quote:

Originally posted by Jon_Snow Can someone give me a cor version? I can't be bothered to read through this thread.

Side A: "hey dude i have a gun and i never killed someone"
Side B: " but broooo anyone can get a gun in US even psychos"
Side A: "thats why we need more guns to protect ourselves"7
Side B: "Youre the laughing stalk of the wooooorld bro"
Side A: "Guns don't kill people bro its the people that kill people"
Side B: "Dude didnt you see that shit in china all the psycho had was a knife and everyone lived because of it"
Side A: "Cause bro didnt know how to use a fuckin knife"
Side B: "Check out these stats"
Side A: " nahhh bro check out THESE stats"

quote:

Originally posted by pkcRAISTLIN uh huh.

that plot is pretty meaningless without a correlation co-efficient (which i can gauge right now is very very low).

there is obviously some correlation, but nothing to really write home about.. you remove 3 or 4 states from that list and you're almost getting an inverse correlation there.

the rule of thumb says 95% correlation or better is meaningful data.

But none the less, the correlation is there. and the point is logical.

nahhhhhhh bro check out thoooose stats

quote:

Originally posted by Halcyon+On+On

I had to do a double take, i didn't catch on to your joke LOL

theres the real one.

http://stlouis.cbslocal.com/2012/12...-toward-patron/

___________________
There are more things in heaven and earth, Horatio,
Than are dreamt of in your philosophy.

quote:

Originally posted by Vivid Boy nahhhhhhh bro check out thoooose stats

Housing prices are on teh rise!

all day everyday. I make my own graphs to prove my bullshit

quote:

Originally posted by Nrg2Nfinit theres the real one.

Quite common to eliminate 2 outlying points (lowest, highest) when plotting lines of best fit though, yeah? sorry its been like 15 years since I did statistical analysis.

___________________

Now with extra singles!
my old stuff, not quite up to snuff - but I still dig it - UPDATED 9/23/2012

quote:

Originally posted by pkcRAISTLIN Quite common to eliminate 2 outlying points (lowest, highest) when plotting lines of best fit though, yeah? sorry its been like 15 years since I did statistical analysis.

If you are referring to the graph that i posted, 0.75 is a very weak correlation. Usually you want anything above 0.95 for significance (and that is very liberal still). regardless, the relationship stays if you remove the outliers. as guns go up so does deaths per 100k. I wouldn't use this as a predictor for deaths if you set "x" as the number of guns.


With regards to removing outliers, you would want to remove anything that falls away from the line of best fit, of course that will alter the line of best fit.

Again the sample size is weak here, variance is high.. correlation overall weak... you basically have 75% confidence in that data set on how it relates to that regression line listed.

Well I don’t think anybody is claiming that access to guns is the only variable.