Question

For the purposes of testing for a correlation, does it matter which variable is labeled X and which is labeled Y?

A. No it does not matter; the estimate of a correlation (and correlation test) is symmetric.

B. Yes, it does matter; the strength of the correlation (and the result of the correlation test) will depend on which you variable you pick as the explanatory variable.

Answer 1

Correlation : It is the measure of linear relationship between two variables

Correlation can be positive, negative, or no correlation.

Generally it can be denoted by ' r'

It can be calculated as

Where , r = 1 then there is a perfect positive correlation

r = -1 then there is a perfect negative correlation

r = 0   then there is a no correlation

The correlation coefficient lies between -1 to 1

Note that the correlation coefficient is symmetrical with respect to X and Y

i.e.   

Therefor Option A ) No it does not matter; the estimate of a correlation (and correlation test) is symmetric is correct