• JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  Bookmark and Share
Master's Dissertation
Full name
Antonio Carlos Simões Pião
Knowledge Area
Date of Defense
Piracicaba, 1989
Title in Portuguese
Análise de modelos lineares em dados de contagens binomiais negativas, usando dados originais ou transformados para normalidade e homocedasticidade
Keywords in Portuguese
Abstract in Portuguese
Simularam-se 1000 ensaios para cada uma das 112 combinações de 4 populações (tratamentos), englobando casos de populações iguais e casos com diferenças em m, k ou ambos. Deve-se ter cuidado ao aplicar transformações de dados, particularmente se não há homogeneidade de k. A estatística C(α) proposta por BARNWAL & PAUL (1988), mostrou alguma robustez para valores não homogêneos de k, conduzindo a resultados equivalentes àqueles obtidos usado dados não transformados. A análise de variância, usado o teste de mínimo qui-quadrado XU2 mostrou ser viesado, superestimando valores quando a matriz de variâncias e covariâncias é desconhecida. Se a matriz de variâncias e covariâncias é conhecida, os resultados são equivalentes a aqueles obtidos dos dados originais. Resultados similares foram obtidos para populações menores, n=10, quando um poder decrescente do teste foi detectado. Foram escolhidos 20 casos e simularam-se 1000 ensaios para cada caso.
Title in English
Linear models analysis of negative binomial counts using original, and transformed data for normality and homocedasticity
Abstract in English
The well-known negative binomial distribution is quite frequently used to interpret counting variables, through different techniques. In order to compare these techniques, four populations of size n=50 were computer-generated for different values of the parameters m and k, using NORMAN & CANNON (1972) procedure. Comparisons of the transformations of variables, as suggested by BARBOSA (1985), were used. For that, 1000 essays were simulated for each one of the 112 combinations of 4 populations. This covered equal and different populations with respect to the parameters m, k or both. In conclusion, for some values of the parameters m and k, there is no necessity of any data transformation, particularly if depending of k. Statistics like C(α) proposed by BARNWAL & PAUL (1988) showed some robustness for non-homogeneous values of k, leading to equivalent results to that ones obtainned using untransformed data. The analysis of variance, using the minimum chi-square test U2 showed to be biased superestimating values when the variance-covariance matrix is unknown. If the variance-covariance matrix is knew the results are equivalent from those obtainned from original data. Similar results were obtainned for smaller populations, n=10, when a decreasing power of the tests was detected. In such a case 20 combinations and 1000 simulations for each combination were performed
WARNING - Viewing this document is conditioned on your acceptance of the following terms of use:
This document is only for private use for research and teaching activities. Reproduction for commercial use is forbidden. This rights cover the whole data about this document as well as its contents. Any uses or copies of this document in whole or in part must include the author's name.
Publishing Date
WARNING: Learn what derived works are clicking here.
All rights of the thesis/dissertation are from the authors
Digital Library of Theses and Dissertations of USP. Copyright © 2001-2023. All rights reserved.