Citing in 140 characters

Using the Israeli SF&F society's time machine* I managed to get an advanced copy of Priem and Costello's paper: "How and why scholars cite on Twitter", which will be presented at the ASIST 2010 conference (22-27 October). As one can learn from the paper's name, it deals with researchers' Twitter citing behavior.

The snowball sampling here is a bit problematic, as the authors themselves admit in the conclusion part. They started with 3 academics from different disciplines, asked them to tweet about the research, and asked new participants to tweet about the research as well. They ended up with 28 academics (faculty, post-docs and Ph.D students).

They took a mixed-method approach, both interviewing the participants and harvesting their tweets. They sampled each participant's last hundred tweets and their final sample included 2,322 tweets containing hyperlinks.

Citations here were defined as "direct or indirect links from a tweet to a peer-reviewed scholarly research online". The citations were divided to direct citations (tweets which led straight to the academic resource) and indirect citations (there's another web page between the tweet and the academic resource). 6% of the tweets in the sample contained citations. A bit more than half (52%) of those citations were direct and the rest were indirect.

Main conclusions

• One of the reasons for indirect t-citings is the nonavailability of the academic resource itself. Instead of linking to a pay-walled paper, the link can be to a blog post discussing the paper.
• T-citings are considered part of an ongoing conversation between colleagues and are part of the researcher's workflow.
• Unlike most citings in traditional academic discourse, t-citings aren't usually used to support an argument, but to filter information. A t-citing can often be considered a recommendation for the resource in question.
• T-citings are faster than a speeding bullet. 15% of the t-citings were tweeted at the publishing day of the papers they referred to, and 39% of the t-citings were made within a week of the paper's publication.

This is the first study I've read about Twitter citings and I believe the authors did a good job in capturing the "water-cooler" essence of t-citings. A bigger sample might allow to study the differences between the t-citing habits of researchers from different disciplines.

*Okay, I emailed the authors.

Priem, J., & Costello K. L. (2010). How and why scholars cite on Twitter ASIST, Oct. 22-27, Pittsburgh, PA, USA.