(This article was originally published at Statistical Modeling, Causal Inference, and Social Science, and syndicated at StatsBlogs.)
Solomon Hsiang writes:
I [Hsiang] have posted about high temperature inducing individuals to exhibit more violent behavior when driving, playing baseball and prowling bars. These cases are neat anecdotes that let us see the “pure aggression” response in lab-like conditions. But they don’t affect most of us too much. But violent crime in the real world affects everyone. Earlier, I posted a paper by Jacob et al. that looked at assault in the USA for about a decade – they found that higher temperatures lead to more assault and that the rise in violent crimes rose more quickly than the analogous rise in non-violent property-crime, an indicator that there is a “pure aggression” component to the rise in violent crime.
A new working paper “Crime, Weather, and Climate Change” by recent Harvard grad Matthew Ranson puts together an impressive data set of all types of crime in USA counties for 50 years. The results tell the aggression story using street-level data very clearly [click to view the graphs]:
All crime increases as temperatures rise from 0 F to about 50 F. It seems reasonable to hypothesize that a lot of this pattern comes from “logistical constraints”, eg. it’s hard to steal a car when it’s covered in snow. But above 60 F, only the violent crimes continue to go up: murder, rape, and assault. The comparison between murder and manslaughter is elegantly telling, as manslaughter should be less motivated by malicious intent.
This seems important to me. Just one graphics tip (to Ronson, not to Hsiang, who is merely reporting this work): The graphs are great, but they’d be even better if they were rearranged slightly. First, put the labels on the top rather than bottom of the graphs. As it is, when I first looked, I thought that the results for Murder were in the second row, not the first row. You can make more space for the titles by removing the boxes around the graphs (in R notation, bty=”l” rather than bty=”o”). The x-axes are too busy, it would be enough to label temperature every 20 degrees rather than 10. (I’d actually prefer Celsius but that’s more of a judgment call.) The zero line should be gray rather than black so as not to so strongly distract from the results. The y-labels would be improved by (redundantly) naming the crime: thus, Murder Rate, Manslaughter Rate, Rape Rate, etc., rather than the identical “Number of Crimes” for each.
Finally, I’d prefer the scales of the y-axis to be directly interpretable. Instead of presenting changes in monthly crime rate per 100,000 persons, I’d present changes in crime rates per crime. For example, if there were 15,000 murders in the U.S. in one year, that’s 15,000/(12*100,000)=0.0125 murders per 100,000 people per month. A change in 0.002 in the upper-left graph then corresponds to 0.16 or 16% of the murder rate. Maybe I got the numbers wrong here; in any case, my point is that percentages of the murder rate will be much more relevant than crime rate per person.
Please comment on the article here: Statistical Modeling, Causal Inference, and Social Science