标题：Choosing Between Two Income Distribution Models with Contaminated Data, (revised version published in the Journal of the Royal Statistical Society, Serie B, 15 (1997), pp. 715-727)
出版社：Suntory Toyota International Centres for Economics and Related Disciplines
摘要：Choosing between two income distribution models typically in-volves testing two non-nested hypotheses, that is hypotheses such thatone cannot be obtained as a special or limiting case of the other. Cox(1961, 1962) proposed a classical testing procedure based on the com-parison of the maximized likelihood functions for the two models. Inthis paper it is shown that such a procedure is not robust in that asingle observation can reverse the decision. A robust version of Cox-type test statistics is proposed which can be used for the comparisonof any parametric model. Its robustness properties as well as otherproperties are shown in simulated examples
关键词：M-estimators; model choice; robust tests; Income distri-;bution; linear regression
