Charting the correlation between crime and religious beliefs.
Calling btny! bthny bad graph alert in aisle six! This is egregious, Economist. Did you take statistics in clown college?
going to pass this one off to abbyjean because it’s my bedtime and looking at this is making my head hurt
this makes my head hurt too.
a few aesthetic things - why are they circles instead of dots? what good does the empty white space in the middle do except make it harder to read? also, what work are the colors doing? they’re not really telling any kind of interesting story, and they’re so similar and pale that they’re hard to easily differentiate. result = clutter without information.
now let’s talk about the two measurements. characterizing the metric of “the difference in percentages of people who believe in heaven and in hell” as “religious beliefs” is, um, a pretty interesting definition of “religious beliefs.” just looking at the title, i’d expect it to measure percentages who believe in god, who attend church regularly, or other measure of religiousity. measuring how many more people believe in heaven than in hell may be some kind of metric, but it’s not what i’d define as “religious beliefs.”
also, reporting religious beliefs as a difference in percentages makes it pretty confusing. take colombia, where 40% more people believe in heaven than in hell. that could mean that 100% of people believe in heaven and 60% believe in hell, or that 40% believe in heaven and 0% believe in hell. those are two very different portraits of a country’s religion! similarly, in the countries where the same proportion believe in heaven and hell - does that mean everyone does? nobody does? half do? that seems relevant to evaluating a country’s religious beliefs.
the crime measurement also does them a disfavor for the association they’re trying to demonstrate, because measuring it in standard deviations from the mean obscures just how different a country that’s two or three standard deviations above the mean is from one that’s half a standard deviation above the mean. since i think their association is hooey because their measure of religious beliefs is baffling and nonsensical, i don’t care about this very much.
finally, it’s obviously trying to imply a causal connection - people who believe in hell are scared to reduce crimes and thus lower national crime rates! - which couldn’t be concluded even if they did demonstrate an association.
i do like that they used red lines to show the zero values on the X and Y axis.