Mask wearing: What predicts the first active move?
In the first mask scenario, participants first had to decide whether to take their mask off or leave it on. The results of a binary logistic regression showed a significant predictive value of five variables for keeping the mask on (see Table S4): i) relationship with the interacting partner (if the interaction partner is a potential co-worker, not a family member, odd’s ratio = 5.357), ii) higher empathy with those vulnerable to the pandemic (odd’s ratio = 1.083), iii) higher germ aversion (odd’s ratio = 1.061), iv) higher age (odd’s ratio= 1.022), and v) higher disbelief in protection of masks (odd’s ratio= 0.925). Further, country of residence had a significant predictive value, where Spain, which was most affected by the pandemic at the time of data collection, was used as a reference group. Specifically, residence in Poland (odd’s ratio = 0.228) and the UK (odd’s ratio = 0.187) decreased the probability of keeping the masks on. While most of the above variables positively predicted keeping the mask on, v) negatively predicted mask-wearing (Table 1). Participant’s sex, belief in COVID-19 related fake news, perceived own infectability, and prior information about the interaction partner had no significant predictive value. Figure S1A, B depicts percentages of participants who decided to take their mask off across countries.
The model (χ2 = 186.889, p < .001) predicts 79.8% of the analyzed cases correctly.
Table 1: Summary of predictors of scenario outcomes
Scenario
|
Outcome
|
Positive Predictors
|
Negative Predictors
|
Not significant predictor
|
Mask scenario I: Active first move
|
Leaving on the mask (safe, cooperative) as opposed to taking it off (putting partner at risk)
|
Interaction is with stranger
Empathy
Germ aversion
Higher age
|
Disbelief in protection of masks
Residence in PL, UK as compared to Spain
|
Infectability
Prior information about partner
Residence in other countries
Sex
Belief in fake news
|
Mask scenario II: Responding to partner taking off their mask
|
Keep on own mask regardless of the other person being without mask (mostly beneficial to partner) as opposed to both partners taking off the mask (equally unsafe).
|
Empathy
Germ Aversion
Higher age
|
Interaction is with family member
Disbelief in protection of masks
Residence in CZ, PL, UK, AT as compared to Spain
|
Infectability
Prior information about partner
Sex
Belief in fake news
|
Demand the partner to put their mask back on (equally safe outcome) as opposed to both partners taking off the mask (equally unsafe).
|
Empathy
Germ aversion
Higher age
|
Interaction is with family member
Disbelief in protection of masks
Residence in CZ, PL, UK, AT as compared to Spain
|
Infectability
Prior information about partner
Sex
Belief in fake news
|
Vaccine scenario
|
Intention to take the vaccine
|
Empathy
|
Financial investment and side effects
Disbelief in protection of masks
Belief in fake news
|
Immunization status of surrounding group
Country
Sex
Germ aversion
Infectability
Age
|
Mask wearing part II: How to react to the partner’s move
In the second part of the mask scenario, participants’ had to respond to their interacting partner taking off the mask. They could choose to A) keep their mask on without protesting, B) take their mask off as well, C) protest and demand the partner to put the mask back on, or D) demand the partner to put the mask back on while taking their own mask off. Here, only a small proportion (< 5 % of answers) of participants chose the option of putting only the other person at risk (D). Therefore, we omitted the category in the model to reduce data dispersion. For descriptive results on that category, see Fig. S1 C, D.
The results of a multinomial logistic regression (full model:χ2 = 217.648, p < .001) with the most unsafe option (both individuals taking off their masks) as a reference group revealed six variables that were significant predictors of the overall outcome behavior: i) higher disbelief in protection of masks (χ2 = 69.172, p < .001), ii) whether the interaction partner is a family member (χ2 = 38.394, p < .001), iii) country of residence (χ2 = 36.5, p < .001), iv) germ aversion (χ2 = 38.394, p < .001), v) age (χ2 = 24.010, p < .001), and vi) empathy with those most vulnerable to the pandemic (χ2 = 10.626, p = .005). Participants’ sex, belief in COVID-19 related fake news, and prior information given about the interacting partner had no significant predictive value. Figure S1C, D depicts the percentage of participants choosing each of the four possible moves. We now continue to characterize predictors of each possible reaction.
Predictors of the choice: Keeping the mask on while tolerating the partner without their mask
The option of keeping the own mask while tolerating the interacting partner wearing no mask had six significant predictors in our model as compared to the riskiest option (both participants taking off their mask) (Table S5).
The three variables i) higher empathy with those vulnerable to the pandemic (odd’s ratio= 1.078), ii) higher germ aversion (odd’s ratio= 1.034), and iii) higher age (odd’s ratio= 1.028) positively predicted choosing this option, whereas iv) higher disbelief in protection of masks (odd’s ratio= 0.929), and v) if the interaction partner was a family member (odd’s ratio= 0.373) negatively predicted this choice as opposed to the reference group. Further, vi) the participants’ country of residence had a predictive value; specifically, residence in Austria (odd’s ratio= 0.454), Poland (odd’s ratio= 0.228), Czech Republic (odd’s ratio= 0.253), and United Kingdom (odd’s ratio= 0.0200) negatively predicted the choice of this option when compared to Spain as the most affected country.
Predictors of the choice: Protesting the partner’s unsafe action
We identified six significant predictors for the behavioral outcome of protesting the partner’s action (Table S6): i) higher empathy with those vulnerable to the pandemic (odd’s ratio= 1.163), ii) higher germ aversion (odd’s ratio= 1.086), and iii) higher age (odd’s ratio= 1.051) positively predicted choosing this option, whereas iv) higher disbelief in masks as protections (odd’s ratio= 0.917), and v) if the interaction partner was a family member (odd’s ratio= 0.203) negatively predicted this choice as opposed to the reference group. In addition, vi) country of residence had a predictive value, where residence in Austria (odd’s ratio= 0.300), Poland (odd’s ratio= 0.213), Czech Republic (odd’s ratio= 0.127), and United Kingdom (odd’s ratio= 0.098) negatively predicted the choice of this option as compared to Spain.
Decision motives to wear or not to wear masks
Participants were given the option to explain their choices in the scenarios in their own words (n= 322 explanations given). We broadly classified the participants who gave these answers into three groups based on the participant’s choice in the scenarios: those always wearing a mask (n= 239 explanations given), ‘switchers’ who wear the mask with the stranger but not with the cousin (n= 61 explanations given), and those never wearing a mask (n =29 explanations given). (Table 2)
For the group always wearing their masks, 88% and 87% of explanations given for decisions with the cousin and stranger respectively mentioned one of seven main motives: i) safety of specific individuals (cousin: 24%; stranger: 30%), ii) general safety concerns (cousin: 18%, stranger: 15%), iii) avoiding conflict and practicality of communication (cousin: 13%, stranger: 9%), iv) personal freedom and individual choice (cousin: 9%, stranger: 6%), v) generalized responsibility during a pandemic (cousin: 8%, stranger: 7%), vi) professionalism at the workplace and legal regulations (cousin: 7%, stranger: 13%), and vii) knowing the interaction partner, or lack of prior knowledge about the interaction partner (cousin: 5%, stranger: 8%). The first and most abundant motive, safety of specific individuals (i), often included safety of the participants themselves (cousin: 93%, stranger: 89%), but also mentions of the partner in the current interaction (cousin: 39%, stranger: 43%), and of others outside the interaction (e.g. elderly grandparents one could carry the disease to; cousin: 25%, stranger: 31%) (subcategories non-mutually exclusive).
For those wearing the mask only in one scenario, typically with the stranger, 86% (stranger) and 77% (cousin) of all answers contained one of the following motives: i) generalized trust in family members (43%; only present in cousin condition), ii) knowing the interaction partner, or lack of prior knowledge about the interaction partner (cousin: 13%, stranger: 33%), iii) ease of social interaction (cousin: 6%, stranger: 16%), and iv) professionalism at the workplace and legal regulations (stranger: 18%, cousin: 3%). Further, some less abundant decision motives only occurred in either the stranger or cousin scenario. In the interaction with the stranger, these were: i) safety of specific individuals (6%), ii) avoidance of conflict (6%), and iii) generalized safety considerations. In the interaction with the cousin, physical discomfort with the mask was identified as the main motive in an additional 8% of the answers.
For the group never choosing to wear a mask, 96% (both cousin and stranger) of the responses were assigned one or multiple of the following three main motives: i) rationalization (e.g. that the situation will be safe or unsafe regardless of the mask; cousin: 48%, stranger: 33%), ii) ease of social interaction (both: 29%), and iii) physical discomfort of masks (cousin: 19%, stranger: 24%). Only for the interaction with the stranger, an additional 10% of explanations contained social reciprocity and joint agreement of both persons as the main decision motive.
Predictors of vaccination intention
Following the mask scenarios, participants were presented with a hypothetical vaccination scenario (see 2.5.). The result of the univariate general linear model (F = 106.568, p = .001, ηp2 = .268) provides us with four significant predictors of vaccination intention (see Table S7): i) the cost of the vaccine (high financial investment and side effects versus neither) (F = 25.162, p < .001, ηp2 = .080), ii) disbelief in mask wearing as a protective measure (F = 17.328, p < .001, ηp2 = .056), iii) belief in COVID-19 related fake news (F = 16.235, p < .001, ηp2 = .053), and iv) empathy with those vulnerable to the pandemic (F = 9.51, p = .002, ηp2 = .032). Immunization status of the surrounding social group (hypothetical co-workers and friends) had no effect, nor did sex, country of residence, the participants’ age or the scales germ aversion and infectability.
Decision motives of groups with high or low vaccination intention and undecided groups
A subset of participants (n=171) provided explanations in their own words for their decisions in the vaccine scenario. The answers were divided to three sub-groups based on the score indicating vaccination intention: pro-vaccine (scores from 61 to 101, n = 121 explanations given), undecided (scores from 40 to 60, n = 24 explanations given), and anti-vaccine, with scores from 1 to 39 (n = 26 explanations given). (Table 2)
For the pro-vaccine group, 76% of all answers contained one or multiple of five main decision motives: i) safety of specific individuals (26%; of these, 97% included protection of the participant, 68% also included protection of others), ii) enabling a return to ‘normality’ (18%), iii) trust in science and official institutions (13%), iv) community-level altruistic motives (10%), and v) trade-off considerations about benefits and risks of vaccines (9%). The remaining 22% of answers were diverse and could not be classified in a more abundant category.
For undecided individuals, 62.5% of all answers mentioned one or multiple of three main motives: i) doubts about the efficiency and safety of the vaccine (40%), ii) safety from the disease (12%), and iii) considerations referencing trust in science and official institutions (12%). The remaining answers could not be classified into categories amounting to more than 5% of all answers.
For the group with the lowest vaccination intention, 77% of the answers could be classified as one of these four motives: i) generalized distrust in vaccines (27%), ii) cost of the vaccine in the presented scenarios (23%), iii) doubts about the effectiveness of vaccines (13%), and iv) fear of side-effects (10%). The remaining answers were diverse
Table 2: Summary of decision motives reported in open-ended answers
Scenario
|
Group based on responses in scenarios
|
Top three decision motives
(% of answers categorized as such)
|
Other decision motives
(% of answers categorized as such)
|
Mask scenarios
|
Always wearing masks
|
Safety of specific individuals
(C: 24%, S: 30%)
Generalized safety concerns
(C: 18%, S: 15%)
Avoiding conflict and practicality of communication
(C: 13%, S: 9%)
|
Personal freedom and individual choice
(C: 9%, S: 6 %)
Generalized responsibility during a pandemic
(C: 8%, S: 7%)
Professionalism at the workplace and legal regulations
(C: 7%, S: 13%)
Knowing the interaction partner, or lack of prior knowledge
(C: 5%, S: 8%)
|
Wearing mask in only one scenario (typically with the stranger)
|
Generalized trust in family members
(C: 43%)
Knowing the interaction partner, or lack of prior knowledge
(C: 13%, S: 33%)
Ease of social interaction
(C: 6%, S: 16%)
|
Professionalism at the workplace and legal regulations
(C: 3%, S: 18%)
|
Never wearing mask
|
Rationalization
(C: 48%, S: 33%)
Ease of social interaction
(C: 29%, S: 29%)
Physical discomfort of masks
(C: 19%, S: 24%)
|
|
|
High vaccination intention
|
Safety of specific individuals (26%; of these, 97% included protection of the participant, 68% also included protection of others)
Enabling a return to ‘normality’ (18%)
Trust in science and official institutions (13%),
|
Community-level altruistic motives (10%)
Trade-off considerations about benefits and risks of vaccines (9%)
|
Vaccine scenario
|
Undecided
|
Doubts about the efficiency and safety of vaccination (40%)
Safety from the disease (12%)
Considerations referencing trust in science and official institutions (12%)
|
|
|
Low vaccination intention
|
Generalized distrust in vaccines (27%)
Cost of the vaccine in the presented scenarios (23%)
Doubts about the effectiveness of vaccines (13%)
|
Fear of side-effects (10%)
|
Note: C= in cousin scenario, S= in stranger scenario.
Vaccination versus mask-wearing and testing intentions in different situations.
Next, we sought to describe participants’ approval for vaccination or mask-wearing and testing in order to attend certain events. Here, participants could indicate on a slider if they would rather wear masks and undergo testing whenever attending one of the listed events or if they would rather get vaccinated against SARS-CoV-2 in order to pursue the activities. For all countries and situations, on average, participants leaned towards vaccination (Table 3). A single ANOVA test of differences in favoring of the long- or short-term solution between countries and events revealed no significant results.
Table 3. Mean ± SD indicating the preference to wear masks and get tested to attend an event (lower values) or to get vaccinated (higher values) on a slider (1 to 101). The values in parenthesis indicate the proportion of individuals who scored above 51 (lean towards vaccination).
Country
|
Appointment with a medical practitioner
|
Work-place with possibility to spread the disease
|
Traveling abroad
|
Eating out (restaurant)
|
Social event (e.g. concert)
|
CZ
|
61±43 (57%)
|
76±39 (73%)
|
78±35 (73%)
|
70±41 (70%)
|
69±42 (66%)
|
PL
|
72±38 (74%)
|
75±35 (81%)
|
76±36 (79%)
|
78±36 (79%)
|
73±39 (75%)
|
UK
|
72±40 (71%)
|
77±36 (77%)
|
78±39 (75%)
|
74±39 (74%)
|
80±33 (82%)
|
ES
|
75±37 (69%)
|
86±30 (87%)
|
88±28 (89%)
|
82±32 (73%)
|
84±33 (78%)
|
AT
|
76±38 (74%)
|
85±31 (84%)
|
86±31 (82%)
|
82±35 (78%)
|
85±32 (86%)
|