Divided perceptions of risk? A new online tool to study the many flavors of polarization

Polarization has become a major concern in behavioral science and popular media, as it may affect many important areas of life. For instance, how polarized are people's perceptions of risks, such as regarding (not) imposing mitigation measures during a pandemic? Answering this question is surprisingly challenging: Whereas multiple theoretical views of polarization and their respective mathematical operationalizations coexist, the latter are often used interchangebly as measures of "polarization". This may be indicative of a jingle fallacy, because it is unknown whether the diverse ways of quantifying polarization in people's perceptions of important societal matters empirically converge. In study 1 we thus ran a reanalysis of a large dataset from the World Values Survey covering diverse topics of societal relevance (N = 93,214), finding only moderate empirical convergence between six operationalizations of polarization. In study 2 we applied the same approach focusing specifically on people's risk perceptions of COVID-19 mitigation measures (N = 768), and found a similar pattern of low convergence between different operationalizations of polarization. Yet, according to one operationalization with a clear threshold for polarization, risk perceptions were polarized in 11 out of 12 experimental conditions. Our findings emphasize the need to carefully consider how polarization is operationalized to avoid broad generalizations, keeping in mind that some operationalizations may speak to specific theoretical conceptualizations. To raise awareness for this concern and support behavioral science researchers in conducting similar analyses with their own datasets, we provide a novel online tool available at https://shiny.cbdr-lab.net/polarization.