EViews.com

Deleted

by putting the sample as when winter=1 and msg=1, then you're running the regression with both winter and msg and the constant equal to 1. Three identical variables. Hence the singularity.

You'll either have to not have that as your sample, or remove both winter and msg from the regression.

Why not just ask Prof. Ellis????

(You're trying to estimate the effect of winter and msg, but you've restricted the sample so these only have one value.)