Calculating correlation with unknown sample size

nche

New Member
#1
Hello, friends. Here is the question that I would like answered. I have literally been sitting here for over an hour trying to find the solution on the web, and I am beyond frustrated (my broken arm doesn't help).

What would be the correlation between the annual salary of males and females at a company if for a certain type of position men always made:

a) $5,000 more than women?
r=

b) 25% more than women?
r=

c) 15% less than women?
r=

I am used to calculating the correlation coefficient with actual values (x and y), but since this one had none except for the different changes, I did not know what to do. How can we determine a, b, and c without actual given values?
 

hlsmith

Less is more. Stay pure. Stay poor.
#2
How did you break your arm?

Where does this question come from and for clearification, there is not other information presented?
 
#3
if for a certain type of position men always made:

a) $5,000 more than women?
(I note the phrase: "men always made")

Sometimes, a little example helps. Here is one for you:

sal_women sal_men
35 40
41 46
42 47
43 48
55 60


If you are one of those who believe that you can't conclude anything from such a small sample, I suggest you to generate a sample of 5000 or 5 millions of observations.

I generated the data like

sal_men = 5 + sal_women

Generate the other two by:

sal_men = 0.25*sal_women

sal_men = (1-0.15)*sal_women

As a bonus question:
suppose that men made 1% more that women:

sal_men = 1.01*sal_women

would the size of the correlation coefficient, tell you anything about the size of the salary difference?

So, how important is the correlation coefficient?

- - -

Hello, friends.
Sorry I can't remember if we had exchanged "friends greetings". Or did you direct this to someone else? But I suggest that you send an friends request to Mr BryanGoodrich. I think he will be enchanted and answer all your questions.