election statistics question

looking at the new jersey 2005 election... i want to see if counties with a higher percentage of "over 65" people correlate to either a higher dem or higher rep voting percentage. i have all the raw numbers, i am using excel. i might have just answered my own question, do i use correlation function? but i'd like an explanation of the math behind whatever function it is that i should use. i took intro to stats a few years ago.


TS Contributor

In order to use the CORREL function (correlation function in Excel), make sure that you have two columns of data, with each row representing a county:

column 1 --> for each county, the percentage of voters over 65 years of age
column 2 --> for each county, the percentage of democrat votes

Assuming %dem + %repub = 1, it would be redundant to have a third column with % republican votes.

The CORREL function will tell you an index of the strength of the relationship between the two variables, from -1 (strong negative relationship) to +1 (strong positive relationship). Close to 0 indicates little if any relationship.

If you square the result from CORREL, you get the coefficient of determination, or the percentage of variance in one variable "explained" by the variance in the other.

There are a whole slew of sites that go into regression / correlation in detail. A good one is HyperStat (ch 15):