I'm just starting my work on this so please go easy on me. I have a bunch of
data in front of me that will help me generate some answers in five
different categories. I have the ability to sort the data geographically,
gender, age etc. and was wondering if anyone could recommend or help me
determine/test if there is a correlation between certain groups? e.g.
females in different geographic areas? What would consist of a normal
correlation test?

HotRod -

If you want to describe relationships among categorical data, you can use a
Pivot Table to get summary counts for two characteristics that will help you
to visually decide whether a relationship exists. For a "test," a chi-square
test of independence might be appropriate.

If the data are numerical, then you could start with an XY (Scatter) chart
to visually describe the relationship between two characteristics, and you
could use the correlation coefficient (worksheet function CORREL) as a
summary measure. For a "test," the Regression tool provides an F-test for
overall fit and t-stats for individual coefficients.

- Mike
http://www.mikemiddleton.com

"HotRod" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed)...
> I'm just starting my work on this so please go easy on me. I have a bunch
> of data in front of me that will help me generate some answers in five
> different categories. I have the ability to sort the data geographically,
> gender, age etc. and was wondering if anyone could recommend or help me
> determine/test if there is a correlation between certain groups? e.g.
> females in different geographic areas? What would consist of a normal
> correlation test?

Mike
Thanks I'll give these a shot and see what gets me the best results.
I've got some reading to do...



"Mike Middleton" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed)...
> HotRod -
>
> If you want to describe relationships among categorical data, you can use
> a Pivot Table to get summary counts for two characteristics that will help
> you to visually decide whether a relationship exists. For a "test," a
> chi-square test of independence might be appropriate.
>
> If the data are numerical, then you could start with an XY (Scatter) chart
> to visually describe the relationship between two characteristics, and you
> could use the correlation coefficient (worksheet function CORREL) as a
> summary measure. For a "test," the Regression tool provides an F-test for
> overall fit and t-stats for individual coefficients.
>
> - Mike
> http://www.mikemiddleton.com
>
> "HotRod" <(E-Mail Removed)> wrote in message
> news:(E-Mail Removed)...
>> I'm just starting my work on this so please go easy on me. I have a bunch
>> of data in front of me that will help me generate some answers in five
>> different categories. I have the ability to sort the data geographically,
>> gender, age etc. and was wondering if anyone could recommend or help me
>> determine/test if there is a correlation between certain groups? e.g.
>> females in different geographic areas? What would consist of a normal
>> correlation test?
>
>