This dissertation is a pooled time series, cross-sectional, quantitative study of the impact of international financial transfer on the level of democracy. The study covers 174 developed and developing countries from 1976 through 1994. Through evaluating the democracy and democratization literature and other studies, the dissertation develops a theory and testable hypotheses about the impact of the international variables foreign aid and foreign direct investment on levels of democracy. This study sought to determine whether these two financial variables promote or nurture democracy and if so, how? A pooled time-series cross-sectional model is developed employing these two variables along with other relevant control variables. Control variables included the presence of the Cold War and existence of formal alliance with the United States, which account for the strategic dimension that might affect the financial transfer - level of democracy linkage. The model also includes an economic development variable (per capita Gross National Product) to account for the powerful impact for economic development on the level of democracy, as well as a control for each country's population size. By addressing and the inclusion of financial, economic, strategic, and population size effects, I consider whether change in these variables affect the level of democracy and in which direction. The dissertation tests this model by employing several techniques. The variables are subjected to bivariate and multivariate analysis including bivariate correlations, analysis of variance, and ordinary least square (OLS) multivariate regression with robust matrix and a lagged dependent variable. Panel corrected standard error (PCSE) was also employed to empirically test the pooled timeseries cross-sectional multivariate model. The dissertation analytical section concludes with path analysis testing which showed the impact of each of the independent variables on the dependent variable. The findings indicate less impact of international financial variables upon the level of democracy than hypothesized. Foreign assistance correlates negatively with economic development levels and has no effect on democracy levels. In contrast, foreign direct investment associates positively to economic development levels and, through increased economic development, contributes to democracy.

Methodologies for data reduction, modeling, and classification of grouped response curves are explored. In particular, the thesis focuses on the analysis of a collection of highly correlated, highly dimensional response-curve data of spectral reflectance curves of wood surface features. In the analysis, questions about the application of cross-validation estimation of discriminant function error rates for data that has been previously transformed by principal component analysis arise. Performing cross-validation requires re-calculating the principal component transformation and discriminant functions of the training sets, a very lengthy process. A more efficient approach of carrying out the cross-validation calculations, plus the alternative of estimating error rates without the re-calculation of the principal component decomposition, are studied to address questions about the cross-validation procedure. If populations are assumed to have common covariance structures, the pooled covariance matrix can be decomposed for the principal component transformation. The leave-one-out cross-validation procedure results in a rank-one update in the pooled covariance matrix for each observation left out. Algorithms have been developed for calculating the updated eigenstructure under rank-one updates and they can be applied to the orthogonal decomposition of the pooled covariance matrix. Use of these algorithms results in much faster computation of error rates, especially when the number of variables is large. The bias and variance of an estimator that performs leave-one-out cross-validation directly on the principal component scores (without re-computation of the principal component transformation for each observation) is also investigated. / Graduation date: 1993

Thesis (M. Phil.)--University of Hong Kong, 1986.

published_or_final_version / abstract / toc / Statistics / Doctoral / Doctor of Philosophy

Thesis (Ph. D.)--Auburn University. / Abstract. Vita. Includes bibliographical references (p. 91-103).

Discriminant analysis uses labelled observations to infer the labels of unlabelled observations in a population. Despite many advances in unsupervised and, to a lesser extent, semi-supervised learning over the past decade, discriminant analysis is often employed using approaches that date back to very well-known work of Fisher in the 1930s. One notable exception is mixture discriminant analysis, where the labels are estimated using parametric finite mixture models, commonly the Gaussian mixture model. The supposed advantage with mixture discriminant analysis is that multiple Gaussian components can be used for each class, hence providing a work around when a class is not Gaussian. This thesis makes several contributions to ``modern" discriminant analysis. Three robust discriminant analysis methods are introduced using mixtures of multivariate t-distributions, mixtures of multivariate power exponential distributions, and mixtures of contaminated Gaussian distributions, respectively. This provides an appealing framework for handling varying tail-weights and peakedness in the classes that may also contain mild outliers. To facilitate the modelling of asymmetric classes, we also explore robust discriminant analysis via finite mixtures of generalized hyperbolic distributions and mixtures of multivariate skew-t distributions. These approaches are tailored towards skewed classes but also have the added advantage of modelling symmetric classes where necessary. Finally, we introduce an approach that combines support vector machines with mixture discriminant analysis. This approach defines class boundaries in the labelled observations and, in some sense, improves mixture discriminant analysis performance. Crucially, in all of our mixture modelling work, we consider the case where the number of components per class is one. The utility of the approaches introduced is demonstrated on simulated and real data sets. / Thesis / Doctor of Philosophy (PhD)

The purpose of this study was to identify variables which discriminate between winning and losing a debate as measured by judges' responses on semantic differential scales. The dependent variable was membership in either the group "wins" or "losses." The independent variables were measured by semantic differential scales related to Delivery, Argument, and Evidence. The analytical procedure used was discriminant function analysis. Such an analysis discriminates maximally between the win and loss groups. Four scale items emerged as discriminating for wins and losses in a debate. The most discriminating variable came from the Argument dimension, specifically the scale item Convincing-Unconvincing. The second most discriminating variable was from the Evidence dimension, that is Strong-Weak. The third discriminating variable was from the Delivery dimension, namely Pleasant-Unpleasant. The last significant variable was also from the Evidence dimension, specifically Valuable-Worthless. The final Lambda of .5314 and the canonical correlation of .6845 indicate that the discriminant function produced a fairly high degree of separation between the win and loss groups.

published_or_final_version / Statistics / Master / Master of Philosophy

The identification of predictor variables that meaningfully contribute to group differences in Descriptive Discriminant Analysis (DDA) has had conflicting guidance in the historical quantitative psychological literature. Early simulation results that tested the bias and power of the standardized coefficients and the structural coefficients were ambiguous, yet a consensus still emerged that the structural coefficients were preferred. This study reviews the historical debate and known statistical weaknesses of both standardized coefficients and structure coefficients, summarizes relevant research and proposes a Monte Carlo study that will test whether the inclusion of standardized coefficients in interpreting DDA results for both the two-group and three-group cases can assist applied researchers in meaningfully ranking variables contributing to group differences. / text

Zhang, Xinmiao. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2007. / Includes bibliographical references (leaves 48). / Abstracts in English and Chinese. / Acknowledgement --- p.i / Abstract --- p.ii / Chapter 1 --- Introduction --- p.1 / Chapter 2 --- Estimation with Known Misclassification Probabilities --- p.3 / Chapter 2.1 --- Model --- p.3 / Chapter 2.2 --- Maximum Likelihood Estimation --- p.5 / Chapter 2.3 --- Statistical Property --- p.6 / Chapter 2.4 --- Mx Estimation --- p.7 / Chapter 2.5 --- Partition Maximum Likelihood (PML) Estimation --- p.9 / Chapter 2.6 --- Starting Value --- p.10 / Chapter 2.7 --- Examples --- p.11 / Chapter 2.7.1 --- Example 1 --- p.11 / Chapter 2.7.2 --- Example 2 --- p.12 / Chapter 2.7.3 --- Example 3 --- p.13 / Chapter 3 --- Estimation by Double Sampling --- p.15 / Chapter 3.1 --- Model and Analysis --- p.16 / Chapter 3.2 --- Statistical Property --- p.17 / Chapter 3.3 --- Mx Estimation and PML Estimation --- p.18 / Chapter 3.4 --- Starting Value --- p.19 / Chapter 3.5 --- Examples --- p.19 / Chapter 3.5.1 --- Example 4 --- p.19 / Chapter 4 --- Simulation --- p.20 / Chapter 4.1 --- Simulation with Known Misclassification Probability --- p.20 / Chapter 4.2 --- Simulation with Double Sampling --- p.22 / Chapter 5 --- Conclusion --- p.24 / Appendix and Tables --- p.26 / References --- p.48