Asymmetric affective polarization regarding COVID-19 vaccination in six European countries – Nature.com
July 11, 2024
A. Measuring opinion-based affective polarization regarding COVID-19 vaccination using thermometer scores
We assessed opinion-based affective polarization regarding COVID-19 vaccination in a multistep process. To begin with, we asked respondents to indicate on an 11-point scale ranging from 0 complete rejection to 10 complete support how strongly they oppose or support COVID-19 vaccination (Mean (M)Sample FRA=6.83, (M)Sample GER=7.8, (M)Sample ITA=7.7, (M)Sample SPA=7.96, (M)Sample SWI=7.26, (M)Sample UK=8.14).
The left panel in Fig.2 shows the mean levels of support for COVID-19 vaccination across our six countries. Generally, support is very high with mean levels of 7 on the scale from 0 to 10. The lowest level of support is found in France and the highest in the UK and Spain. The violin plots in the right panel of Fig.1 again illustrate this finding as the distribution is skewed in favor of COVID-19 vaccination. However, the distribution shows that a consistent minority opposes COVID-19 vaccination in all six countries. That is, the issue has triggered the formation of two groups: one for and one against it.
Mean support for COVID-19 vaccination and its distribution.
In a next step, we applied the most widely used measurement approach for affective polarization, the feeling thermometer8,40, in which respondents are asked to rate their feelings about a specific subject on a temperature scale9,40. First, we dichotomized the variable on the position regarding COVID-19 vaccination and classified respondents who indicated a value from 0 to 4 as opponents of COVID-19 vaccination (anti-vaccination group) and those who indicated a value from 6 to 10 as supporters of COVID-19 vaccination (pro-vaccination group). We excluded respondents who indicated a value of 5 as they were neutral on the issue. Second, we asked respondents of the respective groups to rate their feelings toward a) supporters of COVID-19 vaccination and b) opponents of COVID-19 vaccination on a scale from 5 (very cold and negative) to+5 (very warm and positive). We transformed this scale to range from 0 to 10 and subsequently used the absolute difference between the two ratings to obtain a measure for affective polarization ((M)Sample FRA=5.1, (M)Sample GER=6.05, (M)Sample ITA=6.12, (M)Sample SPA=6.44, (M)Sample SWI=5.32, (M)Sample UK=6.03). The descriptive results are reported below using bar graphs for readability. Formal t-tests are reported in the supplementary material, section C, Tables 1214.
Figure3 shows the thermometer scores for the two groups separated by group membership (supporters vs. opponents) and country. The feelings toward COVID-19 vaccination supporters (left panel in Fig.3) show a clear pattern: Supporters feel very positive and warm toward other supporters ((M)Supp FRA=8.9, (M)Supp GER=9.18, (M)Supp ITA=8.96, (M)Supp SPA=8.93, (M)Supp SWI=8.93, (M)Supp UK=9.05). Conversely, opponents feel somewhat cold and negative toward supporters, with values below the neutral value of 5 ((M)Opp FRA=4.86, (M)Opp GER=4.43, (M)Opp ITA=4.17, (M)Opp SPA=4.79, (M)Opp SWI=4.65, (M)Opp UK=4.47). These values are statistically significant and different at the 95% level.
Thermometer ratings of feelings toward COVID-19 vaccination supporters and opponents by group and country.Notes: Figure3 shows the mean thermometer ratings of feelings toward vaccination supporters and opponents separated by group and country, with 95% confidence intervals. Reading example for France in the left panel: In France, vaccination opponents have an average thermometer rating of feelings toward vaccination supporters of 6.15, while vaccination supporters have an average thermometer rating of feelings toward vaccination supporters of 7.04 on a scale of 010. The difference is statistically significant.
Looking at the thermometer scores for COVID-19 vaccination opponents, we see a mirror image. Opponents feel relatively positive and warm toward other opponents ((M)Opp FRA=6.55, (M)Opp GER=6.73, (M)Opp ITA=6.77, (M)Opp SPA=5.78, (M)Opp SWI=6.85, (M)Opp UK=5.96). Yet, these positive in-group feelings are comparatively lower among opponents than among supporters. It seems that there is less group cohesion among opponents than among supporters of COVID-19 vaccination. Furthermore, supporters express very cold and negative feelings toward opponents, as expected ((M)Supp FRA=2.46, (M)Supp GER=2.07, (M)Supp ITA=2.27, (M)Supp SPA=1.76, (M)Supp SWI=2.73, (M)Supp UK=2.32). All differences between supporters and opponents are statistically significant at the 95% level.
Figure4 reveals the absolute difference between both thermometer ratings, separated by group and country. Supporters show a relatively high average difference in feelings toward the in-group and the out-group ((M)Supp FRA=6.63, (M)Supp GER=7.17, (M)Supp ITA=7.01, (M)Supp SPA=7.35, (M)Supp SWI=6.46, (M)Supp UK=6.86). Opponents also show a difference in feelings for their in- and out-group, but this difference is less pronounced ((M)Opp FRA=2.65, (M)Opp GER=2.95, (M)Opp ITA=3.18, (M)Opp SPA=2.93, (M)Opp SWI=2.87, (M)Opp UK=3.02). All differences in affective polarization are significant at the 95% level. The highest levels of affective polarization are found in Germany and Spain among supporters and in Italy and the United Kingdom among opponents. Both groups express affective polarization regarding COVID-19 vaccination, but it is stronger among the pro- than among the anti-vaccination group. As expected, affective polarization is asymmetric, implying that both in-group attachment and out-group dislike are stronger among the pro-vaccination group than among the anti-vaccination group.
Affective polarization of COVID-19 vaccination supporters and opponents.Notes: Figure4 shows the mean level of affective polarization (thermometer measure) by group and country, with 95% confidence intervals. Reading example: In France, vaccination opponents show an average affective polarization of 2.65 and vaccination supporters of 6.63 on a scale from 0 to 10. The difference is statistically significant.
To further evaluate our findings, we used a second common measure of affective polarization: character trait ratings14,17,40. Here, respondents from the pro-vaccination camp and the anti-vaccination camp (as coded above) are asked to rate various character traits of the twodifferent groups. Although trait ratings are a typical measure of affective polarization, they reflect more than just negative affect but also shed light on the perceived stereotypical appearance of a group9. In this vein, these trait ratings allow us to identify whether respondents assign negative or positive characteristics to their respective in-group and out-group. Research has shown that trait ratings and thermometer scores, although conceptually somewhat distinct, correlate fairly well with each other and show little systematic differences40. In our full sample, the affective polarization scores for both measures correlate at (r)Sample=0.62 (r)Sample FRA=0. 61, (r)Sample GER=0.64, (r)Sample ITA=0.67, (r)Sample SPA=0.57, (r)Sample SWI=0.59, (r)Sample UK=0.65).
In our study, we asked respondents to rate the extent to which two positive character traits (openness to compromise and critical thinking) and two negative character traits (selfishness and narrow-mindedness) apply to a) supporters of COVID-19 vaccination and b) opponents of COVID-19 vaccination on a scale from 1 (does not apply at all) to 5 (fully applies). In addition to the assigned values, we also calculated the absolute differences between the scores assigned to the in- and the out-group for each trait. Subsequently, we combined these differences into an additive score for affective polarization ((M)Sample FRA=1.72, (M)Sample GER=2.08, (M)Sample ITA=1.92, (M)Sample SPA=1.99, (M)Sample SWI=1.87, (M)Sample UK=2.01).
This alternative measure reveals a similar picture of affective polarization regarding COVID-19 vaccination as the feeling thermometer. Figure5 is analogous to Fig.3 and shows the character trait ratings for supporters and opponents of COVID-19 vaccination by group and country. For the sake of readability, we combine the two positive and negative traits each (see supplementary material, section D, Fig.1 for the individual character trait ratings). The upper panel of Fig.5 shows the ratings of the two negative traits combined. As we can see in the upper left panel, supporters do not believe that other supporters are selfish and narrow-minded ((M)Supp FRA=1.87, (M)Supp GER=1.77, (M)Supp ITA=2.15, (M)Supp SPA=2.16, (M)Supp SWI=2.00, (M)Supp UK=1.68). Conversely, opponents tend to assign these negative traits to supporters ((M)Opp FRA=2.63, (M)Opp GER=3.00, (M)Opp ITA=3.33, (M) Opp SPA=2.79, (M)Opp SWI=3.04, (M)Opp UK=2.81). The opposite picture emerges when we look at the upper right panel: Consistent with in-group favoritism, opponents do not believe that other opponents are selfish or narrow-minded ((M)Opp FRA=1.87, (M)Opp GER=1.94, (M)Opp ITA=2.08, (M)Opp SPA=2.49, (M)Opp SWI=1.98, (M)Opp UK=2.23). Supporters, however, believe that opponents are selfish and narrow-minded ((M)Supp FRA=3.83, (M)Supp GER=4.10, (M)Supp ITA=4.08, (M)Supp SPA=3.94, (M)Supp SWI=3.96, (M)Supp UK=3.91). All differences are significant at the 95% level.
Perceived character traits of COVID-19 vaccination supporters and opponents by group and country.Notes: Figure5 shows the mean perceived character traits for vaccination supporters and opponents separated by group and country, with 95% confidence intervals. For example, in the top left-hand panel for France: In France, vaccination opponents perceive vaccination supporters to have negative traits with an average of 2.63 while vaccination supporters perceive vaccination supporters to have negative traits with an average of 1.87 on a scale of 15. The difference is statistically significant.
The lower part of Fig.5 shows the ratings of the positive traits combined. Here, an analogous but less consistent trend is observed compared to the ratings of the negative traits. Supporters assign positive traits to their in-group ((M)Supp FRA=2.99, (M)Supp GER=3.64, (M)Supp ITA=3.45, (M)Supp SPA=3.69, (M)Supp SWI=3.56, (M)Supp UK=3.79). Conversely, opponents do not ascribe these traits to supporters ((M)Opp FRA=2.61, (M)Opp GER=2.48, (M)Opp ITA=2.83, (M)Opp SPA=2.85, (M)Opp SWI=2.49, (M)Opp UK=2.78). All differences between supporters and opponents are significant at the 95% level.
Opponents see themselves as more open to compromise and able to think critically ((M)Opp FRA=2.80, (M)Opp GER=3.48, (M)Opp ITA=3.03, (M)Opp SPA=3.14, (M)Opp SWI=3.27, (M)Opp UK=2.86). Yet, supporters do not think that these positive traits apply to opponents ((M)Supp FRA=2.78, (M)Supp GER=2.21, (M)Supp ITA=2.59, (M)Supp SPA=2.50, (M)Supp SWI=2.49, (M)Supp UK=2.27). All differences are significant at the 95% level, except those for France.
Although less pronounced, Fig.5 provides further evidence that supporters and opponents of COVID-19 vaccination tend to view their in-group positively and their out-group negatively. Figure6, which shows the absolute difference in character trait ratings by group and country, corroborates these observations. While both groups show a difference in ascribed character traits toward the in-group and the out-group, the average difference is again slightly but statistically significantly (at the 95% level) greater among vaccination supporters ((M)Supp FRA=2.04, (M)Supp GER=2.31, (M)Supp ITA=2.09, (M)Supp SPA=2.18, (M)Supp SWI=2.03, (M)Supp UK=2.24) than among vaccination opponents ((M)Opp FRA=1.25, (M)Opp GER=1.49, (M)Opp ITA=1.51, (M)Opp SPA=1.23, (M)Opp SWI=1.61, (M)Opp UK=1.22).
Affective polarization among COVID-19 vaccination supporters and opponents.Notes: Figure6 shows the mean level of affective polarization (according to the character trait ratings) by group and country, with 95% confidence intervals. Reading example: In France, vaccination opponents have an average affective polarization of 1.25, while vaccination supporters have an affective polarization of 2.04 on a scale from 1 to 5. The difference is statistically significant.
Overall, our data suggest a divide around peoples opinions on COVID-19 vaccination. Supporters and opponents dislike each other and tend to attribute positive traits to their in-group and negative traits to the respective out-group. Thus, our data indicates the presence of opinion-based affective polarization regarding COVID-19 vaccination in six European democracies in early 2022. Importantly, however, this affective polarization is asymmetric, as the pro-vaccination group tends to be more polarized than the anti-vaccination group. Naturally, this finding raises the question of potential correlates of this form of polarization. As a first step in this direction, exploratory analyses reported in the supplementary material, section E, Figs. 24 show that older age, lower social trust, higher levels of conscientiousness, and a general support for COVID-19 vaccination are associated with higher levels of opinion-based affective polarization regarding COVID-19 vaccination.
See the rest here:
Asymmetric affective polarization regarding COVID-19 vaccination in six European countries - Nature.com