Questions are being raised about whether the pandemic will lead to a move away from the use of conference proceedings as a major channel for publishing research in some disciplines.

It comes as initial figures suggest that there was a drop in the number of proceedings publications last year, when most academic conferences were forced to make their events virtual because of the Covid crisis.

The latest statistics on the Dimensions database of indexed research, from Digital Science, suggest an overall drop of 90,000 in proceedings published in 2020 to about 375,000.

Daniel Hook, chief executive of Digital Science and also a theoretical physicist, said the drop “makes sense at a high level because I know that a number of conferences that I was taking part in last year were cancelled and not reorganised”.

̽����Ƶ

However, he cautioned, it might be four or five more months before there was enough information on 2020 publications to analyse the full picture, including whether there had been any shift to the use of video recordings in place of written proceedings.

One of the disciplines that would stand to be most affected by any change in such publishing would be computer science, where at least half of all publications indexed in Elsevier’s Scopus database were conference papers from 2015 to 2019.

̽����Ƶ

James Davenport, the Hebron and Medlock professor of information technology at the University of Bath, said that in his subfield of computer algebra he had not detected any impact on publishing in the move to virtual conferences. He also cautioned that some conferences had “post-proceedings” that might not yet be showing up in the data.

However, he said, the impact might have been greater in other subfields, with colleagues saying they had heard of smaller events receiving fewer submissions as authors tried their luck with bigger conferences, which were much cheaper and easier to attend virtually.

There had also been reports of some events originally scheduled for the period when the pandemic first hit not being rearranged online, he said.

Ian Borthwick, head of publishing at the British Computing Society, said their own proceedings publications were “significantly depressed” last year, dropping by about half.

̽����Ƶ

However, he added, submissions to its flagship The Computer Journal had actually risen by about a third, suggesting that there could have been an “adaptive approach happening” with academics choosing other routes for publications.

Holger Hoos, professor of machine learning at Leiden University in the Netherlands, and editor-in-chief of the open access Journal of Artificial Intelligence Research, said there had already been a debate under way before the pandemic about whether there should be a move away from using proceedings as the “primary publication venue” in computer science.

Conference publishing may have become popular in the discipline because of a perception that it was quicker than the journal route, but Professor Hoos said this was no longer necessarily the case, referring to JAIR’s turnaround times as set out by its .

“There are many journals in the life sciences – and even some top-tier journals in computer science – that turn around things a lot faster than computer science conferences,” he said, adding that although the pandemic might not itself cause a shift, a longer-term move “from conference-first to journal-first is under way in my perception and is a good thing for the field”.

̽����Ƶ

simon.baker@timeshighereducation.com