Many of us at the ISU Library are fans of altmetrics. We hold workshops on the topic and love to show people just how cool altmetrics are. But there’s a new problem on the altmetrics horizon: lack of transparency.
One of the reasons I rallied behind altmetrics was because unlike most other research performance metrics (such as JIF, H-index calculations, etc.) I could easily access all of the data and come to my own conclusions about what it means. Altmetrics also has a lot of value-added data in that many of the data points (shares, tweets, blog posts, news coverage, etc.) are commentary or conversational – you can see what people and organizations are saying about your research, not just that they cited it – and that kind of information is often more valuable than sheer numbers.
So, why am I worried then? Well it’s because of a fairly recent change one of the major altmetrics players has made…
The company Altmetric recently started limiting access to the complete data sets for each publication’s “Altmetric score”. In the past I was able to click the “twitter” tab of my detailed report and view all 43 tweets that mentioned my paper – including my own. Now, I can only view the four “most recent” tweets and am greeted with the message:
Intrigued, and a little irritated, I clicked the message and found out that if I wanted access to all 43 tweets 1) the journal I published in needed to be an Altmetric subscriber or 2) my University needed to be an Altmetric subscriber. Now, before I go any further I feel like I should make it clear that I understand the pressures that must have prompted this change at Altmetric; I still like their product; I am not advocating a boycott or anything drastic but I have concerns and frustrations as both an author as a librarian/teacher.
Before I discus my concerns and frustrations with these changes I should outline what being an Altmetric subscriber means. From what I can tell having an institutional subscription lets you browse and view Altmetric’s full data sets for authors affiliated with your institution. In short, it gives the subscribing institution gets access to a customizeable Altmetric data dashboard and also gives all of the staff and students access to their (full) Altmetric data. There’s also a note on how being a subscriber may let you better integrate Altmetric badges in institutional repositories but little detail on if you get anything extra beyond the current free badge. I hope you don’t because if subscribing institutions get some sort of data collection (not just data access) benefit then the whole system will be broken.
Publisher subscriptions on the other hand get a different set of benefits. While an institutional membership unlocks Altmetric’s data around authors a publisher subscription unlocks the data around papers and grants access to this data to both the publisher’s employees and to the public. This article from Nov 2014 in Science is a good example – if you click the Twitter tab you get access to all 184 tweets – not just the “most recent.” Publisher subscribers also get a dashboard and more control over how to share and present Altmetric data with their users… yada yada.
Let me be frank: I think that restricting the benefits of an institutional subscription to only employees and students is pretty crappy. Altmetrics can be a great recruiting tool and I don’t think there’s a good reason that an institutional subscriber shouldn’t be able to have a public dashboard with their membership. If the people who are paying for this data want to show it off then let them.
Weakening the system
As I pointed out earlier one of the largest factors in altmetrics (and Altmetric’s) favor is the breadth and depth of the data that users and authors get to explore about their publications. When this data is limited only to those who can afford it the system is weakened. Limiting access to the data that goes into the calculation of an Altmetric Score makes the system more susceptible to gaming. After all, it’s pretty cheap to pay for tweets , blog posts, etc. which promote a link/article. If the data is closed it will be much harder for reviewers and peers to see if there is something fishy going on. Altmetric Scores could become meaningless quickly because of potential noise in the system. Sure, Altmetric can/does probably run checks against gaming tactics but it’s always easier to see it for yourself than to trust a third party.
Who gets left out?
My biggest concern with these changes is a social justice concern: Altmetric’s pricing model is creating a market where only large publishers and top research universities will get access to altmetric data. There is currently no individual subscription option. Altmetric is offering librarians a a “free librarian account” which I’m guessing is some sort of marketing tool rather than something truly useful (once again, there is VERY little information on the website).
Here’s a short and incomplete list of people and groups that could be excluded with this pricing and distribution model:
- Authors who publish in open access journals hosted by institutional repositories (i.e. not owned by a publisher).
- Small, societal publishers and their authors/readers (some of these folks are already left out as they are not publishing online but that’s a problem that is squarely in their court).
- Most researchers who work for a local, state, or federal government (i.e. usually operating off a very lean operating budget).
- Researcher who work for non-profits (ditto).
- Small industry researchers (ditto ditto).
- Authors, institutions, and publishers located in low-income countries.
- Faculty and administrators at liberal arts and community colleges.
- Developers who use Altmetric as a data source (such as ImpactStory)
Reciprocal sharing… or not
I don’t know the details of how or what Altmetric harvests to make their data sets – I’d love to know more – but they are harvesting some “open” data such as Twitter, Google+, Facebook, etc. and it just feels wrong to limit access to data that users have made open and free for reuse. Yes, I know that Altmetric isn’t really limiting access to it – I could go out and do searches to find it myself – but these data sources also represent one of the largest, and possibly easiest, sources of data that Altmetric harvests. While I don’t think it’s likely, making this a “premium” feature could lead to some users limiting their public interactions on these platforms in order to avoid making data for Altmetric to reuse and make money off of.
In closing, I’m not thrilled by these changes but I understand them. Prior to these changes I had wondered exactly how Altmetric was staying in business and if it was able to turn a profit. Restricting access to their full data provides more motivation for publishers and institutions to subscribe which in turn will keep Altmetric in business but this move comes with some unfortunate trade-offs. I welcome feedback on my critique – especially if I’ve made any false assumptions or mistakes.