By using this site, you agree to our Privacy Policy and our Terms of Use. Close
DaHuuuuuudge on 08 January 2013
Edit Thread
Kasz216 said:
DaHuuuuuudge said:
Kasz216 said:

mean negative correlation.  Trying to put it into terms that he can understand.  Since he geniunely seems to be interested.   Like when people say "Reverse Racism" even though there is no such thing as "Reverse Racism" 

There generally tends to be a negative correlation.  Which is likely... completely unrelated.

As for there being no statistical significance... Look at the last link he posted.

What's the point of talking about Alpha when nobody knows what Alpha is?  (Note for others... see Correlation Coefficent) 

In general, you can often eyeball a linear regression trend line and know when it's not significant.  Unless someone built their graph in a stupid way stretching out either the X or Y.

-_-

The point of talking about alpha-levels is because essencially, they are what statistical significance is. You seem to be confusing correlation coefficient with alpha-levels, an alpha level tells you how likely it will be to reject the null hypothesis, and the correlation coefficient (r-value) shows the level of correlation as well as the nature of correlation (positive vs. negative)

You judge the correlation coefficent with the alpha....

I hate to go all wikipedia on you... which is weird that I'd have too since your a government stats guy but....

http://en.wikipedia.org/wiki/Statistical_significance

 

 

The significance level is usually denoted by the Greek symbol α (lowercase alpha). Popular levels of significance are 10% (0.1), 5% (0.05), 1% (0.01), 0.5% (0.005), and 0.1% (0.001). If a test of significance gives a p-value lower than the significance level α, the null hypothesis is rejected. Such results are informally referred to as 'statistically significant'. For example, if someone argues that "there's only one chance in a thousand this could have happened by coincidence", a 0.001 level of statistical significance is being implied. The lower the significance level chosen, the stronger the evidence required. The choice of significance level is somewhat arbitrary, but for many applications, a level of 5% is chosen by convention.[3][4]

 

 

If a correlation isn't strong enough, it can't pass the Alpha... because it's not strong enough to not be random chance.

I'm honestly not sure if you're just trolling right now. The correlation coefficient isn't mentioned once in the quoted entry. R values are for graphs, alpha values are for inference tests. Alpha levels don't have to be for data that is graphed. I know i shouldn't keep responding, it appears the more wrong you are the more you will reply, but misuse of statistics and condescension are two vehicles for argument i greatly dislike, and you employed them both