TESTING FOR INDEPENDENCE OF IRRELEVANT ALTERNATIVES - SOME EMPIRICAL RESULTS

In a recent article, Zhang and Hoffman discuss the use of discrete cho ice legit models in sociological research. In the present article, the authors estimate a multinomial logit model of U.K. Magistrates Courts sentencing using a data set collected by the National Association for the Care and Resettlement of Offenders (NACRO) and test the independe nce of irrelevant alternatives (IIA) property using six tests. Conduct ing the tests with the appropriate large sample critical values, the a uthors find that the acceptance or rejection of IIA depends both on wh ich test and which variant of a given test is used. The authors then u se simulation techniques to assess the size and power performance of t he tests. The empirical example is revisited with the inferences perfo rmed using empirical critical values obtained by simulation, and the r esultant inferences are compared. The results show that empirical work ers should exercise caution when testing for IIA.