The dizzying pendulum of conflicting information experts have offered during the coronavirus pandemic has revealed the perils of relying too much on computer models for scientific studies.
More than 20 scientists and experts wrote an editorial in the June 24 issue of Nature saying the politicization of COVID-19 research shows the need for change in conducting and reporting computer modeling studies. They also criticized how politicians use the results.
“Pandemic politics highlight how predictions need to be transparent and humble to invite insight, not blame,” the group wrote.
Despite myriad COVID-19 studies, researchers cannot provide precise and reliable numbers and predictions because of the number of unknowns about the disease, the editorial noted. Politicians and policymakers need to pay attention to the inherent uncertainties in computer models. Just like weather reports are helpful but have significant limitations, computer models are useful in understanding COVID-19, but “political rivals often brandish them to support predetermined agendas,” the writers said. Scientists should be frank about the unknowns in their studies and not project more certainty than their model deserves. And politicians must not cherry-pick the results they like.
The writers of the editorial also encouraged scientists to keep their models simple and openly share their biases, noting that modeling techniques are never neutral: “Results from models will at least partly reflect the interests, disciplinary orientations, and biases of the developers.”
Additionally, people need to understand that computer predictions offer possibilities, not certainties. The World Health Organization predicted on May 7 that 83,000 to 190,000 people in Africa could die of COVID-19 in the first year of the pandemic. Researchers came to that number by incorporating 10 uncertain probabilities and increasing each by 10 percent—an arbitrary number with no scientific basis, the editorial writers explained. But people believed those predictions gave an accurate picture of the future. As of Thursday, Africa had seen more than 12,200 deaths from the coronavirus.
Mathematical projections become dangerous if people expect too much of them, the writers said, adding that without appropriate guidelines, “model predictions become Trojan horses for unstated interests and values.”
David F. Coppedge, the founder of the website Creation Evolution Headlines, noted how the editorial writers’ suggestions could affect a wide range of political policies and controversies.
“Take some of the leading political controversies of the day—COVID-19 policy, climate change, transgenderism, Darwinian evolution—and run them through the five recommendations,” he said. “What survives?”
Albert Mohler, president of The Southern Baptist Theological Seminary and a member of WORLD News Group’s board, offered a similar commentary on his podcast The Briefing: “When you’re in the middle of a moral controversy, when you’re in the midst of a political controversy … it is very frustrating to see, but very revealing to understand, that science is invoked in ways that actually aren’t scientific.”