Correlation does not prove causality
This is one of the first rules in statistics.
A very simple word picture to prove the point: We all know that a rising tide lifts all boats. A naive statistician might look at the data and conclude that lifting one boat, perhaps with a crane, would result in all of the boats in the harbor also being lifted.
In the real world, most metrics are not as strongly correlated as the elevation of the boats in the harbor. A problem solver might note correlation between many factors. He/she will first look for causality relationships between those factors that have the strongest correlation.
Gun ownership rates and suicide rates
Mental Health Maintenance
People are social animals. Perhaps suicide rate correlates more strongly with population density. Lets compare.
There is so much separation that the chart is not readable. |
All I can say is "Wow!" |
Low population density means that potential friends are farther away.
Low population density means that mental health professionals are more distant. In some cases, mental health professionals have an active dislike of American West culture and refuse to provide services to the same.
Psychology Professor at U-T refuses to teach classes if guns might be present. In many other professions this would be considered dereliction of duty.
Low population density is often a disincentive for any health professionals to relocate to your area. They have bills to pay and debt to work off.
Summary
Gun ownership rate and suicide rate are linked but it is not a pure "causal" link. They are linked because they both share a relationship with low population density.
People who live in states with low population density tend to own guns because they have more opportunity to use them. Thus, guns are an investment they can exercise. People who live in densely populated states have fewer opportunities to shoot their guns. Consequently, guns tend to be an investment that provides little return.
The data suggests that people in the states with the lowest population are more likely to seek suicide than states with higher population density. One possible reason is due to the difficulty in talking face-to-face with friends and the difficulty in getting professional medical help.
In the interest of even-handedness, I will note that suicide attempts made with guns are more likely to be successful than attempts made with other methods.
I like to stay positive. One thing we can do is to slow down and actively listen to one person today. It is best to listen to the one person who seems to be hurting the most. Regardless of who you choose, let them know that you think the world is a better place because they are here and let them know that we are pulling for them because we are all in this together
Very timely blog post. Thanks. I know someone who can use this advise.
ReplyDeleteAs an engineer I love it when people notice the might need something besides cartesian coordinates. The rest of the states are probably not going to hurt your correlation between suicide rates and population density. Other interesting data: actual numbers of mental health practicioners per capita by state rather than just using population as a proxy, and also suicide rates by any other means other than firearms.
ReplyDeleteI considered some of those metrics. Even if Wyoming had the same level of practicioners, the vast distances still pose a logistical problem for anybody not living in Cheyenne, Casper, or Laramie. The patient is defeated by the distances and crappy winter weather.
DeleteThe other issue is that much more detail would probably qualify as a Master's thesis in Psychology. I don't know if I could live with myself if somebody gave me one of those. Plus, there is the risk of my leading students into perdition should they find themselves temped to plagiarize my work.
How many big Indian reservations does the urban east coast have? What is the rate of suicide by race?
ReplyDeleteThat is the first question that jumped into my head, along with the weather conditions.
Psyche counselling, especially in 'remote' states/places is lacking to non-existent... As you've pointed out. I would also like to see the breakdown by ethnicity. I think that plays a part also!
ReplyDelete