"It is often considered that uncertainty should be swept under the carpet to make the message of scientific research sharper," Valentina Bosetti from Bocconi University and Fondazione Eni Enrico Mattei in Italy told environmentalresearchweb. "We find that, in this specific context, this is not the case, rather more scientific details on the relevant uncertainty appear to be more effective."
Bosetti and colleagues asked 217 policymakers at COP21 to say what they thought the global mean temperature increase would be in 2100. Then they showed them one of three visual presentations of the same climate model forecasts and asked them to update their predictions for global mean temperature increase.
The policymakers adopted the information "very conservatively", according to the researchers, "assigning it less weight than their own prior beliefs". But providing individual model estimates as well as the statistical range resulted in the policymakers taking more account of the data.
"Negotiators and policymakers have strong a priori, in this case about the long term climatic implications of the nationally determined contributions, or NDCs," said Bosetti. "When provided with more detailed information they seem to depart more from their prior [beliefs] and incorporate to a greater extent the scientific information provided."
Around half of the policymakers in the sample were active negotiators. These negotiators tended to take less account of the scientific data than their non-negotiator colleagues, perhaps because they wanted to remain close to their country’s negotiation position, according to the researchers.
"I deeply believe that communicating scientific results is key for science-based policymaking, as [do] my co-authors," said Bosetti. "Yet, as scientists we tend to overlook this ‘last mile’ problem. We concentrate on performing the analysis correctly but we tend to overlook the problem of how to communicate this research to policymakers."
Most of the policymakers surveyed did not assign more than a 20% probability to the 2100 temperature increase being less than 2°C, despite the Paris agreement’s stated goal of keeping temperature rises "well below" 2°C.
The team also surveyed 113 MBA students from seven European business schools who were role playing climate change negotations in May 2016. Although the students held similar beliefs to the negotiators before seeing the data, they altered their views more strongly in response to the new information. The students also responded the same no matter what format they saw the data in.
"It is now evident that tailored experiments are needed to improve the ability of scientific results to be accessible to the final users," said Bosetti. "We … show that, as known, it is important to run experiments on the population of interest, as student experiments may have low external validity."
Science-based policymaking, particularly in the field of climate change, is increasingly based on data generated by multiple models and algorithms, Bosetti explained. The resulting data are affected by multiple sources of uncertainties. "This makes the communication of these increasingly complex scientific projections a task which itself is an important research field," she said. "It is vital to create a bridge between psychologists and scientists developing these projections in order to minimize misunderstandings and failed opportunities to reach key audiences."
Bosetti and colleagues reported their findings in Nature Climate Change.