We mentioned last class that the comparison of ICC and AIC would favor AIC since they were comparing it using an optimization of the likelihood, but AIC is a penalized likelihood function so it should always win. What's a way that we can compare ICC and AIC in an unbiased way?
We mentioned last class that the comparison of ICC and AIC would favor AIC since they were comparing it using an optimization of the likelihood, but AIC is a penalized likelihood function so it should always win. What's a way that we can compare ICC and AIC in an unbiased way?