two times the difference between the log-likelihood of the model which includes all the independent variables and the log-likelihood of the model obtained by deleting one of a specified variable. This test follows the asymptotic 〖χ^2〗_(α,p) distribution for a level of significance of 0.05 and with