Correlation Coefficient

It is rescaled to be between $-1$ to $1$ , and it is also unit free.

$$\rho =\frac{cov(X,Y)}{{\sigma }_X{\sigma }_Y}$$

How to compute $cov(X,Y)$? See Covariance

If $\rho =1$, it is perfect positive linear relation

If $\rho =-1$, it is perfect negative linear relation

(Most of the time the Correlation Coefficient would be between when estimate econometrics problem.)

That is, we use absolute value $|\rho |$ to reflects the strength of linear association between the 2 variables.