A comparison between mixed logit model and latent class logit model for multi-profile best-worst scaling: evidence from mobile payment choice dataset

This paper uses a multi-profile best-worst scaling dataset to compare the mixed logit model and the latent class logit model for mobile payment choice. Three non-nested tests are applied to show the comparison results. The results indicate that the mixed logit model is superior to the latent class logit model in all three tests.