Correlation Help!!

[Monica.scott]

At a recent scotch-tasting party, 5 faculty members and their (opposite sex) spouses were guzzling it back pretty well. One person quite correctly noticed that the men seemed to be drinking more than the women. If the mean number of drinks had by the men was 7, and that for the women was 5, and the standard deviation for the whole group was 2 drinks, what was the correlation between gender and the number of drinks consumed?

 

[bmetzger01]

Monica, have you tried constructing a dataset that would meet the given criteria, and then computing correlation from the raw data? You need 5 data points coded as 0 for female, that average to 5. Then you need 5 more data points coded as 1 for male, that average as 7. Construct as such that the 10 data points together have a standard deviation of 2. Let me get you started..

0 3
0 5
0 5
0 5
0 7
1 ...
1 ...

This may be the long way of doing it, but it does get you the answer. I'm drawing a blank on how you would compute correlation with only the summary statistics.

 

[Monica.scott]

Ok that makes sense. Im not sure if i did this right, but i made up a set of data points for males and females:

Males
3,5,5,5,7
Females
5,7,7,7,9

And made sure the standard deviation was 2 for the whole group (since i suck at this i really hope i did it right and it actually is 2 :/)

Then made a chart and calculated the Correlation Coefficient. I got 0.025. Does this seem right to you or am i doing something wrong?