Statistics Question

Metnut · #1 03-05-2007, 04:46 PM

Not really related to poker, but I think this is the best place for me to get help and i cant find any way to solve a problem like this on the internet. It probably would at least apply to being able to find out what percent chance your winning results would be due to pure chance as opposed to other factors (skill).

How would I be able to figure out the percentage that a certain deviation from the mean would be due to chance? For example, say 5000 people are considered for a position, and 600 get that position. Say of that 5000 considered for the position 400 are hispanic and 5 hispanic people got that position. How can I calculate the percentage chance the reason for the lower amount of hispanics getting the position was purely due to chance as opposed to other external factors?

Metnut · #2 03-05-2007, 08:11 PM

Anyone have an idea?

raistlinx · #3 03-05-2007, 11:16 PM

[ QUOTE ]
Not really related to poker, but I think this is the best place for me to get help and i cant find any way to solve a problem like this on the internet. It probably would at least apply to being able to find out what percent chance your winning results would be due to pure chance as opposed to other factors (skill).

How would I be able to figure out the percentage that a certain deviation from the mean would be due to chance? For example, say 5000 people are considered for a position, and 600 get that position. Say of that 5000 considered for the position 400 are hispanic and 5 hispanic people got that position. How can I calculate the percentage chance the reason for the lower amount of hispanics getting the position was purely due to chance as opposed to other external factors?

[/ QUOTE ]

Wow... 600 people for 1 position... do they work for the government?

DWarrior · #4 03-06-2007, 02:22 AM

p^ = 400/5000 = .08
n = 600
x = 5

the Z-score for that is -6.47

Most stats books only go down to a Z-score of -3.89, which they round as having .0001 probability. -6.47 is way smaller than that, so external factors would be my conclusion.

I just used my TI-89 for the calculation, but here's the explanation:
http://home.xnet.com/~fidler/triton/...op/lsprop1.htm