Political partisanship is a better predictor of the fear of dying of COVID-19 than coronavirus cases and deaths, according to new research that examined search data from Google. The study, published in PLOS One, uncovered significant differences between states that supported Hillary Clinton and states that supported Donald Trump in 2016.
There were two main reasons to conduct the study, according to lead researcher Joan C. Timoneda, an assistant professor of political science at Purdue University.
“First is that COVID-19 has had a major impact on politics,” Timoneda explained. “One’s intuition is that, as a society, a response to a pandemic should not be politicized and that we all should act in accordance with the greater public good. This has certainly not been the case, which is fascinating. Second is that I’ve been working with Google Trends data for awhile and I wanted to see what patterns emerged in terms of searches on Google that we could leverage to explain the pandemic’s politics.”
Google Trends is a service that gathers data on search volume for given words or phrases within a specific location and time. In this case, the researchers were interested in examining Google searches for the phrase “will I die from coronavirus” between February 18 and May 30 of 2020, during the early stages of the coronavirus outbreak in the United States.
The researchers found that searches for “will I die from coronavirus” were linked to whether Donald Trump or Hillary Clinton won a state’s electoral votes in the 2016 general election. Even after controlling for factors such as population density, unemployment rate, and urbanicity, searches for the phrase tended to be higher in states that Clinton won compared to states that Trump won.
Not surprisingly, Google searches for “will I die from coronavirus” were more common in states with more coronavirus cases and deaths. But “the correlation between people’s worries about the virus and the actual number of cases and deaths is weaker than the relationship between worries about the virus and partisanship across U.S. states,” the researchers said.
“We show that the response to the pandemic is highly politicized and that people in states that supported Trump are much less likely to be concerned about dying from the coronavirus, even when controlling for cases and deaths in the state,” Timoneda told PsyPost.
The findings are in line with other research, which has provided evidence that political partisanship plays an important role in responses to the COVID-19 pandemic. For example, another study in PLOS One found that shelter-in-place orders tended to be less effective in states with a greater share of Trump voters.
“Google searches are especially good at capturing people’s sentiment for a given topic, and our study adds evidence to what others found,” Timoneda said. “Interestingly, we focus on the first few weeks of the pandemic, and even then you could already tell that politics played a major part in shaping people’s beliefs about the pandemic.”
But the new study — like all research — includes some limitations.
“Google provides daily trends data, which is good,” Timoneda said. “But they are only available at the state level (as well major metro areas), and thus we cannot leverage abundant district- and county-level data on the virus. This would provide a much nicer test for the theory and allow us to use causal inference techniques that we can’t use at the moment.”
“We will continue working on both Google Trends data as well as the politics of COVID-19,” Timoneda added. “How we emerge from this crisis as a society will dictate how well we can tackle other social crises in the future. It imperative that we understand why people do not act together and why polarized politics contribute to bad public policy in emergencies.”
The study, “Will I die of coronavirus? Google Trends data reveal that politics determine virus fears” was authored by Joan C. Timoneda and Sebastián Vallejo Vera.