How can humanists increase the attention paid to their work, using insights from altmetrics? At the ASIS&T Annual Meeting in November, I debuted a study I’ve been working on with Jeremy McLaughlin (ASIS&T/SJSU) on the characteristics of altmetrics data for humanities scholarship archived on figshare. We shared some research-backed ways to stand out and promote one’s work amidst an ever-growing sea of web-native humanities scholarship. Below, we’ve adapted much of our talk for the purposes of this post. We expect to publish formally on our study in 2016.
As regular readers of this blog likely know, altmetrics are the traces that are left behind when other people share and discuss scholarship online. Research is increasingly showing that this data can help us understand the many ways that our research is making a difference in the world.
However, most research done to date on altmetrics is studying the impacts of science scholarship. Bibliometricians have asked many questions concerning altmetrics for journal articles (a format more important in the sciences than in the humanities), their use in scientific disciplinary communities like astrophysics or biomedicine, and whether or not various altmetrics correlate with citations (an interesting question, to be sure, but certainly one of less importance in disciplines that don’t rely heavily on citations to understand the influence of scholarship).
Where are the humanities in this discussion? And what can humanists learn from the important feedback that altmetrics can provide?
Humanities research is a different animal
Arts & humanities researchers publish more diverse, non-article outputs than their peers in the sciences: books, book chapters, exhibition catalogs, textual datasets, book reviews, images, videos, audio files, and many more.
When it comes to research evaluation, humanities researchers are often not concerned with citations counts for their work. Instead, measures of impact in the humanities might include:
- Prestige of publisher for one’s book
- Peer-review status of work
- Prizes & awards received
- Mass media contributions
- Number of national or international presentations completed
Because of the diversity of formats that humanists use to disseminate their work and their reliance on indicators of impact beyond citations, the humanities stand to uniquely benefit from the insights on impact that altmetrics can offer, being that altmetrics apply to all scholarly outputs–not just journal articles or books–and can help one understand impact among scholarly and non-scholarly audiences alike.
Altmetrics + humanities: a great match
A growing body of altmetrics research is helping humanists understand the many impacts of their publications:
- Humanities-related Wikipedia articles have more contributors (including repeat contributors) and more edits than science-related articles (Ehman, Large & Beheshti, 2008), indicating greater public and educational engagement with humanities scholarship;
- Humanities scholars are more likely to use social media in their research (Rowlands et al, 2011), and are more likely to discuss humanities research online (the traces of which generate altmetrics);
- Humanities scholarship is more often discussed on social media than indexed in citation databases like Web of Science (Haustein, Costas & Larivière, 2015); and
- Humanities research tends to have higher Mendeley readership than it does citation counts (Mohammadi & Thelwall, 2014), and a survey on motivations for using Mendeley indicates that scholars are engaging with humanities research in order to use it in teaching and more.
To these insights, we’re adding our own: how are humanists sharing their “alternative” scholarly outputs like data and software in open access repositories, and what sorts of attention is that scholarship receiving online?
What percentage of humanities research is discussed and viewed online?
For our study, we chose to examine metrics for arts and humanities content deposited to the open repository figshare. For these outputs, we studied the distribution of figshare- and Altmetric.com-sourced altmetrics, including a breakdown of metrics for different content types and disciplines over several years.
We started with 5,564 humanities related outputs on figshare, including content across 20 arts and humanities disciplinary categories, 9 unique document types, and publication dates (January 1, 2012 to July 14, 2015).
Research archived in figshare receives a unique DOI upon deposit, which has allowed Altmetric to more consistently track mentions of that research across the social web, including on Twitter, blogs, Mendeley, and mainstream media sites.
DOIs for all 5,564 humanities figshare deposits were analyzed using Altmetric Explorer. 622 humanities-related deposits (11%) had at least one altmetric count (i.e. one tweet, one mention in the news, etc). An additional 1,303 deposits (23%) were downloaded at least once from figshare and 61 (1%) had figshare “likes” (also described as “shares” on the figshare webapp).
Overall, the number of deposits for humanities-related research have increased over the last few years. However, while the amount of content has increased, the percentage of outputs with an Altmetric.com score has gone down over time (Figure 2).
After refining our dataset to remove skewed elements, we were left with a representative sample of 1,467 deposits for our analysis. Across these humanities-related outputs, 81% received one or more downloads from and 38% had been “liked” on figshare. In terms of altmetrics, 5% of outputs had a mention on a blog, 39% on twitter, and 4% and 3% had been mentioned on Google+ or Facebook, respectively.
We saw negligible amounts of attention from news sites, Reddit, Wikipedia, and Sina Weibo, so we did not include them in our analysis. There was no attention to these outputs at all from other Altmetric.com data sources like CiteULike, Pinterest, or Faculty of 1000. Because Mendeley readership counts are not recorded by Altmetric if a research output isn’t also mentioned elsewhere online, we did not include those numbers in our analysis.
Characteristics of humanities outputs with online attention
Categories & Keywords
Of the 1,467 representative deposits that had received some attention, we saw that content was primarily assigned categories including literature, law, history, philosophy, and digital humanities (see Table 1). One in four outputs was assigned to multiple categories. These categories are user-assigned, but come from a controlled vocabulary provided by Figshare.
We also noted that disciplinarity may influence discovery and potential interest on social media. As shown in Figures 2 and 3, the 396 outputs assigned to two or more categories (which comprised only 27% of all outputs with attention) generated more than 3,600 downloads and 1,023 tweets total–a much higher download-per-output and tweet-per-output ratio than research only assigned one category. (We should note here, however, that metrics like “average downloads per output” can be misleading due to outliers and a number of other factors. Always take these numbers–and the claims based on them–with a grain of salt.)
Outputs archived on figshare are not only assigned categories from a controlled vocabulary, but can also be tagged with user-specified, freetext keywords. Looking at the top 10% of all humanities-related outputs that received various types of attention (figshare likes, figshare views, figshare downloads, and overall online attention as measured by the aggregate Altmetric score), we saw some themes (Figure 5).
Overall, keywords were likely to be “meta” in nature: rather than describing specific subjects, they tended to describe the nature of the research itself (“data,” “digital humanities,” “open science,” etc). That is, with one exception–“comics”–a tag that saw a great deal of online attention due in large part, no doubt, to the online outreach efforts of researcher (and former Altmetric colleague) Ernesto Priego.
Research output types
While papers are by far the majority (49%) of humanities-related outputs archived to figshare (n=1,467; see Table 2), there were a significant number of datasets (22%), filesets (9%), figures (6%), and presentations (6%) also uploaded to the repository.
Despite the large number of papers, filesets and datasets were the most downloaded content types by percentage, presentations had the most blog mentions, and posters and presentations were the most tweeted output types.
What does this mean for how humanists should share and promote their work?
Using our preliminary analyses as a guide, we’ve found two initial hacks that humanists can use to make their research “stand out” online.
Deposit everything (and promote the heck out of it)
The increase in overall humanities-related deposits to figshare and the decrease in social media activity for these deposits over time may point to the fact that a humanities information overload is now happening on the platform. Be prepared to do outreach for anything that you want to draw attention to, rather than just uploading it and hoping it will get noticed. That way, you can be sure that your work doesn’t get lost amidst the “noise”. For tips on promoting your work online, check out the Impactstory “30 Day Impact Challenge” ebook.
Proactively promote your work using the right networks and descriptors
Given the explosion in humanities-related research deposited on figshare, it’s important that humanists promote their deposits on social media wisely, targeting the right sharing platforms and the right demographics who might be interested in their research.
So, what is “right”?
The “right” platform, by and large, is currently Twitter. According to our data, most engagement with humanities-related figshare deposits tends to happen on Twitter, so that is likely an ideal candidate for a place to concentrate your outreach. (It should be noted, however, that this conclusion is based upon Altmetric data, which includes a more curated, limited set of sources than exist overall online. Recent studies have pointed to Mendeley and Goodreads as other important sources for humanities impact data, as well.)
In terms of connecting with the “right” demographics, you can help them find your work by using both broad and specific keywords when describing your archived data–it aids in discoverability and that may affect how often your work is discussed on social media, and by whom. Also consider using more than one category to classify your figshare deposits, as interdisciplinary deposits get more attention overall online.
As for who those demographics are–we don’t yet know! More study is needed.
We’ll be continuing our analysis of the data and writing up these results formally soon. We also plan to release the data underpinning the study on (where else?) figshare. Stay tuned!