Increasing division of labour makes science more fragile

Some team structures make mishaps more likely. Funders should encourage robust working practices even if they are more expensive, say John Walsh and You-Na Lee.

Many famous recent scientific fraudsters had no shortage of coauthors, who were startled to find that the results they were putting their names to were fictitious. But even when everyone acts in good faith, team science—which these days effectively means most science—can be surprisingly fragile. The rise of collaboration has been mirrored by a tenfold rise in retractions over the past 25 years.

This suggests that, while it brings great benefits, collaboration can also introduce structural flaws into how research is done. These become particularly vulnerable when combined with incentives that reward speed and productivity over caution and accuracy.