The average number of authors on scientific papers has increased by more than 50 per cent in the past 20 years, new figures show.

The statistics, produced by Thomson Reuters, show that the papers included in the Science Citation Index now have an average of 4.83 authors, up from 3.18 in 1990.

All branches of science saw an increase, with the biggest rises in papers describing large-scale clinical trials and high-energy physics experiments, which may have hundreds of authors.

Barry Clarke, president of the Engineering Professors' Council, speculated that the rise was due to the increasingly collaborative and interdisciplinary nature of scientific research.

³ÉÈËVRÊÓƵ

He said that in the UK this has been driven by the research councils and the research assessment exercise, which assigns funding on the basis of research quality and which is to be replaced by the research excellence framework.

"Universities are under pressure to increase their research activity because they want to increase their position in the league table. So they respond by increasing the engagement of their staff in research or appointing staff already engaged in research," he said.

³ÉÈËVRÊÓƵ

In some cases researchers may have been named as authors even though they had made little direct contribution to the work because their department wanted to enter them for the RAE, he added.

The Thomson Reuters figures also show rising rates of citations and references, with the latter climbing from an average of less than 23 per paper to more than 34 per paper.

Professor Clarke said this could reflect the fact that peer reviewers had become more insistent on the provision of thorough literature reviews.

Martyn Poliakoff, research professor of chemistry at the University of Nottingham, added that the internet had also made it a lot easier to locate references.

³ÉÈËVRÊÓƵ

"When you had to spend two days in the library looking for a reference, you probably wouldn't bother," he said.

The number of scientific articles indexed by Thomson Reuters increased from fewer than 600,000 in 1990 to more than 1 million in 2009.

Over the same period, the average number of articles published each year jumped from 185 to 3 per journal.

paul.jump@tsleducation.com

See Research: Inflationary times: the changing profile of journal-publishing statistics

³ÉÈËVRÊÓƵ