Can i pick the Model with lower AIC but higher BIC?

[Sohe Aja]

Good day Everyone,

I was comparing two models
Model2 has a lower AIC but a higher BIC.
can i pick model 2 as the better model since its lrtest ( ll(model)) is also lower than model 1.

Model	N	II(null)	II(model)	df	AIC	BIC
Model 1	47,111		-5159.380	8	10334.760	10404.650
Model 2	47,111		-5153.186	12	10330.370	10435.210

thank you for responding.

[Andrew Musau]

The decision rule is easy: select the model with the lowest AIC/ BIC. Therefore, in as far as it is reasonable to select a model based on these criteria, then you can do so and state that you selected the model based on AIC. However, you should be careful not to omit variables that the theory deems important, regardless of what these model selection criteria tell you.

[Sohe Aja]

The decision rule is easy: select the model with the lowest AIC/ BIC. Therefore, in as far as it is reasonable to select a model based on these criteria, then you can do so and state that you selected the model based on AIC. However, you should be careful not to omit variables that the theory deems important, regardless of what these model selection criteria tell you.

Thank you Andrew