How the Covid-19 crisis shaped research collaboration behaviour

The paper explores the impact of Covid-19 on scientists' collaboration behaviour in the 14 countries with the largest research output. The approach is bibliometric, taking the unit of analysis of collaborations as the individual researchers and the co-authorships in their preprints. The time plot of the data confirms a clear discontinuity in the number of preprint depositions after the Covid-19 outbreak. Less evident is the discontinuity in average number of co-authors per preprint, and also in propensity to collaborate. Investigating further, a multivariate econometric analysis shows that for propensity to national collaboration (both intra- and extramural) there has been a positive "effect" from the pandemic, but negative one on international collaboration. The same analysis conducted by country, however, reveals that these effects are significant only in some countries and often with discordant signs.