure of audience engagement for an individual piece of media as it traverses the internet. While this attention can be representative of positive or negative engagement, in either instance, the potential for keen, real-time insights is enormous.

Provided, of course, that these insights translate into something actionable.

Scholars are still divided on the relative value of “likes” or “tweets”, which have only recently insinuated themselves into the vernacular of academia. The intuitive relevance to some seems eccentric or even unbecoming to others.

What can these data actually provide? It is a fair question. One very straightforward application is to take a look at which scientific papers are achieving the highest scores.1

A look back at November of 2015 reveals a mixed bag, with topics ranging from a study measuring the association between religion and altruism in children (attention no doubt driven by the counterintuitive finding of a negative association), to a New England Journal of Medicine paper describing the case of an unfortunate soul who seemingly developed cancer from a tapeworm.2,3

It is not unlikely that you are already aware of these stories. The behavioral study was picked up by nearly 70 different news outlets and tweeted over 1,000 times. The frightening tapeworm was discussed on NPR.
pooh2
And what of the manuscript with the highest recorded Altmetric score to date? A lofty score of 11,152 is awarded to a whimsical manuscript from the Canadian Medical Journal entitled, “Pathology in the Hundred Acre Wood.” Using DSM-IV criteria, Pooh receives a provisional diagnosis of OCD. Piglet, generalized anxiety. Tigger, clearly, ADHD.4

It’s a cute paper that clearly resonates across a number of audiences. But again, what do these high scores actually mean, and more importantly, how might they be replicated?

We first must define the scores themselves. The sources used for the aggregation of these scores typically include measurements across five categories:5

• Usage: Views and downloads
• Captures: Bookmarks, favorites, followers, or subscribers
• Mentions: Blogs, Wikipedia, comments, reviews
• Social Media: Tweets, Google+, Facebook, etc.
• Citations: Web of Science, Scopus, CrossRef, PubMed

If you have encountered the image below, an increasingly common representation referred to as a donut, then you’ve already seen altmetrics in action. The colors correspond to specific media types, and the number at the center is a weighted summation, the higher the better. Anything in excess of 20 is generally considered to be pretty good.
donut A skyscraping score of 11,152 seems akin to lightning in a bottle, though I do note here one common trend that seems relevant. The high-ranking manuscripts I’ve discussed here are all freely available, as opposed to something hidden behind a Draconian paywall. It seems fair to speculate that such availability will certainly be a contributing factor to the ultimate reach of any given manuscript. (Journal editors take note.)

At W2O we seek to not only understand new data, but also to learn how research is best promoted and shared, and how these activities serve to foster interactions between thought-leaders and relevant stakeholders.

Altmetrics provides a tool to identify topics that resonate, and also to determine where audiences discover and share these data. This suggests a possible role for altmetrics beyond that of a simple yardstick after the fact. Can altmetrics be adapted into a tool to inform dissemination strategies for new data, reaching audiences that have already demonstrated their preferences and shown us what works?

It’s something to think about.

References:
1. www.altmetrics.org
2. Decety J, Cowell JM, Lee K, et al. Curr Biol. 2015;25:2951-5.
3. Muehlenbachs A, Bhatnagar J, Agudelo CA, et al. N Engl J Med. 2015;373:1845-52.
4. Shea SE, Gordon K, Hawkins A, et al. CMAJ. 2000;163:1557-9.
5. Melero R. Biochem Med (Zagreb). 2015:25:152-60.