Selecting random parameters in discrete choice experiment for environmental valuation: A simulation experiment

This paper examines two tests commonly used to select random parameters in choice modelling: the Lagrange Multiplier (LM) test as proposed by McFadden and Train (2000) and the t-statistic of the deviation of the random parameter. A simulation exercise based on a real case study is carried out assuming cross-sectional data and two panel data settings. These data structures together with different distributional assumptions allow an examination of the empirical size and power of the two analysed tests. The key results indicate that the power of these tests depends on the data structure as well as on the spread and type of the parameter distribution. Furthermore, the LM test is the only one with empirical size not significantly different from the theoretical value.