Hacker News new | ask | show | jobs
by koonsolo 2465 days ago
From a statistical standpoint, it's logical that the most extreme numbers come from the lowest populations. So if the lowest rates are also in remote rural areas, this says more about statistics than anything else.
1 comments

I'm not sure which of these two points you're trying to make:

1) These specific counties could be flukes because of their small population.

We can account for that by making multiple observations--i.e. look at multiple years of data--and seeing if the counties change rank. The visualization has yearly data for 1980-2014, and the specified counties are consistently high compared to the national average

https://vizhub.healthdata.org/subnational/usa

2) The 'rural' component of these counties isn't necessarily a factor, since the counties with lowest rates may be rural as well.

A bit of searching turns up a clear correlation between population density and suicide:

https://www.researchgate.net/figure/How-population-density-a...

I was trying to make point 1, but you definitely showed this is not the case.

That last graph is a really nice one, as it shows more variance on low populations, but on average still way higher than on high populations.

Thanks for the clarification! :)