Suppose that two models are non-nested and there are \(N=200\) observations. For model 1 , the number of parameters \(q=10\) and \(\ln L=-400\). For model 3 , \(q=10\) and \(\ln L=-380\).

(a) Which model is favored using AIC?

(b) Which model is favored using \(\mathrm{BIC}\) ?

(c) Which model would be favored if the models were actually nested and we used a likelihood ratio test at level 0.05 ?