The spokesperson matters: evaluating the crisis communication styles of primary spokespersons when presenting COVID-19 modeling data across three jurisdictions in Canada

Risk prevention measures are more likely to be accepted if people trust risk managers and their ability to handle the crisis, which often depends on who communicates with the public. During the COVID-19 pandemic, some Canadian officials became the main spokespersons of pandemic response in their jurisdiction, speaking almost daily to the public. We evaluated how the primary official for each jurisdiction chose to communicate about epidemiological modeling with the public and how they used modeling data to support their pandemic decisions. We conducted textual and visual analyses of press conferences held in British Columbia, Manitoba, and Ontario. Then, we asked focus group participants who they trusted the most and the least for information on COVID-19. We identified two main communication styles: compassionate-informative and condescending-evasive. Spokespersons following the former demonstrate a trust-building effort by providing straightforward answers, demonstrating expertise, while showing empathy and risk management competence. Those who followed the latter style predominantly offered superficial and defensive responses, engaging in blame-shifting and politicizing risk. Focus group participants trusted most the spokespersons who follow a compassionate-informative style are considered trustworthy, which could increase compliance with public health measures. However, those who use the condescending-evasive style were seen as less trustworthy. Our results underscore, first, the importance of disassociating political agendas from risk communication and emergency response during public health crises. Second, spokespersons should be trained in risk and crisis communication to engage with reporters and the public positively. Finally, crisis communication should emphasize the scientific evidence behind guidelines, while acknowledging scientific uncertainty.