04-27-2017, 10:34 AM
(04-27-2017, 09:05 AM)fredtoast Wrote: Any statitician will also tell you that looking at "variables" can give you "more exact" information.
Looking at the rate of heart disease in smokers is looking at a smaller sample size than the entire population, but that does not mean the stats about smokers are meaningless and skewed by smaller sample size.
Once again, I'll ignore your cherry picking. This is an inane analogy as you try and compare the impact of a certain behavior on one segment of the entire population versus radically different sample sizes studying the exact same thing. Comparison fail.
Quote:So when disussing the effects of gun laws and gun ownership it is helpful to compare different states with different gun laws and owneership rates. the fact that these states have smaller sample sizes is meaningless because you have to look at the variables to make any educated conclusions.
Sure, just as long as you realize, as any statistician would, that the smaller sample sizes of less populace states are going to give you more skewed data than the more populous states.
Quote:there can be a lot of arguments about other variables that might effect the crime rate in different states but it is ridiculous to claim that statistical analysis does not include breaking down statistics into smaller parts based on variables.
As I already said. However, then using this data to make absolute direct comparisons is flawed in its very concept. I'll give you an actually relevant analogy. Take the top twenty cities by murder rate in the United States.
http://www.worldatlas.com/articles/most-dangerous-cities-in-the-united-states.html
You'll find the overwhelming majority of them are Democratic run cities. Ergo, we must conclude that Democratic governance leads to higher murder rates. There couldn't possibly be other factors at work could there?
