The Secret Life of Articles: From Download Metrics to Downstream Impact

by | Nov 5, 2015 | 0 comments


Lorraine Estelle, Wouter Haak, Carol Tenopir

(L-R) Lorraine Estelle, Wouter Haak, Carol Tenopir

Carol Tenopir, Chancellor’s Professor, University of Tennessee, began this session with an update of her report at last year’s conference on sharing and using downloaded scholarly articles. The research is now completed. The study examined whether there is a way to measure the total use of an article, beyond just number of downloads and whether we can find a more complete measure of value. Many of our implied measures are based on use, but we are not measuring sharing.

The study methodology was focus groups in US and UK and an international survey. Respondents were mostly social scientists and scientists; not many were working in the humanities. The survey got 1,000 responses. Age range of the respondents was 18 to 83; the average age was 43.

The main takeaways from this research are:

  • Sharing of scholarship is a means to an end. This word cloud shows some of the feelings of the respondents towards sharing.

Sharing tag cloud

The vast majority of comments were positive. The main reservations were that researchers want to be sure proper citation is made to their work so they get credit. There were not many comments about the legalities of sharing. Sharing facilitates research and is a natural part of scholarship. Many people ask colleagues for copies to get around paywalls, if they don’t have a copy, don’t have a collection, or can’t get access to online copies. Other reasons include convenience, fitting in with work, and helping others. Scholars working in a research group are more likely to share. We can expect sharing to increase especially as the number of authors per paper increases.

  • Function drives form. Sharing does not always mean full text. Within a research group, researchers tend to just share a link. Full text is shared when access is an issue or it is easy to do. When a link is shared, downloads may go up, sharing the full text decreases the number of downloads.
  • Sharing by email is done about 11 times/download, for teaching it is about 14 times/download. Data for full text sharing are similar. Email is very much the primary way to share. The number of people reached is much higher for sharing in research social networks than by email.
  • Version matters: When sharing articles, the final peer reviewed and published version is much preferred because it is the one that will be cited. Content providers should make their product available legally as easily as pirating is and make it so cheap that it is not worth pirating.
  • The library is key in the sharing process.  Library downloads may underestimate usage. When rereading a shared article, most people use a saved copy.

Conclusions of the research study: Sharing is a natural part of the research process. Convenience is a key concern. Sharing channels are chosen to fit a work style. New policies and measures must fit preferences and likely behavior.

Lorraine Estelle, the new director of Project COUNTER, discussed studying possible approaches to the calculation of sharing metrics. Her key findings include:

  • Lack of global usage data makes quantifying sharing problematic,
  • It is difficult to obtain data on authors’ sharing practices, and
  • Any data obtained is likely to be outdated quickly.

Suggested approaches to measuring sharing data are a data-based approach using data from publishers and citation data, and a survey-based approach. The data-based approach will be difficult because of the large number of channels and a lack of standards. Demographics have a large influence on the data; researchers age 39 or less share data less through formal methods. Social scientists share more via social platforms. Here are some possible alternatives:


Collaborative initiatives with CrossRef: DOI event tracking, distributed usage logging (peer-to-peer system for exchange of usage data).

Wouter Haak, VP, Research Data Management Solutions, Elsevier, discussed the implications of sharing on Elsevier’s products, researcher behavior, and what it means for librarians. Mendeley (now an Elsevier product) is a collaborative sharing platform. Sharing is growing tremendously as is article growth. Elsevier wants to make it so easy to share that researchers will prefer to share on collaborative platforms, so they have taken the following 3 steps:

  1. Created a “my research dashboard” to feed back to the researcher what is happening with their article: how many downloads and views of both Elsevier and non-Elsevier articles have occurred. The feedback starts immediately after publication to get signals of who is reading the article, adding it to a Mendeley library or how many people decided to share it with their group, which is a very strong signal for the researcher.
  2. Providing feedback to librarians by the E-Pic system, which is an institutional dashboard showing what is happening with authors and readers in the institution.
  3. Developed principles of sharing and why they believe it is important to feed back sharing data on researchers. So far, they cannot track email usage, but they are talking to platform providers. Sharing is not bad, Elsevier likes researchers to read their articles, and is not worried much about entitlements. They created principles with 10 publishers, asked the community for feedback, and got 52 responses (see  A new set of voluntary principles has been developed.

Voluntary Principles

Sharing could be allowed within a research group and with the wider public. Commercial research is different; Elsevier is asking those platforms to work with COUNTER and CrossRef to allow feeding back to dashboards of recommendation engines. The task of librarians should be to help researchers with discovery and give good recommendations. So far Nature, Springer, Elsevier, and Wiley have changed their policies.

Next steps are to determine if librarians are comfortable with this. They are seeking input from the community and would like feedback; Google “STM principles” or “STM consultation” to give opinions.


Sign-up Today!

Join our mailing list to receive free daily updates.

You have Successfully Subscribed!

Pin It on Pinterest