Consider a multiple linear regression experiment. I am estimating 2 of the n parameters. I did this using:
1) A joint confidence region for the two parameters an ellipse.
2) Two separate confidence intervals for the two parameters.
and got this plot:
I would like to know when do we have a joint confidence region an (ellipse) actually larger than the two separate confidence intervals and when is it the other way around so not enclosed in the rectangle of the two confidence intervals?
