> # Print results > print(lrtest) Likelihood ratio tests of cumulative link models: formula: link: threshold: m2 categorical_FOD_FODs ~ 0 + Condition_FODs:Gender_quantised_FODs - 1 logit flexible m1 categorical_FOD_FODs ~ 0 + Condition_FODs:Gender_FODs - 1 logit flexible no.par AIC logLik LR.stat df Pr(>Chisq) m2 6 5839.2 -2913.6 m1 13 5847.8 -2910.9 5.4581 7 0.6042 > print(AICs) df AIC m1 13 5847.790 m2 6 5839.248 > print(BICs) df BIC m1 13 5927.194 m2 6 5875.896 which test is ideal to conduct?
The stepwise AIC algorithm selects the model with the lowest AIC value. According to this test we do not reject the null hypothesis, and therefore, we do not select the model m2.