Inferential Implications of Over-Parametrization: A Case Study in Incomplete Categorical Data

P>In the context of either Bayesian or classical sensitivity analyses of over-parametrized models for incomplete categorical data, it is well known that prior-dependence on posterior inferences of nonidentifiable parameters or that too parsimonious over-parametrized models may lead to erroneous conclusions. Nevertheless, some authors either pay no attention to which parameters are nonidentifiable or do not appropriately account for possible prior-dependence. We review the literature on this topic and consider simple examples to emphasize that in both inferential frameworks, the subjective components can influence results in nontrivial ways, irrespectively of the sample size. Specifically, we show that prior distributions commonly regarded as slightly informative or noninformative may actually be too informative for nonidentifiable parameters, and that the choice of over-parametrized models may drastically impact the results, suggesting that a careful examination of their effects should be considered before drawing conclusions.Resume Que ce soit dans un cadre Bayesien ou classique, il est bien connu que la surparametrisation, dans les modeles pour donnees categorielles incompletes, peut conduire a des conclusions erronees. Cependant, certains auteurs persistent a negliger les problemes lies a la presence de parametres non identifies. Nous passons en revue la litterature dans ce domaine, et considerons quelques exemples surparametres simples dans lesquels les elements subjectifs influencent de facon non negligeable les resultats, independamment de la taille des echantillons. Plus precisement, nous montrons comment des a priori consideres comme peu ou non-informatifs peuvent se reveler extremement informatifs en ce qui concerne les parametres non identifies, et que le recours a des modeles surparametres peut avoir sur les conclusions finales un impact considerable. Ceci suggere un examen tres attentif de l`impact potentiel des a priori.

Comparing diagnostic tests with missing data

When missing data occur in studies designed to compare the accuracy of diagnostic tests, a common, though naive, practice is to base the comparison of sensitivity, specificity, as well as of positive and negative predictive values on some subset of the data that fits into methods implemented in standard statistical packages. Such methods are usually valid only under the strong missing completely at random (MCAR) assumption and may generate biased and less precise estimates. We review some models that use the dependence structure of the completely observed cases to incorporate the information of the partially categorized observations into the analysis and show how they may be fitted via a two-stage hybrid process involving maximum likelihood in the first stage and weighted least squares in the second. We indicate how computational subroutines written in R may be used to fit the proposed models and illustrate the different analysis strategies with observational data collected to compare the accuracy of three distinct non-invasive diagnostic methods for endometriosis. The results indicate that even when the MCAR assumption is plausible, the naive partial analyses should be avoided.

Leverage analysis for linear mixed models

We consider a generalized leverage matrix useful for the identification of influential units and observations in linear mixed models and show how a decomposition of this matrix may be employed to identify high leverage points for both the marginal fitted values and the random effect component of the conditional fitted values. We illustrate the different uses of the two components of the decomposition with a simulated example as well as with a real data set.

Missing data mechanisms and their implications on the analysis of categorical data

We review some issues related to the implications of different missing data mechanisms on statistical inference for contingency tables and consider simulation studies to compare the results obtained under such models to those where the units with missing data are disregarded. We confirm that although, in general, analyses under the correct missing at random and missing completely at random models are more efficient even for small sample sizes, there are exceptions where they may not improve the results obtained by ignoring the partially classified data. We show that under the missing not at random (MNAR) model, estimates on the boundary of the parameter space as well as lack of identifiability of the parameters of saturated models may be associated with undesirable asymptotic properties of maximum likelihood estimators and likelihood ratio tests; even in standard cases the bias of the estimators may be low only for very large samples. We also show that the probability of a boundary solution obtained under the correct MNAR model may be large even for large samples and that, consequently, we may not always conclude that a MNAR model is misspecified because the estimate is on the boundary of the parameter space.

Likelihood ratio tests for variance components in linear mixed models

Although the asymptotic distributions of the likelihood ratio for testing hypotheses of null variance components in linear mixed models derived by Stram and Lee [1994. Variance components testing in longitudinal mixed effects model. Biometrics 50, 1171-1177] are valid, their proof is based on the work of Self and Liang [1987. Asymptotic properties of maximum likelihood estimators and likelihood tests under nonstandard conditions. J. Amer. Statist. Assoc. 82, 605-610] which requires identically distributed random variables, an assumption not always valid in longitudinal data problems. We use the less restrictive results of Vu and Zhou [1997. Generalization of likelihood ratio tests under nonstandard conditions. Ann. Statist. 25, 897-916] to prove that the proposed mixture of chi-squared distributions is the actual asymptotic distribution of such likelihood ratios used as test statistics for null variance components in models with one or two random effects. We also consider a limited simulation study to evaluate the appropriateness of the asymptotic distribution of such likelihood ratios in moderately sized samples. (C) 2008 Elsevier B.V. All rights reserved.

Predicting random effects with an expanded finite population mixed model

Prediction of random effects is an important problem with expanding applications. In the simplest context, the problem corresponds to prediction of the latent value (the mean) of a realized cluster selected via two-stage sampling. Recently, Stanek and Singer [Predicting random effects from finite population clustered samples with response error. J. Amer. Statist. Assoc. 99, 119-130] developed best linear unbiased predictors (BLUP) under a finite population mixed model that outperform BLUPs from mixed models and superpopulation models. Their setup, however, does not allow for unequally sized clusters. To overcome this drawback, we consider an expanded finite population mixed model based on a larger set of random variables that span a higher dimensional space than those typically applied to such problems. We show that BLUPs for linear combinations of the realized cluster means derived under such a model have considerably smaller mean squared error (MSE) than those obtained from mixed models, superpopulation models, and finite population mixed models. We motivate our general approach by an example developed for two-stage cluster sampling and show that it faithfully captures the stochastic aspects of sampling in the problem. We also consider simulation studies to illustrate the increased accuracy of the BLUP obtained under the expanded finite population mixed model. (C) 2007 Elsevier B.V. All rights reserved.

Beta-binomial/Poisson regression models for repeated bivariate counts

We analyze data obtained from a study designed to evaluate training effects on the performance of certain motor activities of Parkinson`s disease patients. Maximum likelihood methods were used to fit beta-binomial/Poisson regression models tailored to evaluate the effects of training on the numbers of attempted and successful specified manual movements in 1 min periods, controlling for disease stage and use of the preferred hand. We extend models previously considered by other authors in univariate settings to account for the repeated measures nature of the data. The results suggest that the expected number of attempts and successes increase with training, except for patients with advanced stages of the disease using the non-preferred hand. Copyright (c) 2008 John Wiley & Sons, Ltd.

Functional Analysis of saxophone, the Drosophila Gene Encoding the BMP Type I Receptor Ortholog of Human ALK1/ACVRL1 and ACVR1/ALK2

In metazoans, bone morphogenetic proteins (BMPS) direct a myriad of developmental and adult homeostatic evens through their heterotetrameric type I and type II receptor complexes. We examined 3 existing and 12 newly generated mutations in the Drosophila type I receptor gene, saxophone (sax), the ortholog of the human Activin Receptor-Like. Kinasel and -2 (ALK1/ACVR1 and ALK2/ACVR1) genes. Our genetic analyses identified two distinct classes of sax alleles. The first class consists of homozygous viable gain-of-function (GOF) alleles that exhibit (1) synthetic lethality in combination with mutations in BMP pathway components, and (2) significant maternal effect lethality that can be rescued by an increased dosage of the BMP encoding gene, dpp(+). In contrast, the second class consists of alleles that are recessive lethal and do not exhibit lethality in combination with mutations in other BMP pathway components. The alleles in this second class are clearly loss-of-function (LOF) with both complete and partial loss-of-function mutations represented. We find that one allele in the second class of recessive lethals exhibits dominant-negative behavior, albeit distinct from the GOF activity of the first class of viable alleles. On the basis of the fact that the first class of viable alleles can be reverted to lethality and on our ability to independently generate recessive lethal sat mutations, our analysis demonstrates that sax is an essential gene. Consistent with this conclusion, we find that a normal sax transcript is produced by sax(P), a viable allele previously reported to be mill, and that this allele can be reverted to lethality. Interestingly, we determine that two mutations in the first: class of sax alleles show the same amino acid substitutions as mutations in the human receptors ALK1/ACVR1-1 and ACVR1/ALK2, responsible for cases of hereditary hemorrhagic telangiectasia type 2 (HHT2) and fibrodysplasia ossificans progressiva (FOP), respectively. Finally, the data presented here identify different functional requirements for the Sax receptor, support the proposal that Sax participates in a heteromeric receptor complex, and provide a mechanistic framework for future investigations into disease states that arise from defects in BMP/TGF-beta signaling.

"Den stora stygga vargen?" : En argumentationsanalys om vargens moraliska värde i vargdebatten

The purpose of this paper is to see which ethical content arguments have in the debate about wolfhunting, particularly with regard to the Svensk Jakt respective Tidningen Djurens Rätt's arguments. The method used in this paper is the analysis of arguments in which I investigate the pro and contra arguments of the four authors of the articles that I have chosen to study. I will avail myself of theories from Carruthers and Franklin. Analysis shows that the authors of the Swedish hunting above all have a ecocentric view of the wolf where it is an instrumental value for the ecosystem, and the authors of Djurens Rätt's has especially a biocentric view of the wolf where it may be an absolute value, but they can also extend to the that the wolf has instrumental value in the ecocentric perspective. The discussion has shown that Djurens Rätt's and Svensk Jakt has a common position that is an ecocentric perspective where the wolf has instrumental value, even though they believe that different actions are different law to protect the species and the ecosystem.

Feasibility of the Reintroduction of Wolves in Maine: a GIS Study

The eastern timber wolf (Canis lupus lycaon) once inhabited Maine, as well as the rest of the eastern United States and southern Canada. As a result of human land use and widespread extermination campaigns, wolf numbers dramatically decreased, and by the early twentieth century, no wolves remained in Maine. As large carnivorous and territorial mammals, wolves require contiguous undeveloped areas with abundant prey. This project is a feasibility study that identifies the areas in Maine that are suitable for the reintroduction of wolves. We used GIS modeling to identify contiguous forested areas over 1,000 km2, calculate road and population density, and map the presence or absence of prey throughout the state. These variables were combined in a habitat suitability model to determine the location and amount of suitable wolf habitat in Maine. The northwestern part of the state appears most suitable for wolf reintroduction as it is relatively undeveloped with low road and population densities. There is also a smaller isolated area in Washington County that might be suitable, but further investigation is required.

Hip hop sul : um espaço televisivo de formação e atuação musical

O presente trabalho investigou o programa televisivo Hip Hop Sul, veiculado pela TV Educativa do Rio Grande do Sul, e a sua relação com os músicos que dele participam, buscando entender as funções sócio-musicais e as experiências de formação e atuação musical que o programa desempenha. As técnicas utilizadas na coleta de dados foram a entrevista semi-estruturada, observações e acompanhamento da rotina de trabalho dos produtores do programa e registros em fotografias digitais e em audiovisuais da atuação dos grupos de rap e da produção do programa. A investigação foi desenvolvida na perspectiva dos estudos culturais (Wolf, 1995) e a análise dos dados foi feita com base no método de análise televisiva de Casetti e Chio (1998). O trabalho fundamenta-se na perspectiva da televisão como mediadora de conhecimentos musicais, (Fischer, 1997, 2000a, 2000b; Kraemer, 2000; Nanni, 2000 e Souza, 2000), a partir dos quais foram evidenciados os aspectos musicais formativos e atuantes presentes no Hip Hop Sul. Os aspectos abordados referem-se as experiências musicais dos grupos de rap como telespectadores e participantes do programa. Os resultados dessa pesquisa mostram que para esses grupos, ver a si mesmo ou sentir-se representados na televisão, significa existir, ter uma identidade e sair do anonimato. Estar na televisão é, portanto, uma experiência que modifica o ser e o fazer musical.

Gênesis habermasiana e o caso de Catende: pragmática universal e a economia solidária

O presente trabalho se dedica à aproximação entre a pragmática universal, elemento fundamental da racionalidade comunicativa habermasiana, e os conceitos de economia solidária defendidos por Paul Singer. O trabalho, no que diz respeito à complexa teoria de Habermas, se limita à idéia de reconstrução das condições universais de possível compreensão mútua, portanto, não se dedica ao entendimento das relações semânticas nos procedimentos dialógicos, essência da teoria habermasiana. A reflexão teórica utilizou como instrumento empírico, através do estudo de caso não generalizável, o Projeto Harmonia no município de Catende, Zona da Mata pernambucana. Para este fim, foram utilizados dois estudos de campo: entrevistas com questionário semi-estruturado e pesquisa quantitativa com aplicação presencial de questionários. A pesquisa exploratória identificou quatro categorias de trabalhadores que foram analisados distintamente no levantamento de dados. O trabalho se desenvolve em seis capítulos, além da Introdução. O primeiro capítulo faz uma abordagem sumária das raízes da luta pela emancipação. O segundo capítulo trata da teoria habermasiana. O terceiro, descreve a erspectiva da economia solidária como opção ao capitalismo, e não necessariamente como objeto de ruptura paradigmática. O quarto capítulo se dedica à aproximação entre a economia solidária e o pragmatismo universal de Habermas. O quinto capítulo descreve o estudo de caso na Usina Harmonia. Finalmente, o sexto capítulo se dedica às conclusões. O trabalho conclui que há um laço que uni a teoria abermasiana à perspectiva da economia solitária, no que tange ao ambiente que funciona como reconstrução das condições para o entendimento mútuo. A conclusão se baseia no caso de Catende e não tem a pretensão de generalizar para outras experiências solidárias. Conclui-se também que o fato de observar as condições necessárias para o processo intersubjetivo não significa, necessariamente, que estão sendo desenvolvidos os procedimentos que identificam satisfatoriamente, nos termos de Habermas, a aplicação dos atos da fala, e de todas as características semânticas correspondentes.

Duas teorias econômicas: "consumo" e "retórica": subsídios para o estudo do campo científico da economia no Brasil

Então o panorama no primeiro ano era muito ruim. Tinha o Goffredo (da Silva Teles) que não era esse Goffredo de hoje. Era o Goffredo, vamos dizer, muito influenciado pelo romantismo jurídico, muito conservador. É um bom professor. Eu fui ficando numa "sinuca de bico" porque eu não podia largar a Faculdade. O que ia dizer para o meu pai? Um negócio complicadíssimo. Eu fui levando a faculdade, até que chegou o 20. semestre, eu e o (Luiz) Belluzzo - que fez a Faculdade comigo também - encontramos um amigo nosso que fazia filosofia, que é o Luiz Antônio de Oliveira Lima (professor da FGV-SP). E aí eu conversei: "- Lima, eu não aguento mais essa porcaria dessa faculdade. Quero fazer outra coisa. Aí, ele me disse: " Mas, por que você não vai fazer ciências sociais?" Nem sabia da existência desse curso. "Me explica esse negócio de ciências sociais." E fomos conversando ...então eu e o Belluzzo resolvemos prestar o vestibular para Ciências Sociais Esse sim era um curso estupendo, onde eu aprendi a raciocinar

What is the role of bookbuilding in bond allocation ? : evidence from Brazil

A dissertação tem como objetivo analisar dois aspectos do processo de bookbuilding nas emissões de debêntures no mercado brasileiro. O primeiro aspecto é verificar se o underwriter utiliza, a exemplo do que ocorre no Initial Public Offering (IPO) de ações, o poder discricionário nas alocações das debêntures entre os investidores. O segundo consiste em encontrar as características, tanto do emissor quanto do investidor, que influenciam na eficiência do bidder no processo de bookbuilding. Para realizar os testes empíricos foi utilizada uma base de dados composta por 40 books1 (totalizando 727 bids) fornecidos por um banco de investimento.Verifica-se que o underwriter não beneficia nenhum investidor na alocação final das debêntures. Essa afirmação fica evidenciada quando se calcula a diferença entre alocação final (efetivamente recebida pelo investidor) e alocação teórica (estimada com base no método pro-rata) para os 27 books (totalizando 557 bids) que apresentam demanda superior a oferta. A diferença é nula para 96.6% da amostra, sendo que das 19 observações não nulas, 15 possuem diferença absoluta de uma debênture entre a alocação teórica e a final, resultado explicado em função do arredondamento das alocações.Contrariando a teoria de leilão de titulos públicos, onde autores, como Scott and Wolf (1979), defendem que os investidores devem utilizar o step bid como estratégia ótima de bid, este trabalho verificou que no caso de bookbuilding de debêntures no mercado brasileiro, os investidores usuários de step bid posssuem menos chances de ter seu bid atendido plenamente pelo underwriter. Quando o investidor é um gestor de recursos de terceiros (asset management), aumenta-se a possibilidade de ter sua demanda atendida. O maior sucesso do asset management no bookbuilding deve-se às peculiaridades do mercado brasileiro: (i) somente investidores locais participam dos bookbuilding, já que investidores estrangeiros possuem preferência e incentivos por títulos públicos; (ii) gestores de recursos de terceiros representam 75% da demanda por debêntures; (iii) o mercado de gestão de recursos é concentrado: os 5 maiores gestores concentram 60% da indústria. Com isso os gestores de recursos podem desenvolver uma expertise própria, já que são os principais demandadores e frequentemente participam dos bookbuilding. As características do emissor também influenciam no desempenho dos bidders: as debêntures de baixo e médio risco aumentam a possibilidade do bidder ter seu pedido atendido na íntegra. Além disso, como era esperado, quanto maior for a demanda do título, mais dificil é para o investor conseguir a quantidade desejável.