Archived PEW: ANNUAL FIREARM-RELATED HOMICIDES DOWN NEARLY 50 PERCENT FROM 20 YEARS AGO (breitbart.com)
submitted ago by beren
Posted by: beren
Posting time: 5.1 years ago on
Last edit time: never edited.
Archived on: 2/12/2017 1:51:00 AM
Views: 1485
SCP: 92
101 upvotes, 9 downvotes (92% upvoted it)
Archived PEW: ANNUAL FIREARM-RELATED HOMICIDES DOWN NEARLY 50 PERCENT FROM 20 YEARS AGO (breitbart.com)
submitted ago by beren
view the rest of the comments →
[–] rwbj ago
Assuming you're genuinely interested in how you should approach this I can give you some ideas.
Let's assume you think A might have a causal relationship with B because of some sample. The first thing you want to do is to try to ensure that the relationship is not being caused by some other happening. The best way to do this is to resample. So for instance if the theory is that an increase in gun ownership rates results in a reduced homicide rate then you should be able to see a consistent trend over countries other than just the one you start with also having a similar trend. In this case you're kind of buggered from the start. Gun ownership rates in the US are lower than they've been in decades. Here is the article linked to from Breitbart which states:
What you have in the US is a declining rate of gun ownership, but the households that do own guns own more than ever. So in a country of 100 people instead of 50 owning guns and having 1 each, we have a scenario in the US where we have only 30 people owning guns but they have 3 each. The number of guns has nearly doubled, but the number of owners has actually substantially decreased.
Beyond that the next step you should take is to consider the corollary. If gun ownership rates cause a decrease in gun homicide rates then a decrease in gun ownership rates would cause an increase in gun homicide rates. Here again you have plenty of test cases available since you can simply compare before and after gun homicide rates in countries that did enact gun control. And in none of them has there been any indication of an increase in gun homicide rates following tighter gun controls.
Lastly you might also want to look at the extremes of your samples to see how they match up with your hypothesis. So for instance in countries where gun ownership rates are incredibly high you'd expect to see incredibly low gun homicide rates. In countries where gun ownership rates are incredibly low you'd expect to see incredibly high rates of gun homicides. Again this doesn't match up very well with reality. Most countries with incredibly low gun ownership rates also tend to have incredibly low gun homicide rates. The other side of the scale (countries with very high gun ownership rates) is much more difficult to examine since so far as I know nearly all data on gun ownership I know of is done per capita similar to the United States which means 1 person owning 100 guns in a country of 100 people would show one gun per person, on average, which is misleading and irrelevant.