Chi-square difference is inappropriate here because the models are not nested, correct? I know that information theory-based indices like AIC are appropriate for comparing models, regardless of whether the models are nested. However, is it meaningful to use AIC or similar indices when one model ...