No description available.

Most of empirical modeling involves the use of Ordinary Least Squares regression where the residuals are assumed normal, independent, and identically distributed. In finite samples, these assumptions becomes critical for accurate estimations, however, in macroeconomics in particular, these assumptions are rarely tested. This study addresses the applications of statistical testing methods and model respecification within the context of applied macroeconomics. The first application is a statistical comparison of Gregory Mankiw, David Romer and David Weil’s A Contribution to the Empirics of Economic Growth, and Nazrul Islam’s Growth Empirics: A Panel Data Approach. This analysis shows that the models in both papers are statistically misspecified. When respecified, the functional forms of Mankiw, Romer, and Weil’s models change considerably whereas Islam’s retain the theoretical structure. The second application is a study of the impact of inflation on investment and growth. After instrumenting for inflation with a set of political variables, I find that between approximately 1% and 9% inflation, there is a positive correlation between inflation and investment--the Mundell-Tobin effect may be a valid explanation. I further this analysis to show that treating investment as an exogenous variable may be problematic in empirical growth models. / Ph. D.

This thesis explores the theory and practice of sovereignty. I begin with a conceptual analysis of sovereignty, examining its theological roots in contrast with its later influence in contestations over political authority. Theological debates surrounding God’s sovereignty dealt not with the question of legitimacy, which would become important for political sovereignty, but instead with the limits of his ability. Read as an ontological capacity, sovereignty is coterminous with an existent’s activity in the world. As lived, this capacity is regularly limited by the ways in which space is produced via its representations, its symbols, and its practices. All collective appropriations of space have a nomos that characterizes their practice. Foucault’s account of “biopolitics” provides an account of how contemporary materiality is distributed, an account that can be supplemented by sociological typologies of how city space is typically produced. The collective biopolitical distribution of space expands the range of practices that representationally legibilize activity in the world, thereby expanding the conceptual limits of existents and what it means for them to act up to the borders of their capacity, i.e., to practice sovereignty. The desire for total authorial capacity expresses itself in relations of domination and subordination that never erase the fundamental precarity of subjects, even as these expressions seek to disguise it. I conclude with a close reading of narratives recounting the lives of residents in Chicago’s Englewood, reading their activity as practices of sovereignty which manifest variously as they master and produce space. / Ph. D.

Master of Science / Department of Statistics / Wei-Wen Hsu / The zero-inflated Poisson (ZIP) model consists of a Poisson model and a degenerate distribution at zero. Under this model, zero counts are generated from two sources, representing a heterogeneity in the population. In practice, it is often interested to evaluate this heterogeneity is consistent with the observed data or not. Most of the existing methodologies to examine this heterogeneity are often assuming that the Poisson mean is a function of nuisance parameters which are simply the coefficients associated with covariates. However, these nuisance parameters can be misspecified when performing these methodologies. As a result, the validity and the power of the test may be affected. Such impact of misspecification has not been discussed in the literature. This report primarily focuses on investigating the impact of misspecification on the performance of score test for homogeneity in ZIP models. Through an intensive simulation study, we find that: 1) under misspecification, the limiting distribution of the score test statistic under the null no longer follows a chi-squared distribution. A parametric bootstrap methodology is suggested to use to find the true null limiting distribution of the score test statistic; 2) the power of the test decreases as the number of covariates in the Poisson mean increases. The test with a constant Poisson mean has the highest power, even compared to the test with a well-specified mean. At last, simulation results are applied to the Wuhan Inpatient Care Insurance data which contain excess zeros.

In this thesis, two methods for assessing the functional form of covariates in the Cox proportional hazards model are evaluated. The methods include one graphical check based on martingale residuals and one graphical check, together with a Kolmogorov-type supremum test, based on cumulative sums of martingale residuals. The methods are evaluated in a simulation study under five different covariate misspecifications with varying sample sizes and censoring degrees. The results from both methods indicate that the type of covariate misspecification, sample size and censoring degree affect the ability to detect and identify the misspecification. The procedure based on smoothed scatterplots of martingale residuals reveals difficulties with assessing whether the behaviour of the smoothed curve in the plot is an indication of a misspecification or a phenomenon that can occur in a correctly specified model. The graphical check together with the test procedure based on cumulative sums of martingale residuals is shown to successfully detect and identify three out of five covariate misspecifications for large sample sizes. For small sample sizes, especially combined with a high censoring degree, the power of the supremum test is low for all covariate misspecifications.

No description available.

O ajuste de modelos de regressão beta requer a especificação de uma função de ligação. Algumas funções de ligação úteis são: logito, probito, complemento log-log e log-log. Usualmente, a ligação logito é utilizada pois permite interpretação simples para os parâmetros de regressão. O principal objetivo deste trabalho é avaliar o impacto da especificação incorreta da função de ligação em regressão beta. Estudos de simulação serão usados com esse prop´osito. Amostras da variável resposta serão geradas assumindo uma função de ligação conhecida (verdadeira) e o modelo de regressão beta será ajustado usando a função de ligação verdadeira (correta) e algumas funções de ligação incorretas. Resultados numéricos serão comparados para avaliar o efeito da especificação incorreta da função de ligação sobre as inferências em regressão beta. Adicionalmente, será introduzido um modelo de regressão beta com função de ligação de Aranda-Ordaz, a qual depende de um parâmetro que pode ser estimado através dos dados. / Fitting beta regression models requires the specification of the link function. Some useful link functions for beta regression are: logit, probit, complementary log-log and log-log. Usually, the logit link is used since it allows easy interpretation for the regression parameters. The main objective of this work is to evaluate the impact of misspecification of the link function in beta regression. Simulation studies will be used for this purpose. Samples of the response variable will be generated assuming a known (true) link function, and the beta regression will be fitted using the true (correct) link and some incorrect link functions. Numerical results will be compared to evaluate the effect of misspecification of the link function on inference in beta regression. Also, we will introduce a beta regression model with Aranda-Ordaz link function, which depends on an unknown parameter that can be estimated through the data.

A simulation study was conducted to explore the performance of the linear logistic test model (LLTM) when the relationships between items and cognitive components were misspecified. Factors manipulated included percent of misspecification (0%, 1%, 5%, 10%, and 15%), form of misspecification (under-specification, balanced misspecification, and over-specification), sample size (20, 40, 80, 160, 320, 640, and 1280), Q-matrix density (60% and 46%), number of items (20, 40, and 60 items), and skewness of person ability distribution (-0.5, 0, and 0.5). Statistical bias, root mean squared error, confidence interval coverage, confidence interval width, and pairwise cognitive components correlations were computed. The impact of the design factors were interpreted for cognitive components, item difficulty, and person ability parameter estimates. The simulation provided rich results and selected key conclusions include (a) SAS works superbly when estimating LLTM using a marginal maximum likelihood approach for cognitive components and an empirical Bayes estimation for person ability, (b) parameter estimates are sensitive to misspecification, (c) under-specification is preferred to over-specification of the Q-matrix, (d) when properly specified the cognitive components parameter estimates often have tolerable amounts of root mean squared error when the sample size is greater than 80, (e) LLTM is robust to the density of Q-matrix specification, (f) the LLTM works well when the number of items is 40 or greater, and (g) LLTM is robust to a slight skewness of the person ability distribution. In sum, the LLTM is capable of identifying conceptual knowledge when the Q-matrix is properly specified, which is a rich area for applied empirical research.

A semiparametric proportional likelihood ratio model was proposed by Luo and Tsai (2012) which is suitable for modeling a nonlinear monotonic relationship between the response variable and a covariate. Extending the generalized linear model, this model leaves the probability distribution unspecified but estimates it from the data. In this thesis, we propose to extend this model into analyzing the longitudinal data by incorporating random effects into the linear predictor. By using this model as the conditional density of the response variable given the random effects, we present a maximum likelihood approach for model estimation and inference. Two numerical estimation procedures were developed for response variables with finite support, one based on the Newton-Raphson algorithm and the other one based on generalized expectation maximization (GEM) algorithm. In both estimation procedures, Gauss-Hermite quadrature is employed to approximate the integrals. Upon convergence, the observed information matrix is estimated through the second-order numerical differentiation of the log likelihood function. Asymptotic properties of the maximum likelihood estimator are established under certain regularity conditions and simulation studies are conducted to assess its finite sample properties and compare the proposed model to the generalized linear mixed model. The proposed method is illustrated in an analysis of data from a multi-site observational study of prodromal Huntington's disease.

Two Monte Carlo studies were conducted to investigate the sensitivity of fit indices in detecting model misspecification in multilevel structural equation models (MSEM) with normally distributed or dichotomous outcome variables separately under various conditions. Simulation results showed that RMSEA and CFI only reflected within-model fit. In addition, SRMR for within-model (SRMR-W) was more sensitive to within-model misspecifications in factor covariances than pattern coefficients regardless of the impact of other design factors. Researchers should use SRMR-W in combination with RMSEA and CFI to evaluate the within-mode. On the other hand, SRMR for between-model (SRMR-B) was less likely to detect between-model misspecifications when ICC decreased. Lastly, the performance of WRMR was dominated by the misfit of within-model. In addition, WRMR was less likely to detect the misspecified between-models when ICC was relative low. Therefore, WRMR can be used to evaluate the between-model fit when the within-models were correctly specified and the ICC was not too small.