Logarithmic vs. Linear Visualizations of COVID-19 Cases Do Not Affect Citizens’ Support for Confinement
In public health crises, the media and governments routinely share statistical analyses with the public. In the COVID-19 pandemic, the tool most commonly used to convey statistical information about the spread of the virus has been time-series graphs about the cumulative number of cases. When drawing such graphs, analysts have to make design decisions which can have dramatic effects on citizens’ interpretations. Plotting the COVID-19 progression on a linear scale highlights an exponential “explosion” in the number of cases, whereas plotting the number of cases on a logarithmic scale produces a line with a modest-looking slope. Even if the two graphs display the exact same information, differences in visual design may lead people to different substantive conclusions. In this study, we measure the causal effect of different visualization design choices on Canadians’ views about the crisis. We report results from a survey experiment conducted in April 2020 with a sample of 2500 respondents. We find that no matter how the information is presented, Canadians are united in supporting drastic confinement measures and in accepting that these measures will not be removed soon.