A Monte Carlo Study of Tests for the Independence of Irrelevant Alternatives Property

A plethora of tests for the Independence of Irrelevant Alternatives (IIA) property of Logit models of discrete choice behavior has been proposed in the literature. These tests are based upon asymptotic arguments and little is known about their size and power properties in finite samples. This paper uses a Monte Carlo simulation study to investigate the size and power properties of six tests for IIA in the multinomial Logit model. Our results show that the majority of tests based upon partitioning the choice set appear to have very poor size and power properties in small samples. Tests for IIA based upon the DOGIT model, similarly have poor size properties, but in some circumstances do have reasonable power properties.

[1]  Marc Gaudry,et al.  Testing the Dogit Model with Aggregate Time-Series and Cross-Sectional Travel Data , 1979 .

[2]  Daniel McFadden,et al.  Modelling the Choice of Residential Location , 1977 .

[3]  Maxwell L. King,et al.  Locally optimal one-sided tests for multiparameter hypotheses , 1997 .

[4]  David A. Hensher,et al.  Applied discrete-choice modelling , 1980 .

[5]  R. Brooks,et al.  Economic Motivations for Limited Dependent and Qualitative Variable Models , 1993 .

[6]  M. Gaudry.  Asymmetric Shape and Variable Tail Thickness in Multinominal Probabilistic Responses to Significant Transport Service Level Changes , 1989 .

[7]  M. Ben-Akiva,et al.  EMPIRICAL TEST OF A CONSTRAINED CHOICE DISCRETE MODEL : MODE CHOICE IN SAO PAULO, BRAZIL , 1987 .

[8]  Laszlo Matyas,et al.  A Comparative Analysis of Different Monte Carlo Methods , 1993 .

[9]  Moshe Ben-Akiva,et al.  Incorporating random constraints in discrete models of choice set generation , 1987 .

[10]  Wagner A. Kamakura,et al.  Book Review: Structural Analysis of Discrete Data with Econometric Applications , 1982 .

[11]  Robert F. Bordley.  The dogit model is applicable even without perfectly captive buyers , 1990 .

[12]  J. MacKinnon,et al.  Estimation and inference in econometrics , 1994 .

[13]  C. Hsiao,et al.  Multinomial Logit Specification Tests , 1985 .

[14]  Marcel G. Dagenais,et al.  The dogit model , 1979 .

[15]  R. Laferriere,et al.  SHARE: THE S-1 TO S-5 PROGRAMS FOR THE STANDARD AND GENERALIZED BOX- COX LOGIT AND DOGIT AND FOR THE LINEAR AND BOX-TUKEY INVERSE POWER TRANSFORMATION-LOGIT MODELS WITH AGGREGATE DATA , 1993 .

[16]  D. McFadden,et al.  Specification tests for the multinomial logit model , 1984 .

[17]  An Analysis of the Effect of an Offender's Employment Status on the Type of Sentence Chosen by the Magistrate , 1990 .

[18]  D. McFadden.  Econometric analysis of qualitative response models , 1984 .

[19]  Moshe Ben-Akiva,et al.  CONSTRAINTS ON INDIVIDUAL TRAVEL BEHAVIOR IN A BRAZILIAN CITY , 1986 .

[20]  Y. Tse.  A Diagnostic Test for the Multinomial Logit Model , 1987 .

[21]  Joel L. Horowitz,et al.  Identification and diagnosis of specification errors in the multinomial logit model , 1981 .

[22]  D. McFadden.  Econometric Models of Probabilistic Choice , 1981 .

[23]  J. Hausman.  Specification tests in econometrics , 1978 .

[24]  Y. Tse.  A Proportional Random Utility Approach to Qualitative Response Models , 1989 .

[25]  M. Gaudry,et al.  Dogit and Logit Models of Travel Mode Choice in Montreal , 1980 .

[26]  R. Strauss,et al.  The Prediction of Occupation Using Multiple Logit Models , 1975 .

[27]  Lea Vermeire,et al.  Locally optimal tests for multiparameter hypotheses , 1986 .