This post is a recap of preliminary research that I and Altmetric software developer Abheer Kolhatkar shared at the 3:AM conference in September 2016.
Regionally relevant social networks can tell us a lot about how research is shared and discussed in certain countries. In this post, we describe surprising differences between how Russian-authored and British-authored research is discussed on VKontakte (a major Russian social network) and Facebook (a global social network).
VKontakte is Russia’s #1 social network
A screenshot of a public VK profile
VKontakte (also known simply as “VK”) is the number one visited site in Russia and the twentieth most visited site in the world as of September 2016, according to Alexa. Nearly four in ten VK users are located outside of Russia.
According to Baran & Stock (2015), “VKontakte is perceived by its users as more useful than Facebook, is much more trustworthy, and more enjoyable to use.” Given VK’s perceived usefulness and trustworthiness, as well as its popularity in the Russian-speaking world, we wondered if it was possible that the site was used to discuss Russian-authored research more often than other social networks, and also if research authors’ nationalities affected the rates at which research was shared. So, we decided to dig into the data.
Mentions of research on VK and Facebook
Abheer set up a bot that scraped public-facing VKontakte sites in a similar manner to how Altmetric already scrapes Facebook for mentions. For roughly one month, the bot searched VK for links to a small list of twenty-two “whitelisted” domains (places where we know research is published online), shared at any time. Then, we automatically followed links gathered by the bot and, on resulting webpages, we searched the page’s metadata for DOIs (as a shorthand for understanding if the page contained a research output).
Here’s what we found.
3.6% of research mentioned on VK is authored by Russian-affiliated researchers
Overall, we found 341 links to the twenty-two whitelisted domains mentioned on public VK profiles. Of those, 276 were unique outputs with persistent identifiers that we could use to search the Web of Science to learn the location of authors.
Of the items mentioned on VK, we found that 3.6% (n=10) were authored by Russian researchers, and approximately 8% (n=22) were authored by researchers based in the UK.
0.3% of research mentioned on Facebook is authored by Russian-affiliated researchers
For comparison’s sake, we decided to look and see what discussions were like on Facebook, since Alperin found that, for Latin American research, the amount of discussion was roughly the same as for research from “the Global North”.
There were around 56,000 unique mentions of the twenty-two whitelisted domains that had occurred at any time, found on public Facebook pages during our most recent data refresh in early 2016. We took a random sample of outputs with at least one mention (n=2,303, to have a large enough set to draw conclusions from) and again used Web of Science to discover the authors’ institutional affiliation.
We found that the proportion of UK-authored research mentioned on Facebook was roughly 9% (n=205). However, the proportion of Russian-authored research was much smaller than for Russian research shared on VK–0.3% (n=8), an entire order of magnitude less (and a statistically significant difference)!
Here’s another way of understanding the disparity between mentions on each social network, based on authors’ locations:
Far fewer Russian-authored research articles were mentioned on Facebook than on VK (0.3% vs. 3.6%, respectively). One reason for this discrepancy might be that Russian researchers and/or the Russian-speaking general public are using VK more often than Facebook to discuss Russian-authored research.
Another reason might be that a majority of the domains whitelisted in this small study are from Western, English-language repositories and journals. It’s possible that if we had tracked more Russian domains, we would see a higher percentage of Russian-language works shared on both VK and Facebook.
UK-authored research was shared at roughly similar rates across the two platforms (though no lasting conclusions should be drawn from this fact, given the limitations explained below).
Coming back to our original question–is VK used to discuss Russian-authored research more often than Facebook?–the preliminary data we’ve collected suggests that it is, by a lot (at least across English-language repositories and journals). However, given the small sample sizes and limitations on data collection that we’ve described, these results should be interpreted carefully. More research is needed before we can make any generalizable claims.
What does this mean for altmetrics aggregators?
Given the potential disparity in mentions for Russian-authored research across Facebook and VK, adding regionally-relevant social networks, repositories, and journals to the suite of data that altmetrics aggregators collect would help balance coverage, making it much more representative of international scholarly communication patterns on the whole. In turn, that would improve the coverage and reliability of altmetrics research, much of which is done using altmetrics aggregators.