Everyone likes paying homage to “science,” deferring to it as the ultimate authority. The problem is that “science” is often more of an “art.” You frequently can find a “study” to support a point of view—regardless of how crazy it might seem.

This phenomenon interests me as a computational social scientist and strategist. In a study recently undertaken with Jonathan Rothwell at Gallup, we leveraged several months of nationally representative survey data covering 500-1,000 respondents per day who were asked about their beliefs about the pandemic.

Questions ranged from expected economic disruptions to their social distancing behavior. We linked these data with information about the number of COVID-19 infections and the unemployment rate at the county-level, together with a wide array of demographic characteristics about each person. We discovered that political affiliation emerged as the most important predictor of beliefs about the pandemic—more than local infections, age, race, and even whether an individual was unemployed in some cases!

If politics plays such a large role in forming beliefs about the pandemic at an individual-level, then what are the implications at a state and local level? We found that differences in political affiliation at a state-level explained many of the differences in the timing and types of state policy decisions. For example, a percentage point increase in the fraction of voters voting for Trump in the 2016 election is associated with a 1.2 percentage point decline in the probability that the state adopts a stay-at-home order. Moreover, increases in the number of state infections were not correlated with the passage of stay-at-home orders after controlling for the 2016 election vote share.

The problem is that the increasing polarization of the pandemic, and the resulting state and local responses, has led to worse policy decision-making.

Let me explain. Although some governors and mayors might say that they are making “data-driven decisions,” the actual data demonstrates decisions have been far from data-driven. Put bluntly, our results suggest that many of the public health policies, such as stay-at-home orders and non-essential business closures, have had little effect in mitigating the virus over the longer run. They have, however, generated significant economic consequences. Instead, more middle-of-the-road policies, such as mask mandates, appear to have had more success in combating the virus with limited economic damage.

One of the areas in which the increasing politicization of the pandemic response has been particularly problematic has been the reopening of schools. Given that 32% of workers have at least one child under the age of 14, reopening the economy requires a sustainable solution for parents with young children—online learning is not a serious and scalable option at these young ages.

Unfortunately, the adoption of stay-at-home orders has had significant adverse effects on the market for child care. In additional research with Chris Herbst and Umair Ali, we found that the adoption of these recent laws led to a 16% decline in early care and education (ECE) job postings. Moreover, even though job postings for the rest of the economy have recovered to trend, ECE job postings remain depressed at 30% of their trend levels. Although there have been some signs of improvement, the reality is that the market for child care has been hit hard, making it difficult for many parents to maintain their usual hours at work while simultaneously caring for their children.

Increasing evidence has emerged indicating children are at much lower risk of contracting and spreading the virus, according to both the Washington Post and Wall Street Journal. And yet, opposition against school reopening remains vociferous. For example, teacher unions have voiced their opposition to school reopening (they also, in an ironic paradox, are against too much remote teaching). However, even if we keep schools closed, we have to cope with the adverse effects on childhood development and the resulting long-term scarring effects on these children that stems from cutting off their learning pipeline and access to socialization in particularly important developmental years.

Unfortunately, many of the current decisions at state and local levels have been driven by politics, rather than data. The fact that we’re in an election year also does not help. But, by drawing attention to inconsistencies and sub-optimal decision-making, constituents can hold their lawmakers accountable and vote them out of office in the election if they so choose.

The United States has so much potential when its citizens unite and work through challenges together. If we can jettison the partisanship and drill down to the core issues, we can make meaningful steps forward. Even if these steps forward don’t solve all our problems, we’ll be in a better position if they just solve a few problems one-by-one, starting with the looming issue of reopening schools in the fall and integrating students back into the classroom.