Bookmetrix: Demonstrating the power of books in academic research
An interview with Martijn Roelandse
Until recently, the evaluation of research output and impact has focused largely on journal articles. Much less attention was paid to the relative impact of research published in books prior to the launch of Bookmetrix in 2015. Developed in partnership with metrics provider Altmetric, the platform hosts data for more than 260,000 books and over 4.5 million chapters.
The vision for this platform was developed by former neuroscientist and now Head of Publishing Innovation at Springer Nature, Martijn Roelandse. Roelandse was working as Publishing Editor on the Springer Neuroscience list, when he had the idea to create both the metric and technology that would clearly show the impact of research published in books. Here he talks more about the rationale for developing Bookmetrix and newer ways of evaluating research impact.
How important is the role played by books in research dissemination today?
Books, like journal articles, are widely used by academics today. They’re cited in other works, discussed on Twitter, mentioned on news websites and in blogs, and above all, they’re heavily downloaded. Springer Nature titles published in 2016 saw average downloads of more than 5,000 per book, so there’s no doubt that this format is very much alive and kicking in academia.
The transition from print to online in book publishing over the past decade has generated bigger audiences and has also made measuring impact easier. Bookmetrix is trying to improve the transparency of research impact in this format.
How much are the boundaries of research evaluation expanding beyond citation metrics?
As blogs, social media, news and policy websites become more prominent sources of research, evaluation methods are expanding. The European Union, along with other funders and research organisations, are shifting the academic emphasis from publishing research quickly to sharing as much data as possible. To date there hasn’t been a global consensus on standard measures or benchmarks for these newer channels, but more funders and governments are actively thinking about the best ways of accessing, measuring and assessing all of this information.
What were your key motivations for developing a metric to evaluate the impact of books?
As a Publishing Editor, I’ve often heard authors say: ‘Publishing a book won’t add any weight to the evaluation of my research or university assessment’. I’ve also heard a lot of published authors complain in the past that they didn’t know what kind of impact their book had made or gained the recognition from faculty or funders that they’d hoped for. This was frustrating, not only for authors but for us as publishers, and it was my primary motivation for developing Bookmetrix. I wanted to offer authors as many metrics as possible to evaluate the true reach and impact of their work. Creating a citation metric for books was one step towards that goal, but I also wanted to give them a fuller picture with download data, online mentions, reviews and other altmetrics.
You’ve seen a lot of authors choose not to publish in books because of a lack of impact metrics. Do you expect this to change anytime soon?
What’s interesting is that we’re starting to see more authors move to Springer Nature because of Bookmetrix - this is especially true in Germany. And since the Palgrave merger with Springer, we’ve been able to provide metrics for social sciences books. The significance of this can’t be overestimated. Books are the preferred format in humanities and social sciences disciplines, and it’s really important for authors in these fields to be able to demonstrate how much their titles are being used, discussed, reviewed and cited.
The metrics themselves make a compelling case for the role books are playing in academic research today and are convincing more authors to publish in this format. In 2016, more than 25% of our books had one or more citations within a year and we had 0.91 citations on average per title. The scores were even better for altmetrics.
What do we know about the relative performance of books versus journals in terms of how much they are cited and how long they retain their influence?
What often surprises people is that the average number of citations for a book significantly exceeds the number of citations a journal article receives. On average, a book will pick up 20 citations over its lifetime, versus an average of 10.8* citations for a journal article. And the cited half-life for a book falls somewhere between 20-30 years, which is a lot longer than journal articles. Books are also receiving a growing number of online mentions even for archive content.
*Source: Thomson Reuters Essential Science Indicators Database, 2000-2010
What other metrics do you include in Bookmetrix and how much are these recognised by the wider academic community?
Altmetrics such as downloads and online mentions are starting to gain momentum and credibility in the research community. The number of tweets about a piece of research is now considered by many as an indication of how significant that research is. We include all the following data types in Bookmetrix: book and chapter citation counts; online mentions; readership counts; book reviews and chapter downloads.
Which new features did you release with the 2.0 launch last year and what was the rationale behind them?
With version 2.0, we wanted to expand the capabilities of Bookmetrix for authors, but we also wanted the platform to better support librarians and researchers. This meant including features that would help readers discover new books, and institutions assess the impact of e-book collections they subscribe to. We’ve introduced a benchmarking feature on the Book Discovery Page that lets users see the performance of a title versus the discipline average.
We also launched the Collection Discovery Page, which helps librarians by letting them take a full collection and see its overall performance for a single year using the Collection Citation Performance (CCP) metric. The CCP metric is comparable to the Impact Factor for journals. For the first time librarians can draw realistic comparisons between the impact of journals versus book collections, which is really powerful.
How do you see the Bookmetrix platform evolving in the future?
I want to provide librarians with more of the metrics they want and need to accurately measure research impact at their institution, make more informed purchasing decisions, and better measure return on investment. We’ll be adding new metrics to the platform based on feedback and requests from the library community, and more affiliations data to enable in-depth reporting around the publications history of individual institutions.
Martijn Roelandse is Head of Publishing Innovation at Springer Nature.