Polling commissioned from YouGov for the Center for Countering Digital Hate’s recent report, The Anti-Vaxx Industry, surveyed attitudes in Britain and the US towards a future vaccine, as well as self-reported social media and legacy media consumption.
Data were also captured on a range of respondents’ sociodemographic characteristics and political values. We therefore conducted multiple regression analysis to assess more precisely the media consumption-vaccine attitudes link, taking potential confounders into account.
Because Covid-19 vaccine intentions were measured using a four-step response scale, we used an ordinal linear regression specification. Since a reasonably high proportion of respondents replied ‘don’t know’, we then repeated the analyses, this time grouping together those who said they would probably or definitely not have a vaccine with those who said ‘don’t know’.
The following findings are particularly interesting. In the British sample, older respondents are more likely to say they would probably or definitely be vaccinated than younger respondents. This may be driven by social generation rather than age itself: older generation members tend to be more trusting of health and government authorities, and they may also be more concerned to protect their health, particularly if they have chronic medical conditions.
We also found that those with Left political values (proxied by a measure of how the respondent voted in December 2019) are significantly more likely to say they would probably or definitely be vaccinated than those with Right-wing political values (proxied by voting Conservative in December 2019). Those who report they voted Leave rather than Remain are significantly more likely to give a more sceptical response in the first model, although when we include the don’t knows, this difference lost significance.
In each of these cases, it’s implausible that a particular vote choice causes differences in vaccine attitudes. It is more likely that differences in cultural and political values, on both social and economic questions, drive both electoral choices and vaccine attitudes. Ideally, future research should include a range of political values measures to test whether and how vaccine attitudes are becoming bound up with political ideology in Britain, as is already well-established for the US.
We also found important variation by ethnicity. The British sample was not large enough to distinguish between different minorities, and so we could only differentiate White and other than White respondents. We found that 30 percent of White respondents gave a more hesitant response or said that they did not know, compared with 43 percent of BAME respondents. In the full model of strength of hesitancy, BAME respondents remained more likely to give a more hesitant response controlling for third variables. In the second model, where we group the relatively hesitant with the don’t knows, the difference between BAME and White respondents was shy of significance at the conventional levels. However, given our small sample size for respondents who are not White, this is not strong evidence that no difference exists. If a difference does exist, it may be that BAME respondents are more likely to perceive that they have already had the virus and therefore have no need of a future vaccine, but it may also reflect greater concern or lower trust in scientific professionals. It suggests that public health communicators need to think urgently about how to connect with members of different communities.
We conducted similar analyses on the US data, although with slightly different measures relevant to the American context. In these models, we do not observe significant age effects, although we do observe that female respondents are on average more vaccine-hesitant than men, controlling for all other variables.
There are significant regional effects: those living in the Northeast appear less vaccine-hesitant than those in the Midwest, South or West, differences which are statistically-significant. We also find that African Americans are more vaccine-hesitant on these measures than White Americans. This may reflect historical distrust of medical professionals, exemplified by the notorious Tuskegee study run between 1932 and 1972.
Moreover, those who report they voted for Hillary Clinton in 2016 are significantly less vaccine-hesitant than those who supported a third candidate or did not vote. Those who report they voted for Donald Trump are significantly more vaccine-hesitant in the first model, although the result loses significance in the second. We repeated the analyses replacing vote choice with a political ideology measure, and in both cases (the ordinal logistic regression modelling strength of hesitancy, and the binary logistic regression modelling choosing either a more hesitant or ‘don’t know’ response) higher self-rated conservativism is positively and significantly associated with hesitancy.
Again, it is an open question as to whether partisanship causes differences in vaccine attitudes. Political identity has become increasingly central to people’s core sense of who they are, particularly in the US. Professor Matthew Hornsey at the University of Queensland, together with co-authors, has recently published research suggesting that a combination of political conservativism and conspiracist ideation contributes to vaccine concern. Moreover, they find evidence for political leadership making a difference in shaping, and not only reflecting, vaccine attitudes.
In the US case, it is also notable that more frequent legacy media Covid-19 news consumption appears to do little to counteract the distrust associated with more frequent social media news consumption. This provides a contrast with the British sample, where legacy media consumption appears more protective of vaccine confidence.
The difference may arise because the legacy media ecosystems are different in both contexts. It may be that media coverage of Covid-19 news is more politicised in the US where the lines between hard news reporting and entertainment and editorial have become more blurred, and is relatively more potent in influencing opinion than is yet the case in Britain.
Earlier research offers answers. R. Kelly Garrett, Professor of Communication at Ohio State University, together with co-authors Brian Weeks and Rachel Neo, conducted research suggesting that online media sources encouraged consumers to hold to their beliefs regardless of evidence, even if the reader or viewer knows and understands this evidence. Holding to beliefs inconsistent with evidence or expert advice may be expressive, or shore up a sense of personal group identity.
The challenge for public health communicators is to take the importance of identity for misperceptions seriously, while media regulators should attend to how these misperceptions are generated and disseminated.
Please note this is a preprint, so it is a preliminary piece of research that has not yet been through peer review and has not been published in a scientific journal.
Alsan, Marcella, and Marianne Wanamaker. “Tuskegee and the Health of Black Men.” The Quarterly Journal of Economics 133, no. 1 (February 1, 2018): 407–55. https://doi.org/10.1093/qje/qjx029.
Garrett, R. Kelly, Brian E. Weeks, and Rachel L. Neo. “Driving a Wedge Between Evidence and Beliefs: How Online Ideological News Exposure Promotes Political Misperceptions.” Journal of Computer-Mediated Communication 21, no. 5 (September 1, 2016): 331–48. https://doi.org/10.1111/jcc4.12164.
Hornsey, Matthew J., Matthew Finlayson, Gabrielle Chatwood, and Christopher T. Begeny. “Donald Trump and Vaccination: The Effect of Political Identity, Conspiracist Ideation and Presidential Tweets on Vaccine Hesitancy.” Journal of Experimental Social Psychology 88 (May 1, 2020): 103947.
Mason, Lilliana. “Ideologues without Issues: The Polarizing Consequences of Ideological Identities.” Public Opinion Quarterly 82, no. S1 (April 11, 2018): 866–87.
McAndrew, Siobhan, and Daniel Allington. “Mode and Frequency of Covid-19 Information Updates, Political Values, and Future Covid-19 Vaccine Attitudes.” PsyArXiv. August 14 2020. https://psyarxiv.com/j7srx.