Negative R-squared fixed panel regression

#1
Hi,
I did a fixed effects regression for my panel data. My R square value is 0.05 and my adjusted R square value is negative (-0.10). How do I interpret this?
 

hlsmith

Not a robit
#2
Can you describe your study context and model in more detail. I don't believe a negative R^2 value is possible, well it is not within the 0-1 confines of the metric. Let us hear more about your data and model!
 

Dason

Ambassador to the humans
#3
Their r-square isn't negative. The adjusted r-square is. And for a good model it won't be negative but the adjusted r-square definitely can be negative.
 

hlsmith

Not a robit
#4
Yes, I follow that and was going to mention it as well but was first interested to see how many terms they had in their model along with sample size and what the panel consisted of, etc. :)

This same thing happens in meta-analyses and I^2 statistics, where they can land below zero.