Introdução A mortalidade infantil (MI) no Município de São Paulo (MSP) apresenta queda, principalmente a partir da década de 80, entretanto é possível que existam diferenças regionais importantes entre Subprefeituras uma vez que estas apresentam características sócio-ambientais que podem influenciar neste indicador. Objetivo Descrever e analisar a evolução da MI no período de 2000 a 2007, segundo Subprefeituras do MSP. Métodos Estudo ecológico longitudinal, com 31 unidades de análise (Subprefeituras). Utilizou-se, para a análise estatística, o modelo de regressão linear multinível, considerando-se, como variável resposta, o CMI e oito anos de observação (2000 a 2007). O modelo incluiu variáveis relacionadas aos serviços de saúde. Resultados A queda da MI não ocorre de modo homogêneo entre as Subprefeituras evidenciadas pelas diferentes inclinações das retas e interceptos observados e estimados. Após a análise pelo modelo multinível observou-se redução da MI no período de 18,8% com declínio médio de 0,300/00nv ao ano Pelo modelo, 51% da variabilidade da MI se explica por características contextuais das Subprefeituras. No período de estudo, o CMI aumenta: 0,0560/00nv para cada 1% de aumento na proporção de mães com pré-natal inadequado, 0,2140/00nv para cada 1% de aumento na proporção da população exclusivamente usuária SUS, 0,0390/00nv para cada aumento na taxa de leitos obstétricos SUS. O CMI diminui: 0,1910/00nv para cada 1% de aumento na proporção de recuperação da vitalidade do nv. Conclusão A MI apresenta tendência de queda no período de 2000 a 2007 de modo não homogêneo segundo Subprefeitura. As variáveis que apresentaram associação com a MI são: o ano de observação, proporção de nascidos vivos de mães que realizaram até 6 consultas pré-natal (pré-natal inadequado); taxa de leitos obstétricos do Sistema Único de Saúde (SUS); proporção da população exclusivamente usuária do SUS e proporção de recuperação da vitalidade do nascido vivo. Na região periférica do MSP onde se encontram as maiores proporções da população exclusivamente usuária SUS, é também onde se apresentam os maiores CMI. / Introduction - Infant mortality (IM) in São Paulo (MSP) has declined, especially from the 80s. However, there may be important regional differences between Districts as their socio-environmental characteristics may influence this indicator. Objective - To describe and analyze IM trend in the period from 2000 to 2007, according to the Districts of MSP. Method - Ecological longitudinal study comprising 31 units of analysis (Districts). Linear multilevel regression model was used for statistical analysis. Infant Mortality Rate (IMR) and eight years of observation (2000-2007) were used as dependent variables. The model included variables related to health services. Results The decrease in IM does not occur homogeneously between Districts as evidenced by the different slopes and intercepts of the observed and estimated lines. A multilevel model showed an 18.8% reduction in IM in the period with an average decline of 0,300/00 living born (lb) per year. According to the model, 51% of the IM variability can be explained by contextual features of districts. During the study period, IMR increases: 0,0560/00lb for every 1% increase among mothers with inadequate prenatal care, 0,2140/00lb for every 1% increase among users of the Unified Health System (UHS), 0,0390/00lb for each increase in the UHS obstetric beds rate. IMR decreases 0,1910/00lb for every 1% increase in the vitality proportion of recovery of live births. Conclusion - IM shows a declining trend over the period from 2000 to 2007 in a non-homogeneous way according to District. The variables that were associated with IMR: year of observation, proportion of live births from mothers who had up to 6 prenatal visits (inadequate prenatal care), Unified Health System (UHS) obstetric beds rate, proportion of UHS users and vitality recovery of live birth proportion. The highest IMR is found in peripheral region of MSP where the largest proportion of UHS users is found.

This study investigates the team aspects of a process improvement project situated in an Internet banking system maintenance and modification phase. To investigate how team processes could be improved in this context, four interventional change techniques were introduced and implemented during an action research study in which the researcher was included as one of the team members and participated during group meetings and discussions. Internet banking is an important Internet-delivered service which is expected to provide benefits for both commercial banks and bank customers. Internet banking allows bank customers to have the freedom to perform their financial activities at their convenience. Developing, maintaining, and improving Internet banking systems requires large amounts of investment to maintain high levels of Internet banking service quality, and the maintenance and modification phase of the overall lifecycle cost is a considerable part of this investment. Therefore, in order to ensure high levels of usability, reliability, and quality for these Internet banking services, commercial banks need to make significant investments in the maintenance and modification phases of their Internet banking systems' lifecycle. The four interventional change techniques used in this study were: departmental participation, equal participation, holistic scenario, and management support. The four techniques were found to be influential in developing process improvements in the maintenance and modification phase of Internet banking systems. These techniques generated several significant improvements which directly affected the way team members managed their work. The significant contributions of these interventional change techniques were: the creation of cross-functional multilevel teams, development of effective departmental participation and communication techniques, extended scope and knowledge by the team members of Internet banking systems, an increase in team learning and understanding, techniques to change problem structure, and an end to end problem-solving approach. ii These contributions also directly improved the performance of the Internet banking systems maintenance team, and there was a significant improvement in the outcomes of the Internet banking systems maintenance and modification phase.

就台灣經濟、科技、風土民情及就業人口結構變動的狀況來看市場潛力極 大。但就社會現象與社會價值觀而論，存在著許多衝突。就業者而言，除 了已上軌道穩定成長的傳銷公司外，傳銷事業的經營策略、行銷策略等皆 有待強化。本研究基於國內多層次傳銷事業逢勃發展及國內對多層次傳銷 事業，有關事業面的相關研究欠缺。而多層次傳銷本質為一種行銷工具， 故本研究欲從理論及實務二方面了解多層次傳銷，並將多層次傳銷事業的 行銷策略加以分類。本研究印證國內多層次傳銷事業的意義與原理、在行 銷理論中的定位、多層次傳銷事業的成功要件，並將傳銷事業的行銷策略 加以比較，發現在國內傳銷事業在新產品決策方面、對直銷組織的管理、 產品的選擇與保證、促銷與廣告等決策存在類似的行銷策略。同時，傳銷 事業的行銷策略，也受在公司血統、公司規模、產品性質及差異化定位策 略的影響。

This study explored forms of distributed leadership in schools and the relationship between these forms and various personal and organizational factors. Though all empirical constructs used in this study were carefully grounded in the existing literature on the topic, the organization and empirical investigation of these constructs in a systemic manner is a distinct contribution of the study. The research was conducted in one large Ontario district that had been promoting distribute leadership for years. The evidence consisted of the surveys of 92 principals, 52 vice-principals, and 2190 teachers from the district’s elementary schools. Various statistical analyses, including descriptive, correlation, regression, and multi-level modeling analyses, were used to analyze the gathered data. In the study schools, administrators and teachers perceived leadership for high-priority school initiatives being mostly distributed in an egalitarian and coordinated manner. They also associated this coordinated way of distribution with wider distribution of leadership. At the same time, teachers often experienced leadership being distributed by principals in an authoritative manner as well. This finding indicates that principal delegation remains a dominant form of leadership and should be studied alongside more egalitarian distributed forms. This research brings to light the differences in principals’, vice-principals’, teacher-leaders’, and classroom teachers’ experiences with distributed leadership. It is recommended for researchers to take into account these differences in their studies of school leadership. District and school leaders should also consider staff’s varied leadership experiences while developing distributed leadership forms in schools. The study also suggests that school conditions have more effect on the way leadership gets distributed in schools than personal beliefs and experiences of principals, vice-principals, and teachers do. This implies that in order to develop extensive and coordinated distributed leadership forms, which are believed to be more productive, school administrators should devote considerable effort to encourage teachers’ engagement in leadership activities, build cultures of trust and collaboration, develop structures supporting shared decision making, ensure staff’s commitment to shared goals and provide incentives and relevant resources.

This study explored forms of distributed leadership in schools and the relationship between these forms and various personal and organizational factors. Though all empirical constructs used in this study were carefully grounded in the existing literature on the topic, the organization and empirical investigation of these constructs in a systemic manner is a distinct contribution of the study. The research was conducted in one large Ontario district that had been promoting distribute leadership for years. The evidence consisted of the surveys of 92 principals, 52 vice-principals, and 2190 teachers from the district’s elementary schools. Various statistical analyses, including descriptive, correlation, regression, and multi-level modeling analyses, were used to analyze the gathered data. In the study schools, administrators and teachers perceived leadership for high-priority school initiatives being mostly distributed in an egalitarian and coordinated manner. They also associated this coordinated way of distribution with wider distribution of leadership. At the same time, teachers often experienced leadership being distributed by principals in an authoritative manner as well. This finding indicates that principal delegation remains a dominant form of leadership and should be studied alongside more egalitarian distributed forms. This research brings to light the differences in principals’, vice-principals’, teacher-leaders’, and classroom teachers’ experiences with distributed leadership. It is recommended for researchers to take into account these differences in their studies of school leadership. District and school leaders should also consider staff’s varied leadership experiences while developing distributed leadership forms in schools. The study also suggests that school conditions have more effect on the way leadership gets distributed in schools than personal beliefs and experiences of principals, vice-principals, and teachers do. This implies that in order to develop extensive and coordinated distributed leadership forms, which are believed to be more productive, school administrators should devote considerable effort to encourage teachers’ engagement in leadership activities, build cultures of trust and collaboration, develop structures supporting shared decision making, ensure staff’s commitment to shared goals and provide incentives and relevant resources.

All Norwegian regions are represented with permanent offices and are engaged in different activities in the EU capital. This thesis investigates the regional and network level of EU policy-making, and asks the questions of what Norwegian regions are doing in Brussels; if are they are able to influence EU policy; and what effect participation in transnational policy networks have on their abilities to influence EU policy. To answer this, it applies a theoretical framework comprised of multi-level governance and the policy network approach to conduct an analysis of empirical data collected through interviews with different actors in Brussels. The present thesis argues that participation in transnational policy networks improve Norwegian regions' abilities to influence EU policy through being 'representations of scale'.

This dissertation is an empirical study of the political benefits of decentralization. It examines the effects of decentralization on citizens’ evaluations of the political system. Despite the large number of empirical studies on the costs and benefits of decentralization, most studies focus on economic benefits (typically in terms of fiscal efficiency) and pay little attention to potential political benefits. This dissertation seeks to fill these gaps by explicitly modeling the role decentralization plays in shaping citizens’ attitudes toward a political system. Drawing on work in political behavior and decentralization, a theoretical framework is developed to explain the manner in which citizens’ attitudes are shaped by election outcomes and their post-electoral win-loss status in multi-tier government. This dissertation not only offers a general argument with which to understand how a decentralized political structure may lead to greater stability in a democratic regime, but also offers guidance to policymakers on whether decentralization should be pursued as an option for institutional reform.

No description available.

My doctoral research dissertation focuses on two aspects of functional data analysis (FDA): FDA under spatial interdependence and FDA for multi-level data. The first part of my thesis focuses on developing modeling and inference procedure for functional data under spatial dependence. The methodology introduced in this part is motivated by a research study on inequities in accessibility to financial services. The first research problem in this part is concerned with a novel model-based method for clustering random time functions which are spatially interdependent. A cluster consists of time functions which are similar in shape. The time functions are decomposed into spatial global and time-dependent cluster effects using a semi-parametric model. We also assume that the clustering membership is a realization from a Markov random field. Under these model assumptions, we borrow information across curves from nearby locations resulting in enhanced estimation accuracy of the cluster effects and of the cluster membership. In a simulation study, we assess the estimation accuracy of our clustering algorithm under a series of settings: small number of time points, high noise level and varying dependence structures. Over all simulation settings, the spatial-functional clustering method outperforms existing model-based clustering methods. In the case study presented in this project, we focus on estimates and classifies service accessibility patterns varying over a large geographic area (California and Georgia) and over a period of 15 years. The focus of this study is on financial services but it generally applies to any other service operation. The second research project of this part studies an association analysis of space-time varying processes, which is rigorous, computational feasible and implementable with standard software. We introduce general measures to model different aspects of the temporal and spatial association between processes varying in space and time. Using a nonparametric spatiotemporal model, we show that the proposed association estimators are asymptotically unbiased and consistent. We complement the point association estimates with simultaneous confidence bands to assess the uncertainty in the point estimates. In a simulation study, we evaluate the accuracy of the association estimates with respect to the sample size as well as the coverage of the confidence bands. In the case study in this project, we investigate the association between service accessibility and income level. The primary objective of this association analysis is to assess whether there are significant changes in the income-driven equity of financial service accessibility over time and to identify potential under-served markets. The second part of the thesis discusses novel statistical methodology for analyzing multilevel functional data including a clustering method based on a functional ANOVA model and a spatio-temporal model for functional data with a nested hierarchical structure. In this part, I introduce and compare a series of clustering approaches for multilevel functional data. For brevity, I present the clustering methods for two-level data: multiple samples of random functions, each sample corresponding to a case and each random function within a sample/case corresponding to a measurement type. A cluster consists of cases which have similar within-case means (level-1 clustering) or similar between-case means (level-2 clustering). Our primary focus is to evaluate a model-based clustering to more straightforward hard clustering methods. The clustering model is based on a multilevel functional principal component analysis. In a simulation study, we assess the estimation accuracy of our clustering algorithm under a series of settings: small vs. moderate number of time points, high noise level and small number of measurement types. We demonstrate the applicability of the clustering analysis to a real data set consisting of time-varying sales for multiple products sold by a large retailer in the U.S. My ongoing research work in multilevel functional data analysis is developing a statistical model for estimating temporal and spatial associations of a series of time-varying variables with an intrinsic nested hierarchical structure. This work has a great potential in many real applications where the data are areal data collected from different data sources and over geographic regions of different spatial resolution.

Multi-level inverters play a significant role in high power drive systems for induction motors. Interest in multi-level inverters started with the three-level, neutral point clamped (NPC) inverter. Now there are many topologies for higher number of levels such as the, flying capacitor and cascaded H-bridge etc. The advantage of multi-level inverters is the reduced voltage stress on the switching devices, lower dv/dt and lower harmonic content. The voltage space vector structure in a multi-level inverter has a hexagonal periphery similar to that in a two-level inverter. In the over-modulation region in multi-level inverters, there is the presence of lower order harmonics such as 5th and 7th in the output voltage, and this can be avoided by using a voltage space vector scheme with more than six polygonal voltage space vectors such as 12, 18, 24 etc. These polygonal voltage space vectors can be generated by using multi-level inverter topologies, by cascading two-level inverter structures with asymmetric DC-links. This thesis deals with the development of 12-sided and 18-sided polygonal voltage space vector schemes for induction motor drives. With the 12-sided polygonal structure, all the 5th and 7th harmonic orders and 6n±1, n=1, 3, 5.. are absent throughout the modulation range, and in the 18-sided voltage space vector scheme, 5th, 7th, 11th and 13th harmonics are absent throughout the modulation range. With the absence of the low order frequencies in the proposed polygonal space vector structures, high frequency PWM schemes are not needed for voltage control. This is an advantage over conventional schemes. Also, due to the absence of lower order harmonics throughout the modulation range, special compensated synchronous reference frame PI controllers are not needed in current controlled vector control schemes in over-modulation. In this thesis a method is proposed for generating 12-sided polygonal voltage space vectors for an induction motor fed from one side. A cascaded combination of three two-level inverters is used with asymmetrical DC-links. A simple space vector PWM scheme based only on the sampled reference phase amplitudes are used for the inverter output voltage control. The reference space vector is sampled at different sampling rates depending on the frequency of operation. The number of samples in a sector is chosen to keep the overall switching frequency around 1kHz, in order to minimize switching losses. The voltage space vectors that make up the two sides of the sector in which the reference vector lies, are time averaged using volt-sec balance, to result in the reference vector. In the proposed 12-sided PWM scheme all the harmonics of the order 6n±1, n=1, 3, 5... are eliminated from the phase voltage, throughout the modulation range. In multi-level inverters steps are taken to eliminate common-mode voltage. Common-mode voltage is defined as one third of the sum of the three pole voltages of the inverter for a three phase system. Bearings are found to fail prematurely in drives with fast rising voltage pulses and high frequency switching. The alternating common-mode voltage generated by the PWM inverter contributes to capacitive couplings from stator body to rotor body. This generates motor shaft voltages causing bearing currents to flow from rotor to stator body and then to the ground. There can be a flashover between the bearing races. Also a phenomenon termed EDM (Electro-discharge machining) effect occurs and may damage the bearings. Common-mode voltage has to be eliminated in order to overcome these effects. In multi-level inverters redundancy of space vector locations is used to eliminate common-mode voltages. In the present thesis a 12-sided polygonal voltage space vector based inverter with an open-end winding induction motor is proposed, in which the common-mode voltage variation at the poles of the inverter is eliminated. In this scheme, there is a three-level inverter on each side of the open-end winding of the induction motor. The three-level inverter is made by cascading two, two-level inverters with unequal DC-link voltages. Appropriate space vectors are selected from opposite sides such that the sum of the pole voltages on each side is a constant. Also during the PWM operation when the zero vector is applied, identical voltage levels are used on both sides of the open-end windings, in order to make the phase voltages zero, while the common-mode voltage is kept constant. This way, common-mode voltage variations are eliminated throughout the modulation range by appropriately selecting the voltage vectors from opposite ends. In this method all the harmonics of 6n±1, n=1, 3, 5.. and triplen orders are eliminated. In the 12-sided polygonal voltage space vector methods, the 11th and 13th harmonics though attenuated are not eliminated. In the 18-sided polygonal voltage space vector method the 11th and 13th harmonics are eliminated along with the 5th and 7th harmonics. This scheme consists of an open-end winding induction motor fed from one side by a two-level inverter and the other side by a three-level inverter comprising of two cascaded two-level inverters. Asymmetric DC-links of a particular ratio are present. The 12-sided and 18-sided polygonal voltage space vector methods have been first simulated using SIMULINK and then verified experimentally on a 1.5kW induction motor drive. In the simulation as well as the experimental setup the starting point is the generation of the three reference voltages v, vB and vC . A method for determining the sector in which the reference vector lies by comparing the values of the scaled sampled instantaneous reference voltages is proposed. For the reference vector lying in a sector between the two active vectors, the first vector is to be kept on for T1 duration and the second vector for T2 duration. These timing durations can be found from the derived formula, using the sampled instantaneous values of the reference voltages and the sector information. From the pulse widths and the sector number, the voltage level at which a phase in the inverter has to be maintained is uniquely determined from look-up tables. Thus, once the pole voltages are determined the phase voltages can be easily determined for simulation studies. By using a suitable induction motor model in the simulation, the effect of the PWM scheme on the motor current can be easily obtained. The simulation studies are experimentally verified on a 1.5kW open-end winding induction motor drive. A V/f control scheme is used for the study of the drive scheme for different speeds of operation. A DSP (TMS320LF2407A) is used for generating the PWM signals for variable speed operation. The 12-sided polygonal voltage space vector scheme with the motor fed from a single side has a simple power bus structure and it is also observed that the pole voltage is clamped to zero for 30% of the time duration of one cycle of operation. This will increase the overall efficiency. The proposed scheme eliminates all harmonics of the order 6n±1, n=1, 3, 5…for the complete modulation range. The 12-sided polygonal voltage space vector scheme with common-mode elimination requires the open-end winding configuration of the induction motor. Two asymmetrical DC-links are required which are common to both sides. The leg of the high voltage inverter is seen to be switched only for 50% duration in a cycle of operation. This will also reduce switching losses considerably. The proposed scheme not only eliminates all harmonics of the order 6n±1, n=1, 3, 5…for the complete modulation range, but also maintains the common-mode voltage on both sides constant. The common-mode voltage variation is eliminated. This eliminates bearing currents and shaft voltages which can damage the motor bearings. In the 18-sided polygonal voltage space vector based inverter, the 11th and 13th harmonics are eliminated along with the 5th and 7th. Here also an open-end winding induction motor is used, with a two-level inverter on one side and a three-level inverter on the other side. A pole of the two-level inverter is at clamped to zero voltage for 50% of the time and a pole of the three-level inverter is clamped to zero for 30% of the time for one cycle of operation. The 18-sided polygonal voltage space vectors show the highest maximum peak fundamental voltage in the 18-step mode of 0.663Vdc compared to 0.658Vdc in the 12-step mode of the 12-sided polygonal voltage space vector scheme and 0.637Vdc in the six-step mode of a two-level inverter or conventional multi-level inverter (where Vdc is the radius of the space vector polygon). Though the schemes proposed are verified on a low power laboratory prototype, the principle and the control algorithm development are general in nature and can be easily extended to induction motor drives for high power applications.