Skip to main content

Verified by Psychology Today

Why People Are Not Budging Despite Mask-Wearing Nudging

Neither Republicans nor Democrats are likely to change their minds about masks.

Key points

  • Attitudes regarding mask-wearing behavior are highly polarized.
  • We tested Republicans and Democrats separately on whether we could increase the rates at which people wear masks.
  • We found that behavior change is difficult in such a polarizing environment and nudges alone might not be enough.

Co-written with: Elena Giulia Clemente, Michele Gelfand & Dylan Pieper

Despite the availability of effective vaccines for the virus that causes COVID-19, it remains important to public health to wear a mask. Studies show that wearing a mask is effective at reducing transmission of the virus and that masks play an important role in reducing infections, even alongside high rates of vaccine compliance. However, much like with vaccines, there is a deep political divide on wearing masks in the United States, especially as mask mandates become more relaxed. A recent poll revealed that 37 percent of Republicans said they will wear masks all or most of the time when in public indoor settings compared to 77 percent of Democrats.

Source: Svetlana_Khoruzhaia/Shutterstock

Behavioral scientists have made a number of attempts to promote wearing a mask. Some experimental interventions have succeeded in encouraging people to wear a mask by communicating that COVID-19 is a threat to one’s community and evoking empathy through stories about how masks protect older adults who are vulnerable to COVID-19. Other research has found a lack of evidence that interventions influenced by psychology encourage wearing a mask more than a standard health message. However, these studies were not intended to evaluate the effects of the interventions on Democrats and Republicans, nor did they tailor the interventions to these groups. There is also a lack of insight into whether behavioral scientists are good at predicting what interventions will work in polarizing contexts, which we know from existing research can impede efforts to change behavior.

Masking Polarization

To address the issue of political partisanship on mask-wearing, we designed two studies. First, we designed an experiment with interventions that promoted wearing a mask with a total of seven different messages tailored to the motivations of Democrats and Republicans. Second, we designed a prediction study in which we asked academics, behavioral scientists, and laypeople to predict the effects of the seven interventions. Both studies were large and representative of the U.S. population as well as matched for demographics across political parties.

In our first study (by Gelfand, Li, Stamkou, Pieper, Denison, Fernandez, Choi, Chatman, Jackson & Dimant), participants read a baseline message that masks are important, or one of seven messages reflecting different factors specific to COVID-19—including protecting oneself, protecting one’s community, fulfilling a patriotic duty, feeling disgust (desire for purity), reviving the economy, addressing the deaths caused by the virus, and trusting scientific evidence that masks are effective. Consistent with public polling, we found that Republicans had significantly more negative attitudes toward masks, lower intentions to wear them, and were less likely to sign or share pledges to wear a mask on social media than Democrats. However, none of our messages significantly affected Republicans’ or Democrats’ attitudes, intentions, or behaviors compared to the baseline condition. The interventions were not able to change deep-seated partisanship on wearing masks.

Eugen Dimant
Effect of the intervention messages on attitudes, intentions, and pledges to wear a mask across political parties.
Source: Eugen Dimant

In our second study (by Dimant, Pieper, Clemente Dreber & Gelfand, in collaboration with the Behavior Change for Good Initiative (BCFG), the Behavioral Insights Team (BIT), ideas42, and the Center for Advanced Hindsight), we evaluated whether academics, behavioral scientists, and laypeople were able to predict the results of the interventions to promote wearing a mask. To do this, we set up a survey in which respondents were shown the intervention messages from the original survey. They were then asked to predict the effectiveness of each intervention relative to the control baseline, separately for Democrats and Republicans. Data for academics and practitioners was collected through social media and academic mailing lists and a collaboration with multiple behavioral science units to reach practitioners.

We found that forecasters are able to predict the high baseline attitudes and intentions towards wearing a mask, both for Democrats and Republicans. However, they could not accurately infer the effectiveness of the interventions. In particular, they expected the interventions to have a larger effect on Democrats than on Republicans, relative to a no-intervention baseline. They overestimated the effectiveness of the intervention on Democrats relative to the actual effect, while they were more accurate when predicting results for Republicans.

We also found differences in forecasting abilities across the three groups of forecasters. Both behavioral scientists and academics do better than laypeople. Nevertheless, none of them could predict the remarkably limited impact of nudges on mask-wearing.

Eugen Dimant
Prediction error across the three types of forecasters: laypeople, practitioners, and academics.
Source: Eugen Dimant

Bridging the Divide

Despite the moral framing behind them, none of the nudges persuaded Republicans or Democrats to wear masks, compared to a control condition. The interventions were not able to override the deep-seated partisanship in the U.S. The interventions may have been too weak for a strong context such as the COVID-19 pandemic, or may have happened too late, once political polarization over mask-wearing was already established. Moreover, neither behavioral scientists nor academics were able to predict the ineffectiveness of these messages, although they were still more accurate than the general population.

One final insight from our work is that—given the polarizing nature of this topic—nudges might simply not be enough and more social and structural changes are needed. As Richard Thaler put it: “Nudge is part of the solution to almost any problem, but is not the solution to any problem.”


Gelfand, M., Li, R., Stamkou, E., Pieper, D., Denison, E., Fernandez, J., … Dimant, E. (2021, September 24). Persuading Republicans and Democrats to Comply with Mask Wearing: An Intervention Tournament.

Dimant, Eugen and Pieper, Dylan and Clemente, Elena Giulia and Dreber, Anna and Gelfand, Michele Joy, Politicizing Mask-Wearing: Predicting the Success of Behavioral Interventions Among Republicans and Democrats (October 1, 2021). Available at SSRN: