How would you find the correlation coefficient from given Sum of Squares values (Regression and Error) and regression values - slope of the estimate and standard error?
Any help would be appreciated.
$\begingroup$
$\endgroup$
2
-
$\begingroup$ Results from simple linear regression or multiple linear regression? $\endgroup$– user158565Dec 9, 2018 at 23:11
-
$\begingroup$ Single linear regression $\endgroup$– Hannah L.Dec 9, 2018 at 23:18
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
For simple linear regression, the regression coefficient and correlation coefficient have following relation (when sum of square (SS) due to regression and error are know).
correlation coefficient = sign(slope) * sqrt[SS(regression)/SS(Total)]
SS(Total) = SS(regression)+SS(Error).
sign(slope) = 1 if slope >0. = -1 if slope < 0, and =0 if slope = 0.
answered Dec 9, 2018 at 23:25