Governance of the healthcare systems is one of the most important challenges forWestern countries. Within this, an accurate assessment of the quality is key to policy makers and public managers, in order to guarantee equity, effectiveness and efficiency. In this thesis, we investigate aspects and methods related to healthcare evaluation by focussing on the healthcare system in Lombardy (Italy), where public and private providers compete with each other, patients are free to choose where to be hospitalized, and a pay-for-performance program was recently implemented. The general aim of this thesis is to highlight the role of statistics within a quality evaluation framework, in the form of advancing the statistical methods used to measure quality, of evaluating the effectiveness of implemented policies, and of testing the effect that mechanisms of competition and cooperation can have on the quality of a healthcare system. We firstly advance a new methodological approach for measuring hospital quality, providing a new tool for managers involved in performance evaluations. Multilevel models are typically used in healthcare, in order to account for the hierarchical structure of the data. These models however do not account for unobserved heterogeneity. We therefore propose an extension of the cluster-weighted models to the multilevel framework and focus in particular on the case of a binary dependent variable, which is common in healthcare. The resulting multilevel logistic cluster-weighted model is shown to perform well in a healthcare evaluation context. Secondly, we evaluate the effectiveness of a pay-for-performance program. Differently from the existent literature, in this thesis we evaluate this program on the basis of five health outcomes and across a wide range of medical conditions. Availability of data pre and post-policy in Lombardy allows us to use a difference-in-differences approach. The statistical model includes multiple dependent outcomes, that allow quantifying the joint effect of the program, and random effects, that account for the heterogeneity of the data at the ward and hospital level. The results show that the policy has overall a positive effect on the hospitals' performance. Thirdly, we study the effect of pro-competition reforms on the hospital quality. In Lombardy, competition between hospitals has been mostly driven by the adoption of a quasi-market system. Our results show that no association exists between hospital quality and competition. We speculate that this may be the result of asymmetric information, i.e. the lack of transparent information provided to citizens about the quality of hospitals. This is bound to reduce the impact of pro-competition reforms on quality and can in part explain the conflicting results found in the literature on this subject. Our results should motivate a public disclosure of quality evaluations. Regardless of the specifics of a system, hospitals are altruistic economic agents and they cooperate in order to improve their quality. In this work, we analyse the effect of cooperation on quality, taking the network of patients' transfers between hospitals as a proxy of their level of cooperation. Using the latest network models, we find that cooperation does lead to an increase in quality and should therefore be encouraged by policy makers.

Os modelos multiníveis são uma classe de modelos úteis na análise de bases de dados com estrutura hierárquica. No presente trabalho propõem-se os modelos multiníveis com resposta Weibull, nos quais são considerados interceptos aleatórios na modelagem dos dois parâmetros da distribuição da variável resposta. Os modelos aqui propostos são flexíveis devido a que a distribuição dos interceptos aleatórios pode der escolhida entre uma das seguintes quatro distribuições: normal, log--gama, logística e Cauchy. Uma extensão dos modelos é apresentada na qual é possível incluir na parte sistemática dos dois parâmetros da distribuição da variável resposta interceptos e inclinações aleatórias com distribuição normal bivariada. A estimação dos parâmetros é realizada pelo método de máxima verossimilhança usando a quadratura de Gauss--Hermite para aproximar a função de verossimilhança. Um pacote em linguagem R foi desenvolvido especialmente para a estimação dos parâmetros, predição dos efeitos aleatórios e para a obtenção dos resíduos nos modelos propostos. Adicionalmente, por meio de um estudo de simulação foi avaliado o impacto nas estimativas dos parâmetros do modelo ao assumir incorretamente a distribuição dos interceptos aleatórios. / Multilevel models are a class of models useful in the analysis of datasets with hierarchical structure. In the present work we propose multilevel Weibull models in which random intercepts are considered to model the two parameters of the Weibull distribution. The proposed models are flexible due to random intercepts distribution can be chosen from one of the four following distributions: normal, log-gamma, logistics and Cauchy. An extension of the models is presented in which we can include, in the systematic part of the two parameters of the distribution, random intercepts and slopes with a bivariate normal distribution. The parameter estimation is performed by maximum likelihood method using the Gauss Hermite quadrature to approximate the likelihood function. A package in R language was especially developed to obtain parameter estimation, random effects predictions and residuals for the proposed models. Additionally, through a simulation study we investigated the misspecification random effect distribution on estimated parameter for the proposed model

La tesis se inscribe en un doble ámbito científico formado por la Estadística y la Pedagogía. El objetivo de la tesis es la investigación de modelos estadísticos y estrategias de análisis que puedan ser de utilidad en el seguimiento de sistemas de evaluación complejos. Su motivación se encuentra en la necesidad de analizar las Pruebas de Aptitud para el Acceso a la Universidad (PAAU), que regulan el acceso a la universidad en España, desde la perspectiva de la ciencia estadística. La validez y fiabilidad de los exámenes COU (Curso de Orientación Universitaria) y PAAU han merecido una atención especial a lo largo de la investigación. Asimismo, se analizan con detenimiento las principales fuentes de variación de dichas notas: las diferencias entre centros de secundaria y el proceso de corrección de las pruebas PAAU.En la Introducción, una vez resumidas las características del sistema de evaluación de las pruebas PAAU y discutido el papel de la estadística en el tratamiento de datos en educación, se establecen los objetivos concretos de la tesis, a la luz de las necesidades existentes y de los trabajos de investigación realizados hasta el momento.El Capítulo 1 ilustra las diferencias entre los exámenes COU y las pruebas PAAU. Se aborda el estudio de la asociación entre ambas puntuaciones. La modelización de la variación de la nota PAAU individual por medio de modelos de regresión coeficientes aleatorios permite evidenciar (y medir) las diferencias entre centros de secundaria en cuanto a los estándares utilizados en COU. Este primer capítulo contiene una detallada introducción a los modelos de coeficientes aleatorios, también llamados modelos de nivel múltiple, que posteriormente se aplicaran en los capítulos 2 y 4, en la versión de modelos multivariantes de componentes de la varianza. El segundo capitulo, en un enfoque que complementa el anterior, se centra en el estudio de las medias (de COU y de PAAU) de cada centro, en la estructura de covarianza entre ambas. Como resultado relevante cabe citar la aplicación a la selección de la combinación más eficiente. El Capítulo 3 se ha dedicado enteramente a la calidad del sistema de corrección de los exámenes PAAU. La modelización presentada ha permitido evaluar el impacto de los correctores en términos de la varianza debida a las diferencias en el grado de severidad y a la varianza generada por la inconsistencia. Para la obtención de los datos se ha requerido del diseño de experimentos. Dichos experimentos, que han evidenciado una serie de puntos débiles del sistema, deberían ser realizados de manera sistemática cada año en una estrategia de mejora de la calidad del proceso. El Capítulo 4 estudia la covarianza del conjunto de notas PAAU tanto a nivel estudiante como a nivel centro, ofreciendo nuevos elementos de reflexión para la validez de dichas pruebas. El Capítulo 5 resume la aplicación de varias propuestas de la tesis a la primera convocatoria de las pruebas PAAU-LOGSE. El Capítulo 6 incluye las conclusiones de la tesis así como una serie de propuestas de seguimiento y mejora de la calidad global del sistema.

Longitudinal data arise when subjects are followed over time. This type of data is typically dependent, due to including repeated observations and this type of dependence is termed as within-subject dependence. Often the scientific interest is on multiple longitudinal measurements which introduce two additional types of associations, between-response and cross-response temporal dependencies. Only the statistical methods which take these association structures might yield reliable and valid statistical inferences. Although the methods for univariate longitudinal data have been mostly studied, multivariate longitudinal data still needs more work. In this thesis, although we mainly focus on multivariate longitudinal binary data models, we also consider other types of response families when necessary. We extend a work on multivariate marginal models, namely multivariate marginal models with response specific parameters (MMM1), and propose multivariate marginal models with shared regression parameters (MMM2). Both of these models are generalized estimating equation (GEE) based, and are valid for several response families such as Binomial, Gaussian, Poisson, and Gamma. Two different R packages, mmm and mmm2 are proposed to fit them, respectively. We further develop a marginalized multilevel model, namely probit normal marginalized transition random effects models (PNMTREM) for multivariate longitudinal binary response. By this model, implicit function theorem is introduced to explicitly link the levels of marginalized multilevel models with transition structures for the first time. An R package, bf pnmtrem is proposed to fit the model. PNMTREM is applied to data collected through Iowa Youth and Families Project (IYFP). Five different models, including univariate and multivariate ones, are considered to forecast multivariate longitudinal binary data. A comparative simulation study, which includes a model-independent data simulation process, is considered for this purpose. Forecasting independent variables are taken into account as well. To assess the forecasts, several accuracy measures, such as expected proportion of correct prediction (ePCP), area under the receiver operating characteristic (AUROC) curve, mean absolute scaled error (MASE) are considered. Mother&#039 / s Stress and Children&#039 / s Morbidity (MSCM) data are used to illustrate this comparison in real life. Results show that marginalized models yield better forecasting results compared to marginal models. Simulation results are in agreement with these results as well.

Thesis (M.A.)--University of Notre Dame, 2009. / Thesis directed by Sean Kelly for the Department of Sociology. "December 2009." Includes bibliographical references (leaves 59-62).

Not available / text

A relatively recent increase in the popularity of evidence-based activism has created a higher demand for statisticians to work on human rights and economic development projects. The statistical challenges of revealing patterns of violence in armed conflict require efficient use of the data, and careful consideration of the implications of modeling decisions on estimates. Impact evaluation of a complex economic development project requires a careful consideration of causality and transparency to donors and beneficiaries. In this dissertation, I compare marginal and conditional models for capture recapture, and develop new hierarchical models that accommodate challenges in data from the armed conflict in Colombia, and more generally, in many other capture recapture settings. Additionally, I propose a study design for a non-randomized impact evaluation of the Millennium Villages Project (MVP), to be carried out during my postdoctoral fellowship. The design includes small area estimation of baseline variables, propensity score matching, and hierarchical models for causal inference.

School Effectiveness Research (SER) is concerned with efforts to better understand the effectiveness enhancing relationship between student and school variables and how these variables primarily influence academic achievement (Scheerens, 2004). However, one identified methodological shortcoming in SER is the absence of cross-cultural perspectives (Kyriakides, 2006). This is a concern as what may prove effective in one nation does not necessarily mean that it can be easily and seamlessly imported into another with the same results. This study looked at the relationships between science self-beliefs and academic achievement in science across all nations who participated in the Programme for International Student Assessment (PISA) in 2006. It further explored the variance accounted for by cultural, social and economic capital (the elements of the PISA socioeconomic status variable) for each country in PISA 2006 when predicting scientific literacy. Lastly, it used hierarchical linear modeling (HLM) to analyze data from PISA 2006 for nations experiencing high rates of immigration (i.e., Germany, Spain, Canada, the United States, Australia and New Zealand). The outcome measures used for these countries were achievement scores in science, mathematics and reading. The variables examined at the student level were science self-efficacy, science self-concept, immigrant status and socioeconomic status. The variables examined at the school level were student level aggregates of school proportion of immigrants and school socioeconomic status. In the correlation analysis between science literacy and either science self-concept of science self-efficacy, findings suggest that at the student level, students with both higher science self-concept and higher science self-efficacy tend to achieve higher academically. However, at the country level the relationship was negative between self-concept and academic achievement in science (i.e., countries with higher science self-concept tend to achieve lower on scientific literacy). When the variables that comprised each of the cultural, social, and economic components of SES were regressed on scientific literacy for the PISA sample, cultural capital accounted for 16% of the variance in scientific literacy scores compared to 14% for social capital, 13% for the composite Economic Social and Cultural Status (ESCS), and 12% for economic capital. In the HLM null models, the intraclass correlations for the all countries except for Germany ranged from .16 to .29 (Germany’s was between .57 and .68). In the final models, at level-1 country, immigrant status tended to negatively influence achievement (i.e., non-native students are predicted to have lower performance), while science self-efficacy and science self-concept positively influenced achievement. The student level ESCS variable also impacted achievement positively. At the school level, level-2, school mean ESCS or school proportion of immigrants were found to significantly influence the level-1 predictors; however, a good deal of variability across nations was observed. The findings from this study demonstrate that there are some distinct national differences in the relationships between science self-beliefs, immigrant status and academic achievement.

Thesis (Ph.D.) -- University of Western Sydney, 2004. / "A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy" Bibliography : leaves 171-186.

Thesis (Ph. D.)--Ohio State University, 2008. / Title from first page of PDF file. Includes bibliographical references (p. 151-153).