Extended Data Fig. 3: Results of Experiment 1 (A,B) and Experiment 2 (C, D) stratified by gender.

This figure presents the disaggregated results of the two conjoint experiments stratified by gender. Gender was assessed by respondents’ self-reports in the survey. Therefore, we used the term "gender" throughout this work because of the self-reported nature of the data. We included only male and female gender as categories in this analysis since the number of participants who indicated ‘non-binary/other’ gender was too small to obtain reliable results for this group (depicted in Supplemental File 1). Panels A and B show the results of Experiment 1 by gender. Panel A refers to the binary measure of campaign evaluations, whereas Panel B depicts the ratings of the likelihood of getting vaccinated. Overall, there are only small differences between gender groups, and confidence intervals mostly overlap. One notable difference is that men evaluated vaccination campaigns with cash incentives as more appealing than women. Also, regarding the likelihood of getting vaccinated, men seemed to respond somewhat more strongly to positive incentives, although the confidence intervals between the gender groups overlap in this case. Panels C and D show the results of Experiment 2 by gender. Again, we find only small differences across gender groups. In fact, confidence intervals overlap for all of the treatment conditions. ‘GenderP’ means the gender of the survey participants as opposed to the gender of the testimonial in the case vignettes. We calculated Average Marginal Component Effects (AMCEs). Data are presented as AMCE estimates +/−95% confidence intervals. Exact p-values are shown in Supplemental Files 13–16. In total, 6,357 respondents participated in both experiments (Austria: n = 3,187; Italy: n = 3,170). For further information on the number of male and female respondents included in the sample, see Supplemental File 1. We also report the gender composition of the vaccination status groups in Supplemental File 2.