Today was the weekly Cardiff University Astrolunch talk and I was the speaker. My topic: the effect of the internet on astronomy research papers. With the rise of astro-ph and NASA’s ADS (astrophysics data system) paper database, the Internet has really changed the way that researchers access and distribute information. Here is the content of my talk, reformatted to make it bit easier to read online:

How We Got Online

We begin with a quick history lesson. This is merely to help place into context, some of the facts and figures I will be detailing.

1991: arXiv begins.
1993: Mosaic graphical web browser puts people online, at home.
1994: Netscape Navigator is released becomes standard web browser until Microsoft launch Windows 95 with Internet Explorer built in. Browser wars ensue.
1995: NASA ADS begins online.
1998: Google begins, rise of the search engine.
2000: Dotcom bubble bursts, paves way for new generation of technologies including blogging, social networking etc.
2002: Web becomes ubiquitous in Western culture. This date is obviously only approximate.

It’s All About Numbers: ADS

The ADS receives around 3,000,000 readers a month. These are papers actually read, not including hits that result in no followed link.

Over 1,000,000 unique users every month.

Around 30,000 regular users, who access the service more than 10 times a month.

Fig 1 - The way ADS is used. Data taken from January and February 2008. Shows the things people do with articles they find on ADS.

Fig 2 - The way ADS is found. Data taken from January and February 2008. Shows the ways that users find NASA’s ADS.

It’s All About Numbers: arXiv and astro-ph

arXiv receives around 1,000,000 website hits every working day.

More than 900 papers are submitted to astro-ph every month.

Submitting your paper in the final minutes before the 4pm deadline (US Eastern Time) will increase your citation rate.

Living in the United States increases your chances of achieving this effect.

Fig 3 - Shows the way different journals have adopted astro-ph by looking at the percentage of papers from those journals which are also published on astro-ph.

Fig 4 - The characteristic way that a paper gets read. Initially papers are read on astro-ph when they first go up. After they have appeared on ADS, users tend to prefer to read the refereed article there.

Fig 5 - Typical example from Monthly Notices of the Royal Astronomical Society. Papers published in 2007, that were also published on astro-ph received 2.4x as many citations as those that were not. This factor is called the astro-ph Impact Factor.

Fig 6 - Variation of the astro-ph Impact Factor over time for four different journals. Note that the trend is for astro-ph to become more influential as time goes by.

Fig 7 - Impact of astro-ph on Citations for Different Journals. These data points are the averages over the past six years.

Why is astro-ph Influential?

Dietrich (2008a) propose three mechanisms for the impact of astro-ph and other online archives.

Open Access - Because the access to articles is unrestricted by any payment mechanism authors are able to read them more easily, and thus they cite them more frequently.

Early Access- Because the article appears sooner it gains both primacy and additional time in press, and is thus cited more.

Self-selection Bias - Authors preferentially tend to promote (in this case by posting to the internet) the most important and, thus, the most citable articles.

You Can Game the System

Papers appearing at the top of the astro-ph listing each day are seen by more people and thus cited more often. The last submission before the 4pm (Eastern US) deadline appears most prominently in the next mailing and will be listed first. Thus, if you time it right, you can try to place your paper near to the top of the list! The result is that some authors tend to promote their most important works and, thus, most citable articles, by placing them at prominent positions. This not a strong effect but it is also not uncommon.

Being based in the USA the submission deadline preferentially puts those authors at the top of the listing whose working hours coincide with the submission deadline. This is also only a slight effect, and according to Dietrich (2008b) it is more than cancelled out by the author placement effect described above.

In Summary

In a few years, use of astro-ph will be almost universal. The astro-ph preprint archive has had a marked affect on the way academic papers are read and cited. Statistics appear to show astro-ph having an increasing impact, probably as more people use the service more often with improved Internet access.

ADS as a global service introduces biases toward submissions at particular times of day. astro-ph does not diminish readership for publications, nor does it significantly increase use of the ADS. Journals which use astro-ph the most do not see a significantly better impact compared with those that use it less. But all journals see an impact.

Submissions purposefully made near the deadline, artificially increase citation counts but cancel out geographical bias toward North America.

References

Dietrich, J. P., 2008a, “The Importance of Being First: Position Dependent Citation Rates on arXiv:astro-ph”, Publications of the Astronomical Society of the Pacific, Volume 120, issue 864, pp.224-228 - [Link]

Dietrich, J. P., 2008b, “Disentangling Visibility and Self-Promotion Bias in the arXiv:astro-ph Positional Citation Effect”, Publications of the Astronomical Society of the Pacific, Volume 120, issue 869, pp.801-804 - [Link]

Henneken, E. et al., 2006, “Effect of E-printing on Citation Rates in Astronomy and Physics”, Journal of Electronic Publishing, vol. 9. - [Link]

Hennken, E. et al., 2008, “Use of Astronomical Literature - A Report on Usage Patterns”, eprint arXiv:0808.0103 - [Link]

Metcalfe, T. S., 2005, “The Rise and Citation Impact of astro-ph in Major Journals”, Bulletin of the American Astronomical Society, vol. 37, p.555-557 - [Link]