Published
UNA COMPARACIÓN ENTRE LA INFERENCIA BASADA EN LAS ESTADÍSTICAS DE WALD Y RAZÓN DE VEROSIMILITUD EN LOS MODELOS LOGIT Y PROBIT VÍA MONTE CARLO
Keywords:
Modelo logit, modelo probit, error de especificación, simulación de Monte Carlo, test de razón de verosimilitud, test de Wald (es)Logit model, Probit model, Specification errors, Monte Carlo methods, Likelihood ratio test, Wald’s test (en)
Downloads
1Departamento de Estatística, Universidade Federal de Pernambuco, Cidade Universitária,
Recife/PE, 50740-540, Brasil. E-mail: artur@cox.de.ufpe.br
2Departamento de Estatística, Universidade Federal de Pernambuco, Cidade Universitária,
Recife/PE, 50740-540, Brasil. E-mail:hernando@cox.de.ufpe.br
Presentamos un estudio que evalúa y compara el desempeño de puebas de hipótesis e intervalos de confianza basados en la estadística de Wald con los basados en la estadística de razón de verosimilitud para los modelos logit y probit. Esta comparación se hace a través de las tasas de cobertura de los intervalos de confianza, tasas superior e inferior de los intervalos de confianza, y la potencia de la prueba de significancia. Se emplearon métodos de simulación de Monte Carlo. También se compararon las estadísticas de Wald y de razón de verosimilitud en los modelos logit y probit en presencia de errores en la especificación del modelo.
Palabras Clave: Modelo logit, modelo probit, error de especificación, simulación de Monte Carlo, test de razón de verosimilitud, test de Wald.
In this paper we present a study which evaluates and compares the performance of hypothesis testing and confidence intervals based on Wald"s statistic with those based on the likelihood ratio statistic for probit and logit models. We compare the rate of coverage of the confidence intervals, the maximum and minimum confidence interval coverage rate, and the power of the significance tests employing Monte Carlo simulation methods. The Wald and likelihood ratio statistics are also compared for the logit and probit models in the presence of errors in model specification.
Keywords: Logit model, Probit model, Specification errors, Monte Carlo methods, Likelihood ratio test, Waldás test.
Texto completo disponible en PDF
Referencias
1. Doornik, J. (2001), Ox: An Object-Oriented Matrix Language, 4 edn, Timberlake Consultants Press, London.
2. Hauck, W. W. & Donner, A. (1977), "Wald"s test as applied to hipotesis in logit analysis", JASA 72, 851- 853.
3. Hosmer, D. & Lemeshow, S. (1989), Applied Logistic Regression, Wiley & Sons, New York.
4. Jennings, D. E. (1986), "Judgind inference adequacy logistic regression", JASA 81, 471- 476.
5. McCullagh, P. & Nelder, J. (1989), Generalized Linear Models, Chapman & Hall, New York.
How to Cite
APA
ACM
ACS
ABNT
Chicago
Harvard
IEEE
MLA
Turabian
Vancouver
Download Citation
Article abstract page views
Downloads
License
Copyright (c) 2005 Revista Colombiana de Estadística
This work is licensed under a Creative Commons Attribution 4.0 International License.
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).