There are far more cities with Dem mayors than Republican mayors, so comparing raw counts is pretty meaningless. (Similar to counting comparing total homicide instead of per capita.) There are 10 Republican mayors amongst the 50 largest cities, and furthermore, these are heavily concentrated at the low end.
It turns out the relationship between crime and population is non linear, and that nonlinearity is also true outside of the US and it's political context.
So, what you're seeing is bigger cities with higher per capita rates due to the underlying relationship between population density and crime. It also happens to be the case that in the US, urban areas strongly prefer Democratic mayors.
Quoting you: "There are far more cities with Dem mayors than Republican mayors, so comparing raw counts is pretty meaningless. (Similar to counting comparing total homicide instead of per capita.)"
Quoting me: "I did not compare raw counts, I compared per capita counts."
I did not learn anything I did not know before re-reading.
You're using raw counts /of cities/ in a top ten list, when the actual leadership of cities is heavily skewed towards Dem mayors. I point out that this is similar to comparing raw counts instead of per-capita.
Let v = high violence, d = dem mayor, r = republican mayor.
You seem to want to compare P(v|d) to P(v|r). We have little to no data for comparison because there are so few examples of cities with republican mayors. Those that exist are amongst the smaller American cities. This means you need to disentangle the effect of city size on violence rates from the effect of the mayor's party.
I repeat:
There are 10 Republican mayors amongst the 50 largest cities, and furthermore, these are heavily concentrated at the low end.
It turns out the relationship between crime and population is non linear, and that nonlinearity is also true outside of the US and it's political context.
So, what you're seeing is bigger cities with higher per capita rates due to the underlying relationship between population density and crime. It also happens to be the case that in the US, urban areas strongly prefer Democratic mayors.
https://en.m.wikipedia.org/wiki/List_of_mayors_of_the_50_lar...
It turns out the relationship between crime and population is non linear, and that nonlinearity is also true outside of the US and it's political context.
https://crimesciencejournal.biomedcentral.com/articles/10.11...
So, what you're seeing is bigger cities with higher per capita rates due to the underlying relationship between population density and crime. It also happens to be the case that in the US, urban areas strongly prefer Democratic mayors.