Apr 302012
 

In performing a linear regression, we must often ask ourselves, when should we remove outliers, when should we scale variables, centre variables, or transform them? This article looks at what transformations are available and when to use them;

Scaling IV or DV

This will adjust the gradient and the intercept. In a multiple linear regression scalling the various IVs, we most likely do this to ensure they have roughly the same mean/SD so that the unstandardised gradients are easier to interpret.

Continue reading “Linear Regression: When to adjust variables” »