Online access changes citation patterns
Although the internet gives scientists instant access to a wealth of academic journals, recent findings from the University of Chicago, in the USA, indicate that researchers are now citing fewer papers than ever before. What’s more, these citations are selected from fewer and more recent articles.
With more than one million articles now available free on the internet, sociology professor James Evans wanted to find out exactly how online access shapes scientific research. With funding from the US National Science Foundation, he analysed a database of more than 34 million articles and compared their online availability from 1998 to 2005 with the number of times they were cited from 1945 to 2005.
His results revealed that, as more journal issues moved to the internet, fewer articles were cited – and the ones that were cited tended to be more recent. For example, for every additional year of back issues a journal posted online, his calculations showed, on average, 14 per cent fewer distinct citations to that journal.
Evans also found that scholars now seem to concentrate citations on specific journals and articles, as cited by their peers. As he says: ‘More is available, but less is sampled. And what is sampled is more recent and located in the most prominent journals.’
Discipline differences
But not all researchers are the same. Evans’ investigations indicated that this trend varied across academic disciplines. For example, life scientists and medical researchers tended to reference fewer articles, as did researchers in the humanities and mathematics. In contrast, researchers in business studies, law and other social sciences referred to a wider range of papers.
So why are many researchers overlooking the wealth of information in the past and referring to fewer papers? Evans asserts that the answer lies in the differences between library and online indexes.
‘In “bricks and mortar” libraries or print journals, articles are poorly indexed so you have to read through titles, abstracts and parts of the article to know what the article is about,’ he explains. ‘When online, researchers are more likely to search for particular terms and click through hyperlinks and references to other articles.’
As Evans points out, this means that, when online, researchers can easily sift out material that is less interesting, irrelevant or dated. What’s more, researchers can now access popular opinion much more quickly. Evans believes that this also influences their research and the citations they select.
‘Researchers are now spending less time trying to figure out what other articles mean to them and are more likely to pick up on articles other people have already looked at,’ he says.
And while Evans doesn’t believe this ‘herd behaviour’ heralds the end of independent literature reviews, he is worried that if new research isn’t picked up quickly by scholarly communities, it could be discounted before it has been seriously evaluated.
‘With science and scholarship increasingly going online, findings and ideas that don’t receive attention very soon will be forgotten more quickly than ever before,’ he warns.
The open-access effect
A separate US study has also cast doubts on hopes that open access will boost citations.
Researchers at Cornell University concluded that open-access academic articles get read more often but don’t generate more citations. In their study, they randomly assigned 247 articles in 11 scientific journals, to free access. They measured how many times these articles were downloaded, the number of unique visitors to each article and how many times each article was cited.
‘There were definitely more article downloads for freely-accessible articles,’ said Philip Davis, a former science librarian who designed the study. ‘Yet nearly half of these downloads were by internet-indexing robots like Google, crawling the web for free content.’
‘The established dogma is that freely-available scientific articles are cited more because they are read more,’ said Davis. ‘We found that openaccess publishing may reach more readers than subscription-access publishing, but there is no evidence that freely-accessible articles are cited any more than subscription-access articles.’
The University of Chicago findings appeared in Science. The Cornell study was published in the British Medical Journal.