A new study clarifies when attempts to debunk misinformation about science are most likely to work.

Social psychologists and communication scholars Man-pui Sally Chan and Dolores Albarracín of the University of Pennsylvania conducted the meta-analysis, a quantitative synthesis of prior research, which involved 60,000 participants in 74 experiments. Each experiment either assessed belief in misinformation about science or introduced misinformation about science as accurate and then introduced corrections for the misinformation.

Although on average the corrections failed to accomplish their objectives, they worked better when the issue in the correction was emotionally more positive than the misinformation, the correction matched the ideology of the recipients, the issue was not politically polarized, and the correction provided abundant details as to why the earlier claims were false.

The researchers found that “attempts to debunk science-relevant misinformation were, on average, not successful,” says Chan, the lead author of the paper in Nature Human Behaviour and a research associate at the Annenberg School for Communication at Penn. “Therefore, most of the science-relevant misinformation goes uncorrected even when a debunk is presented. People believe in the misinformation as much before as after the debunk. This is quite notable, because corrections in other domains, such as reports about an accident or political event, do reasonably well, as shown by past research. However, this does not occur in the domain of misinformation about science.”

The researchers conducted their study with two goals in mind. The first was gauging whether the misinformation can be corrected; the second was determining which types of corrections fare better than others.

To achieve those goals, the team began by figuring out whether negative or neutral misinformation is easier to correct. Their investigation confirmed that positive misinformation, which makes people “feel good about themselves, their future, or the world more generally,” the study says, is more challenging to correct than negative misinformation.

Read the full article about science misinformation by Martin Repetto at Futurity.