On Thu, 30 Nov 2000 17:51:26 GMT, [EMAIL PROTECTED] wrote:
> This can't be that complicated, but I can't figure it out myself.
>
> I need to determine whether the distribution of the total state
> population into 21 counties is the same as the distribution of voters.
> I have two lists of percentages, and they look similar by eye. How do
> I determine whether the differences are significant?
>
They are different.
Now what do you mean by "significant"?
I suppose that one meaning might be,
"Does a contingency chi-squared test show enough difference that there
is some decent statistical power to detect and measure a mediating
variable (such as age)?"
In that case, the simple formula uses the raw Ns in each cell, in the
2x21 table, Voters vs. non-voters for each county.
--
Rich Ulrich, [EMAIL PROTECTED]
http://www.pitt.edu/~wpilib/index.html
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
