Roles Conceptualization, Formal analysis, Writing – original draft, Writing – review & editing * E-mail: christoph.schmidt-petri@kit.edu (CSP); cschroeder@diw.de (CS) Affiliation Karlsruhe Institute of Technology, Karlsruhe, Germany ⨯
Roles Conceptualization, Formal analysis, Writing – original draft, Writing – review & editing * E-mail: christoph.schmidt-petri@kit.edu (CSP); cschroeder@diw.de (CS) Affiliations DIW Berlin / SOEP, Berlin, Germany, Freie Universität Berlin, Berlin, Germany
Several vaccines against COVID-19 have now been developed and are already being rolled out around the world. The decision whether or not to get vaccinated has so far been left to the individual citizens. However, there are good reasons, both in theory as well as in practice, to believe that the willingness to get vaccinated might not be sufficiently high to achieve herd immunity. A policy of mandatory vaccination could ensure high levels of vaccination coverage, but its legitimacy is doubtful. We investigate the willingness to get vaccinated and the reasons for an acceptance (or rejection) of a policy of mandatory vaccination against COVID-19 in June and July 2020 in Germany based on a representative real time survey, a random sub-sample (SOEP-CoV) of the German Socio-Economic Panel (SOEP). Our results show that about 70 percent of adults in Germany would voluntarily get vaccinated against the coronavirus if a vaccine without side effects was available. About half of residents of Germany are in favor, and half against, a policy of mandatory vaccination. The approval rate for mandatory vaccination is significantly higher among those who would get vaccinated voluntarily (around 60 percent) than among those who would not get vaccinated voluntarily (27 percent). The individual willingness to get vaccinated and acceptance of a policy of mandatory vaccination correlates systematically with socio-demographic and psychological characteristics of the respondents. We conclude that as far as people’s declared intentions are concerned, herd immunity could be reached without a policy of mandatory vaccination, but that such a policy might be found acceptable too, were it to become necessary.
Citation: Graeber D, Schmidt-Petri C, Schröder C (2021) Attitudes on voluntary and mandatory vaccination against COVID-19: Evidence from Germany. PLoS ONE 16(5): e0248372. https://doi.org/10.1371/journal.pone.0248372
Editor: Valerio Capraro, Middlesex University, UNITED KINGDOM
Received: October 19, 2020; Accepted: February 25, 2021; Published: May 10, 2021
Copyright: © 2021 Graeber et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Our analyses rely on the German Socio-Economic Panel (SOEP), an independent scientific data infrastructure established in 1984. We, as users, cannot send the data to the journal and make them publicly available, as this is against SOEP's statutes (and most likely against the statutes of all providers of micro data). However, this should not be a hurdle, as researchers from scientific institutions around the globe can access the data (free of costs) once they have signed a user contract. The scientific use file of the SOEP with anonymous microdata is made available free of charge to universities and research institutes for research and teaching purposes. The direct use of SOEP data is subject to the provisions of German data protection law. Therefore, signing a data distribution contract is the single precondition for working with SOEP data. The data distribution contract can be requested with a form which can be downloaded from: http://www.diw.de/documents/dokumentenarchiv/17/diw_01.c.88926.de/soep_application_contract.pdf.
Funding: The data collection of the SOEP-CoV Study was financially supported by the German Federal Ministry of Education and Research. We acknowledge support by the KIT-Publication Fund of the Karlsruhe Institute of Technology. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Great efforts have been made worldwide to develop a vaccine against COVID-19. When we first drafted this article, in October 2020, 35 different potential vaccines were in clinical trials and 145 were still in the pre-clinical stage. In February 2021, several vaccines have been approved in many countries and are being rolled out, 74 are in clinical trials, and 182 are in the pre-clinical stage [1].
These developments are very encouraging, as a wide availability of vaccines is seen by many as a prerequisite for a return to a “normal” pre-COVID-19 type of social and economic life. With the growing availability of vaccines comes the hope that coercive measures such as restrictions on international trade, contact restrictions, and travel bans, etc., which cause enormous economic and social costs, may soon be removed and will not need to be reimplemented.
Of course, any vaccine is only an effective contribution to a return to normal life if a sufficiently high number of people are actually vaccinated, yielding herd immunity. If so, vaccination secures a public good: protection from COVID-19 for everyone. From a microeconomic perspective, this raises a well-known problem, free-riding: If the vaccination is freely available but not obligatory, then citizens’ individual decisions determine the extent to which this public good is made available. In order to make that decision, they will weigh their own costs and benefits. These costs include the time sacrificed, physical unpleasantness, possible side effects of a vaccination, etc. The benefits to a particular individual are primarily, but not necessarily exclusively, the reduction in risk to that person’s own health or material well-being. From a welfare perspective, if individuals do not take into account the positive externalities on third parties that their own vaccination triggers, there will be an undersupply of the public good. Following [2, 3], individuals’ utility function may also include other-regarding preferences and hence yield a direct benefit from contributions to a public good. In our context, people could therefore benefit from a ‘warm glow of vaccinating’, because by vaccinating themselves they also reduce the risks of others. But even so there is certainly no guarantee that the social optimum will be reached [4] or that a sufficiently high number of people will freely choose to get vaccinated.
It is frequently argued that vaccination should be made mandatory because of the free-rider problem [5]: While vaccinated individuals have incurred private costs in terms of discomfort or money and receive the private benefit of a reduced risk of getting the disease, the major collective benefit, the reduced incidence of disease, is public. If enough other people produce the public benefit, and the circulation of the virus decreases accordingly, an individual might rationally decide to free-ride on others’ decisions. A policy of mandatory vaccination would prevent this.
[6] argue that such a policy would not be necessary: “If vaccinations are perfect, then if one is vaccinated he or she does not care whether others are vaccinated, so there is no longer any public good problem” ([6], p. 70). Hence there would not be a case in favor of mandatory vaccination, as under such a policy, individuals who would have favored not to be vaccinated are made worse off, while those who anyway would get vaccinated are not better off.
However, by definition, ‘perfect’ vaccination means that everyone vaccinated is perfectly immune [6]. In the current situation, it can neither be taken for granted that a perfect vaccination is being or will be provided soon, nor that everyone who wants to also will have the possibility to be vaccinated (both financially and in terms of health). If perfect vaccination is not feasible, however, mandatory vaccination is not dominated by a laissez faire solution [6].
Extensions of this theoretical public good analysis emphasize the relevance of behavioral aspects not typically considered in classical models. The empirical literature also highlights a number of factors that matter for vaccine uptake. For instance [7], show that social norms matter for an individual’s willingness to get a vaccination and that such norms can suppress vaccine uptake even in the presence of frequent disease outbreaks. Further [8], show that the design of public vaccination policies should also take intergroup interactions into account. Other-regarding preferences can explain voluntary vaccination uptake, as argued by [9]. For example [10], show that the presence of individuals who cannot get vaccinated, like babies and the elderly, increases the willingness to get vaccinated. The static model in [6] also does not reflect interactive processes [9, 11]. show that vaccination is the individually best response until a certain vaccination rate is reached in the population and becomes a social dilemma only from this vaccination rate until herd immunity is maximized. Communicating the social benefits of vaccination can have positive effects, particularly when this protects vulnerable groups, but it can also invite free-riding [12]. Those people who cannot get vaccinated themselves for medical reasons are particularly vulnerable: they cannot protect themselves even if they wanted to and, hence, depend on their fellow citizens to protect them by preventing the spread of the virus through their vaccination. Children, too, need to be considered separately. Since they cannot give informed consent to a voluntary vaccination themselves, they might have to be protected from their parents (who might be unwilling to get them vaccinated) in case of particularly serious diseases (see [13, 14]).
There is, in summary, hope that the public goods problem may be overcome, as social and behavioral science offers a wide array of potential policy options to influence people’s perceptions and reactions to the pandemic (for an extensive up-to-date overview, see [15]). It is not clear, however, how the research on well-established vaccines carries over to the current pandemic, and recent developments seem to indicate that the willingness to get vaccinated against the novel coronavirus is currently rather low. We therefore chose to investigate two fundamental questions at the opposite extremes of the spectrum of policy options: would a sufficient number of people voluntarily undergo vaccination to achieve herd immunity? Or would a mandatory vaccination against COVID-19 be acceptable to achieve herd immunity?
A legal duty to be vaccinated against COVID-19 could be an alternative to other coercive measures if one assumes that a high-risk, unregulated, laissez faire approach is not a realistic policy option: it seems irresponsible to lift all restrictions because the virus would soon spread through the entire population. Coercive measures of some kind therefore seem inevitable. Mandatory vaccination could be preferable to other coercive measures, provided the interference with bodily integrity would be considered less socially costly in the long run than the effects of prolonged lockdowns. Emotions run high where vaccination policies are concerned, but because mandatory vaccination might become a realistic scenario, it is worth investigating what the general population thinks about such a policy.
It is important to emphasize that a legal duty to vaccinate against COVID-19 would not imply a legal (or even moral) duty to vaccinate against other diseases. The novel coronavirus is a special case in many respects: In contrast to influenza, for example, the population does not have a background immunity from past infections. In addition, many infected people do not show symptoms (a recent meta-study estimates this to be one in six infected [16]) and, hence, cannot protect others from being infected through voluntary self-quarantining. Thus, people with COVID-19 represent a much higher risk of infection for others than, for example, people who come down with influenza, assuming that these would normally stay at home. Therefore, a vaccination against COVID-19 is much more important from the social perspective than e.g. a vaccination against influenza: not for self-protection, but to protect other people from unintentional infection. Although classic liberal positions (cf. [17]) would reject a paternalist legal obligation to protect oneself through vaccination, they plausibly would favor a policy of mandatory vaccination in the case of COVID-19 to protect others from being harmed. In modern philosophical discussions, even some libertarians are in favor of mandatory vaccination against serious diseases for similar reasons (see [18] and for an overview [19]).
Though there are philosophical reasons supporting a policy of mandatory vaccination, we want to emphasize that we are not advocating it as a concrete policy option for Germany at this moment. Our aim is to understand whether the general public would consider such a policy acceptable, or which sections of the population, and why. To this end, we study the willingness to get vaccinated and the acceptance of a policy of mandatory vaccination against COVID-19 in June and July 2020 in Germany. We use unique real time survey data from a sub-sample (SOEP-CoV) of the German Socio-Economic Panel (SOEP, see [20]). A set of questions about vaccination was part of the later stages of SOEP-CoV, an ongoing research project initiated in April 2020. This so-called ‘vaccination module’ included questions on the willingness to get vaccinated voluntarily and the acceptance of a policy of mandatory vaccination against COVID-19. In addition, individuals could indicate reasons for their preference regarding the second question. Using the rich data of the SOEP, pre-pandemic income, education, household context, personality, political preferences etc., which can be directly linked with SOEP-CoV, we are able to provide a detailed picture on who intends to get vaccinated and who does not.
The most important result of our study is that about 70 percent of adults in Germany would get vaccinated voluntarily against COVID-19 if a vaccine without significant side effects was available. Further, about half of adults in Germany are in favor, and half against, a policy of mandatory vaccination against COVID-19. The approval rate for mandatory vaccination is significantly higher among those who would get vaccinated voluntarily (around 60 percent) than among those who would not get vaccinated voluntarily (27 percent). However, 22 percent of the individuals would disapprove of both a voluntary and a mandatory vaccination and 8 percent can be characterized as ‘passengers’ (they are not willing to get vaccinated but do support a policy of mandatory vaccination, but they might not all be ‘free-riders’ in the standard sense). In this group, surprisingly, 86 percent state that, without a mandatory vaccination, too few individuals would get vaccinated and about 87 percent indicate that most people underestimate how dangerous COVID-19 is. In general, the willingness to get vaccinated is significantly lower for female, younger, and less educated respondents as well as those with lower income. A policy of mandatory vaccination is rejected with higher probability by women and favored by older people and those living in the eastern federal states.
The German Socio-economic Panel (SOEP) is among the largest and longest-running representative panel surveys worldwide and is recognized for maintaining the highest standards of data quality and research ethics [20]. In 2020, the survey covers about 30,000 adults in 20,000 households. Since the same individuals and households participate in the study every year, life courses of the respondents can be tracked and intertemporal analyses can be carried out at the individual and at the household level. The data contain information on the respondents’ household situation, education, labor market outcomes, and health, among others (see [20, 21]).
To better understand the effects of the corona pandemic, a special survey called SOEP-CoV was conducted within the framework of the SOEP, which consisted of a random sample of about 6,700 SOEP respondents, (see [21, 22]). SOEP-CoV was surveyed in nine staggered tranches from early April to the end of July 2020 and collected data on the following topics: a) Prevalence, health behavior, and health inequality; b) Labor market and gainful employment; c) Social life, networks, and mobility; d) Mental health and well-being; and e) Social cohesion. Over time, some new question modules were introduced within these five thematic complexes. These included the ‘vaccination module’ (see questionnaires available under www.soep-cov.de/Methodik/).
The ‘vaccination module’ went into the field with tranches 7 to 9, in June and July 2020, and covered a total of 851 persons aged 19 years and older. At that moment, major research efforts were being undertaken, but it was not clear whether any vaccine would actually be found. The module hence starts with a question on the hypothetical willingness to get vaccinated against COVID-19:
Argument 1: Others’ willingness to get vaccinated without mandatory vaccination
Argument 2: Misperception of risks
Argument 3: Legitimacy of a policy of mandatory vaccinations in general
Argument 4: Other reasons (without listing these reasons explicitly)
The first three arguments are of particular relevance for political decision-making. Although there is quite a lot of research on the reasons people have not to get vaccinated themselves, there is much less research on what people think about policies of mandatory vaccinations, and up to present–at least to our knowledge–none on the application to the special case of the novel coronavirus. As the reasons for the individual decision need not carry over to the policy assessment, and given the previously discussed particularities of the coronavirus, we focused on factors that are both of theoretical importance and under discussion in the general public. It would be interesting, for instance, if many people did not have the intention to get vaccinated themselves, yet believed that enough other people would get vaccinated so that mandatory vaccination would not be required. Similarly, it would be surprising if people wanted to get vaccinated yet believed that others overestimated the dangerousness of the virus. Finally, we wanted to see whether people considered mandatory vaccinations potentially legitimate at all.
Since SOEP-CoV is a random sample from the SOEP population, the SOEP-CoV data 2020 can be linked with the regular SOEP data of previous years. Thus, attitudes toward vaccination against COVID-19 that were collected during the pandemic can be linked to the characteristics of the respondents before the outbreak of the pandemic (e.g., income or educational level). Since these characteristics were collected before the pandemic, they can be considered unaffected by the pandemic event and, hence, exogenous (S1 File provides definitions of all dependent and independent variables used in the empirical analyses).
The response rate in the vaccination module was high. Altogether, only 4.58 percent of the 851 respondents did not answer the question about voluntary vaccination and 3.41 percent did not answer the question about mandatory vaccination. Of those who supported (objected to) mandatory vaccination, 0.26 (1.82) percent did not provide at least one motive in the follow-up question. Hence, bias from item non-response should be small and we did not correct for it. As the focal variables are coded dichotomously (yes = 1; no = 0), there was no need to remove outliers in them from the database.
To derive population-wide estimates, the SOEP-CoV data is equipped with frequency weights. The weighting of SOEP-CoV follows the standard weighting used in SOEP [23, 24]. Based on the SOEP household weights, weights for all persons in the participating households were generated via a marginal adjustment step and corrected for selection effects. Furthermore, the data were corrected for the fact that some SOEP subsamples were excluded from the SOEP-CoV study from the outset. To address potential selection effects and adjust frequency weights accordingly, we followed the two-step procedure recommended in [25]:
Since the vaccination questions are answered once by each respondent, our empirical strategy is between-person. Uni- and bivariate results for our focal variable, attitudes toward vaccination, are presented as weighted means or percentages. Assessments of differences in attitudes or characteristics between-groups rely on two-tailed t-tests, with statistical significance evaluated at pp p
The multivariate analyses to statistically explain vaccination preferences rely on logit regressions. The purpose of the logit regressions is to estimate the probability that a respondent, i, with a certain set of characteristics, Xi, will fall into a specific one of the two categories. It can be understood as finding the β coefficients that best fit (1) where y is a is a zero-one dummy. We use the logit model to explain attitudes towards both voluntary and mandatory vaccinations. The dummy yi is set equal to one if a respondent supports voluntary, respectively mandatory, vaccination. Stata fits the logit model using the standard Maximum Likelihood estimator. It considers the binary nature of the outcome variable assuming that a) the logit link function is correct, b) the model is correctly specified, and c) observations are independent. The robust variance estimator which we use below is robust to assumptions a) and b). Since the logit regressions are supposed to describe structures in the data, we do not use sample weights. For all our empirical analyses, we used Stata version 15.
For the questions on voluntary vaccination (A) and mandatory vaccination (B), four groups in the population may be distinguished:
Overall, 70 percent of adults in Germany would voluntarily get vaccinated against the coronavirus, provided a vaccine without significant side effects was available (Table 1: groups 2 and 4). This value corresponds exactly to the results of [26]. From May till September 2020, the COVID-19 snapshot monitoring (COSMO) at the University of Erfurt showed relatively constant values of between 60 and 66 percent; it was only in April that it showed an exceptionally high value of 79 percent, and it has now decreased further (cf. [27], p. 76; an overview of previous studies on the willingness to get vaccinated in Germany is provided in S2 File.). Overall, these studies paint a consistent picture, with a slight decline in the willingness in the second half of 2020.