Coase-Sandor Working Paper Series in Law and Economics

Framing Vaccine Mandates: Messenger and Message Effects

Christopher Buccafusco
Daniel Hemel


In September 2021, President Biden announced that the Occupational Safety and Health Administration (OSHA) would require all employers with 100 or more employees to ensure that their workers are fully vaccinated against Covid-19 or show a negative test for the virus at least once a week. The policy has been widely characterized in the media as “President Biden’s vaccine mandate,” though it could be described with equal accuracy as “OSHA’s testing mandate” (since OSHA, rather than Biden, officially promulgated the policy, and once-a-week testing and vaccination are both valid compliance options). Some commentators have speculated that reframing the policy as a testing mandate (with a vaccination option) rather than a vaccine mandate (with a testing option) would boost public support. This study seeks to gain empirical insight into how framing effects shape attitudes toward vaccination and testing policies.

In October 2021, we presented a nationally representative sample of 1500 U.S. adults with different descriptions of the same vaccinate-or-test requirement. We find that recharacterizing “President Biden’s vaccine mandate” as “OSHA’s testing mandate” yields a substantively and statistically positive effect on support, boosting the policy’s net approval margin by approximately 13 percentage points. The effect of reframing is particularly strong among self-identified Republicans, who overwhelmingly oppose the policy when it is framed as President Biden’s vaccine mandate but are more evenly split when the policy is framed as OSHA’s testing mandate. Further analysis reveals that the positive effect is driven by the change in the messenger frame (i.e., switching the promulgator of the policy from President Biden to OSHA). By contrast, changing the message frame from a vaccination requirement (with a testing exception) to a testing requirement (with a vaccination exception) has little independent effect on respondents’ attitudes.

Our results suggest that messenger framing can have meaningful effects on public opinion toward a policy even after the policy is widely known. Beyond the Covid-19 context, our study points to a potential cost of presidential administration when partisan divisions are deep. Our results suggest that framing a regulatory policy as an extension of the president can elicit strong—and in this case, negative—reactions that may be avoidable if the same policy is framed as the work of a bureaucratic agency.