Worth keeping in mind that polls measuring support for violence come with several caveats. For example a 2021 study assessed how question wording, respondent disengagement and illegitimate answers affected the results, finding that accounting for these factors lowered estimates of support for violence from 18.5% to 2.9%, or at most 6.9%. Even then there are arguments to be made about how social desirability bias and the specific scenarios presented can actually lead to underestimations.

https://youtu.be/B2MB2re24oA?t=328