how would you compare the models from one another?
Answers
Answer:
I have a set of results from mathimatical model along 1000 m for every100 m and for different dates. What is the best way to statistically compare the performance of this model with the standard model?
Answer:
There are some questions about this site on the related question of how to compare models when the datasets are completely different. For example this question, or this other question, for which only answer at present suggests that AIC might possibly be appropriate. However, I understand from p.80 of Anderson and Burnham (2004) that “AIC Cannot Be Used to Compare Models of Different Data Sets”.
But what if the datasets are not the same, but one is a subset of the other?
My question is intended to be general but I have a specific scenario in mind. The scenario I have in mind is one in which there are two two-level multilevel models.
Model A is applied to a dataset of 150,000 cases. Model B is applied to a dataset of 2,000 cases, of which all cases are drawn from the 150,000.
The models are identical, except that Model B has one additional binary Level 1 predictor.
If the answer to the question depends on whether the model is a multilevel one or not, I’ll edit the question title to include the word “multilevel”.
Anderson, D., & Burnham, K. (2004). Model selection and multi-model inference. Second. NY: Springer-Verlag.