Quote:
Originally Posted by aowen
[You must be logged in to view images. Log in or Register.]
This aside, my assessment, which must be the third time I have said this, is that crime rates are affected by much more than just gun laws, such as education levels, poverty levels, and other social standards. However, I said that gun laws are an element of the cause. Also, you seem to have some grand problem with correlation. Correlation pretty much asserts a probability. If there was a correlation that gun laws made no impact whatsoever on homicides by firearm and burglaries with firearms, I would take that seriously, but evidence suggests the opposite. And I have connected the dots between these statistics to a degree, but there's not much of a point in continuing the argument until people are willing to read them.
|
No, correlation does not assert a probability of causation. At first I thought you were just being deceitful in trying to use statistics to confuse and overwhelm people. It's a pretty common way to manipulate the ignorant masses. As you continue to write, it becomes clear you are actually part of the ignorant masses. My first clue was when you started dropping pseudoscience terms like retrocausality in an attempt to sound smart. Now you don't seem to understand the problems with inferring causation from correlation at all. You see, if A and B happen, A might cause B, B might cause A, C might cause A and B, or there may be no link whatsoever. A stronger correlation does not help us in determining which of these is the case. And clearly in any complex system, a single effect may be influenced by multiple causes. None of this permits you to infer causation from correlation.
To put it in terms of this debate, let's assume for a minute that there is a strong correlation between increased gun laws and reduced murder rate. It could be the case that adding a new gun law will directly reduce the murder rate. Or it could be the case that as murder rates drop due to other reasons, people feel less that they need guns and thus enact more laws. Or it could be the case that as a nation of people start to feel more close-knit as a society, they both kill each other less and accept tighter gun restrictions. The strength of the correlation tells us nothing about which are causes and which are effects.
Determining causation is far more difficult, particularly in the social sciences, where scientific experiment is for all practical purposes impossible. I don't think I have the answer on the best way to do this, though it's clear that statistical correlation is not even close to the right way. I'm sure you'll just keep on using it though. It's pretty easy to fool people into believing you when you can cite scientific-sounding sources. Most people simply aren't tuned in enough to the nature of knowledge to have any chance at defending against such an assault on logic.
09-21-2013, 02:10 PM
Linear Mode
