Unconditional heteroskedasticity

You have a bunch of error terms: one for each observation. Some are big, some small, some positive, some negative.
The error variance is the variance of these error terms.
 
Back
Top