[R-sig-ME] Comparing two models with different sample sizes

[Simon Harmel]

Dear Colleagues,

I had to remove three extremely outlying (in terms of residuals, cook's
distances, and hat values) observations from my model.

Now, can I compare the fit of my initial model that has three more
observations with my outlier-corrected model using AICc?

Thanks,
Simon

	[[alternative HTML version deleted]]