The specification of cross-sectional models is usually solved following a traditional procedure, highly supported by practitioners. In the first step, a simple model is proposed that will be subsequently improved with different elements if the evidence so advises. This procedure expedites the econometric solution and fits well into the Lagrange Multiplier approach, which contributes to explain its current popularity. However, there are other methods that could also be used, and some of them are considered in this paper. Specifically, we turn our attention to the Vuong test, developed in the context of the Kullback-Leibler information measure. This test represents an intermediate solution between the complexity inherent in the Wald test and the simplicity of the Lagrange Multiplier principle.
