Suppose the least-squares regression line for $y$ and $x$ is $y = kx$. Given that $0 < k < 1$, can we say anything about the means of $y$ and $x$? Can we infer that $\bar{y} < \bar{x}$?
Asked
Active
Viewed 25 times
1 Answers
0
As I recall, the least squares coefficient for $y = kx$ is $k = \frac{\sum xy}{\sum x^2} $.
This doesn't say anything about the means of $x$ and $y$, but does imply imformation about the means of $xy$ and $x^2$.
marty cohen
- 107,799