In three studies (N = 1,055), we investigated the determinants and consequences of the perception of infection risks during the early and later outbreak of COVID-19 in Germany. Individuals’ perceived COVID-19 infection risk was reasonably in line with experts’ assessment but changed over time. The probability of the rare event of getting severely sick and hospitalized was overestimated. Overestimation increased in the exponential growth phase of the pandemic and later on decreased again, showing an inverse U-shaped pattern. Individuals showed biases in their risk perception concerning overconfidence and the underestimation of exponential growth of infection cases in the early phase of the pandemic. Forecasts were more accurate after the growth curve had flattened. Risk perceptions increased with perceived dread and tended to increase with perceived control over infection, the evaluation of scientific and own knowledge about the pandemic. Approval for the introduction of stricter governmental measures and acceptance of future vaccination measures was mainly influenced by rational utilitarian factors of risk perception (probabilities and utilities of outcomes). These rational influences were mediated by dread, but dread had an additional potentially irrational effect. Adherence to governmental recommendations was mainly driven by dread and positive expected long-term consequences of the measures. To a smaller degree, adherence increased with perceived personal consequences of infection and decreased with negative expected short-term consequences of these measures. Implications for theory development are discussed and recommendations to handle virus outbreaks are derived.

The COVID-19 pandemic poses one of the largest behavioral change challenges in the last decades. Because currently, there is no widely available pharmaceutical treatment available to contain the spread of infection, governments worldwide rely – at least to some extent – on behavioral recommendations aimed at reducing spread. The success of this strategy is dependent on the number of people that follow the recommendations. Most recommendations need people to change their behavior or adopt a new behavior. We propose that such behavioral change, with direct costs and delayed benefits, is a source of conflict and mixed feelings. This ambivalence negatively affects adherence to such recommendations. We present three studies that support our hypotheses: the more ambivalent people are about the recommendations, the less they follow them. We also examined the effect of the mixed emotions of hope and worry on adherence and find that it positively relates to adherence. Our findings replicated both in a U.S. sample as well as a representative German sample. Our work is the first to investigate the role of ambivalence in large-scale behavior change and highlight the importance of understanding the conflict that comes with changing behavior. We discuss implications for policy and communication.