Thesis (Ph.D.)--Ohio University, August, 2006. / Title from PDF t.p. Includes bibliographical references.

Orientador: Luiz Carlos de Freitas, Dalton Francisco de Andrade / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Educação / Made available in DSpace on 2018-08-10T21:12:26Z (GMT). No. of bitstreams: 1 Stevao_ChristianeBellorioGennarideAndrade_M.pdf: 5247957 bytes, checksum: 947de91bb797b7b1b693ceab7297dd14 (MD5) Previous issue date: 2008 / Resumo: Este trabalho teve como problema pesquisa mostrar os resultados obtidos aplicando a Teoria da Resposta ao Item aos dados do Projeto GERES Campinas, o qual utiliza todos itens de um teste como âncora para a criação da escala de proficiência, e assim comparar os resultados com a forma clássica de se elaborar escalas, o qual usa critérios específicos para aceitação de um item como item âncora. Para isso trabalhamos com os dados da pesquisa GERES, um estudo longitudinal com alunos de 1ª a 4ª série do ensino fundamental na cidade de Campinas, que teve seu início em março de 2005, nas três redes de ensino, Estadual, Municipal e Particular. Apresentamos o desempenho dos alunos nas três primeiras aplicações, no teste de matemática, e mostramos que há divergência entre as duas técnicas / Abstract: This work had as a research issue to show the results obtained by applying the Theory of the Response to the item to the database from the GERES' Project of Campinas. This Project utilizes all the items of a test as an anchor to make the scale of proficiency, and then to compare the obtained results with the classical way to elaborate scales. The latter uses specific criteria for the approval of an item as an anchor item. Therefore, we have analyzed the data from GERES' Project, which consists of a longitudinal study with 1st to 4th grade students of Elementary School of Campinas City (State, Municipal and Private School System) that started in March, 2005. We present the performance of the students in the Mathematic Test considering the first three applications of GERES Project and as a conclusion we found a divergence between the two methods / Mestrado / Ensino, Avaliação e Formação de Professores / Mestre em Educação

Thesis (PhD)--Stellenbosch University, 2012. / ENGLISH ABSTRACT: Multi-state models are used in this dissertation to model panel data, also known as longitudinal or cross-sectional time-series data. These are data sets which include units that are observed across two or more points in time. These models have been used extensively in medical studies where the disease states of patients are recorded over time. A theoretical overview of the current multi-state Markov models when applied to panel data is presented and based on this theory, a simulation procedure is developed to generate panel data sets for given Markov models. Through the use of this procedure a simulation study is undertaken to investigate the properties of the standard likelihood approach when fitting Markov models and then to assess its shortcomings. One of the main shortcomings highlighted by the simulation study, is the unstable estimates obtained by the standard likelihood models, especially when fitted to small data sets. A Bayesian approach is introduced to develop multi-state models that can overcome these unstable estimates by incorporating prior knowledge into the modelling process. Two Bayesian techniques are developed and presented, and their properties are assessed through the use of extensive simulation studies. Firstly, Bayesian multi-state models are developed by specifying prior distributions for the transition rates, constructing a likelihood using standard Markov theory and then obtaining the posterior distributions of the transition rates. A selected few priors are used in these models. Secondly, Bayesian multi-state imputation techniques are presented that make use of suitable prior information to impute missing observations in the panel data sets. Once imputed, standard likelihood-based Markov models are fitted to the imputed data sets to estimate the transition rates. Two different Bayesian imputation techniques are presented. The first approach makes use of the Dirichlet distribution and imputes the unknown states at all time points with missing observations. The second approach uses a Dirichlet process to estimate the time at which a transition occurred between two known observations and then a state is imputed at that estimated transition time. The simulation studies show that these Bayesian methods resulted in more stable results, even when small samples are available. / AFRIKAANSE OPSOMMING: Meerstadium-modelle word in hierdie verhandeling gebruik om paneeldata, ook bekend as longitudinale of deursnee tydreeksdata, te modelleer. Hierdie is datastelle wat eenhede insluit wat oor twee of meer punte in tyd waargeneem word. Hierdie tipe modelle word dikwels in mediese studies gebruik indien verskillende stadiums van ’n siekte oor tyd waargeneem word. ’n Teoretiese oorsig van die huidige meerstadium Markov-modelle toegepas op paneeldata word gegee. Gebaseer op hierdie teorie word ’n simulasieprosedure ontwikkel om paneeldatastelle te simuleer vir gegewe Markov-modelle. Hierdie prosedure word dan gebruik in ’n simulasiestudie om die eienskappe van die standaard aanneemlikheidsbenadering tot die pas vanMarkov modelle te ondersoek en dan enige tekortkominge hieruit te beoordeel. Een van die hoof tekortkominge wat uitgewys word deur die simulasiestudie, is die onstabiele beramings wat verkry word indien dit gepas word op veral klein datastelle. ’n Bayes-benadering tot die modellering van meerstadiumpaneeldata word ontwikkel omhierdie onstabiliteit te oorkom deur a priori-inligting in die modelleringsproses te inkorporeer. Twee Bayes-tegnieke word ontwikkel en aangebied, en hulle eienskappe word ondersoek deur ’n omvattende simulasiestudie. Eerstens word Bayes-meerstadium-modelle ontwikkel deur a priori-verdelings vir die oorgangskoerse te spesifiseer en dan die aanneemlikheidsfunksie te konstrueer deur van standaard Markov-teorie gebruik te maak en die a posteriori-verdelings van die oorgangskoerse te bepaal. ’n Gekose aantal a priori-verdelings word gebruik in hierdie modelle. Tweedens word Bayesmeerstadium invul tegnieke voorgestel wat gebruik maak van a priori-inligting om ontbrekende waardes in die paneeldatastelle in te vul of te imputeer. Nadat die waardes ge-imputeer is, word standaard Markov-modelle gepas op die ge-imputeerde datastel om die oorgangskoerse te beraam. Twee verskillende Bayes-meerstadium imputasie tegnieke word bespreek. Die eerste tegniek maak gebruik van ’n Dirichletverdeling om die ontbrekende stadium te imputeer by alle tydspunte met ’n ontbrekende waarneming. Die tweede benadering gebruik ’n Dirichlet-proses om die oorgangstyd tussen twee waarnemings te beraam en dan die ontbrekende stadium te imputeer op daardie beraamde oorgangstyd. Die simulasiestudies toon dat die Bayes-metodes resultate oplewer wat meer stabiel is, selfs wanneer klein datastelle beskikbaar is.

This study encompasses 25 kilometers of the Chehalis River in Washington, USA that currently has sections under a Total Maximum Daily Load (TMDL) plan for stream temperature impairments that exceed 18°C, a regulatory standard set at the time of the listing to protect salmonid spawning, rearing, and migration. Using information integrated from stationary data loggers (n=22) that collected stream temperature information from August 4-September 10, 2017, and longitudinal thermal profiling performed on July 29-30, August 4-5, and September 9-10, 2017, this study aimed to quantify the spatial distribution of stream temperature, evaluate relative consistencies of the riverine thermal regime over time, and identify which independent variables (land cover, aspect, canopy cover, impervious surfaces, channel width, discharge and air temperature) are correlated with stream temperature metrics using Spearman's rank correlation and stepwise linear regression modeling. Stream temperature was found to be strongly correlated with all air temperature metrics. The strongest model from stepwise linear regression (R2 = 0.711) found width, shrub/scrub, mixed forest, and cultivated crop land cover to be the strongest explanatory variables with the seven day average of the daily maximum stream temperature (7DADMaxTw) at the 22 sites. Tributaries had overall cooler average maximum stream temperatures than main stem sites. Thermal profiling identified seven cold-water patches (defined as the cumulative stream temperature ≥1°C cooler than the surrounding water). Integrating longitudinal thermal profiling and stationary data loggers allows resource managers to understand spatiotemporal stream temperature trends and influences and can assess more effective mitigation strategies to combat rising stream temperatures.

To investigate the effect of environmental pollution on the health of children in the Durban South Industrial Basin (DSIB) due to its proximity to industrial activities, 233 children from five primary schools were considered. Three of these schools were located in the south of Durban while the other two were in the northern residential areas that were closer to industrial activities. Data collected included the participants' demographic, health, occupational, social and economic characteristics. In addition, environmental information was monitored throughout the study specifically, measurements on the levels of some ambient air pollutants. The objective of this thesis is to investigate which of these factors had an effect on the lung function of the children. In order to achieve this objective, different sample survey data analysis techniques are investigated. This includes the design-based and model-based approaches. The nature of the survey data finally leads to the longitudinal mixed model approach. The multicolinearity between the pollutant variables leads to the fitting of two separate models: one with the peak counts as the independent pollutant measures and the other with the 8-hour maximum moving average as the independent pollutant variables. In the selection of the fixed-effects structure, a scatter-plot smoother known as the loess fit is applied to the response variable individual profile plots. The random effects and the residual effect are assumed to have different covariance structures. The unstructured (UN) covariance structure is used for the random effects, while using the Akaike information criterion (AIC), the compound symmetric (CS) covariance structure is selected to be appropriate for the residual effects. To check the model fit, the profiles of the fitted and observed values of the dependent variables are compared graphically. The data is also characterized by the problem of intermittent missingness. The type of missingness is investigated by applying a modified logistic regression model missing at random (MAR) test. The results indicate that school location, sex and weight are the significant factors for the children's respiratory conditions. More specifically, the children in schools located in the northern residential areas are found to have poor respiratory conditions as compared to those in the Durban-South schools. In addition, poor respiratory conditions are also identified for overweight children. / Thesis (M.Sc.)-University of KwaZulu-Natal, Pietermaritzburg, 2006.

This dissertation research aims to develop systemic methods to analyze mental disorder and social media use data in young adults in a dynamic way. The first part of the dissertation is a comprehensive review on modeling methods of longitudinal data. The second part describes the methods that we used to identify symptom clusters that can characterize treatment trajectories and to predict responses of anti-depressants for depression patients. Manhattan distance and bottom-up hierarchical clustering methods were used to identify the symptom clusters. Penalized logistic regressions were conducted to identify top baseline predictors of treatment outcomes. The third part presents of Tweedie distribution application with generalized linear models and growth mixed models for analyzing association between social media use patterns and mental health status. The fourth part is future work and research directions.

This dissertation examines the predictability of user engagement on online social platforms by integrating theoretical perspectives from the literature on media and technology habits with principles of context-aware computing. It presents three studies, each targeting a different facet of technology-mediated communication, from social media use in general to more granular behaviors like active and passive use and instant messaging. The first chapter proposes a novel approach to the study of social media habits through predictive modeling of sequential smartphone user behaviors. Using longitudinal smartphone app log data, it examines the predictability of app engagement as a way to capture a critical yet previously neglected aspect of media and technology habits: their embeddedness in repetitive behavioral sequences. The study employs Long Short-Term Memory (LSTM) and transformer neural networks to demonstrate that social media use follows predictable patterns over time and that its predictability varies substantially across individuals. T he second chapter shifts focus to the potential of context-aware modeling as a holistic yet parsimonious and privacy-preserving approach to predicting user engagement on online social platforms. Analyzing over 100 million Snapchat sessions from nearly 80,000 users via deep LSTM neural networks, the study demonstrates the predictability of active and passive use based on past behavior and a notable improvement in predictive performance upon integrating momentary context information. Features related to connectivity status, location, temporal context, and weather were found to capture non-redundant variance in user engagement relative to features derived from histories of in-app behaviors. The findings are consistent with the idea of context-contingent, habit-driven patterns of active and passive use, highlighting the utility of contextualized representations of user behavior for predicting user engagement on online social platforms. The third chapter investigates the predictability of attentiveness and responsiveness in instant messaging on a large online social platform. Utilizing metadata from over 19 million messages, the study examines the predictive power of a wide range of predictor groups, including message attributes, user attributes, and momentary context, as well as historical communication patterns within ego networks and dyadic relationships. The findings echo the overarching theme that habitual behaviors and contextual factors shape user engagement. However, in this case, dyad-specific messaging histories account for the overwhelming share of explained variance, underlining the socially interdependent nature of user engagement in instant messaging. Collectively, the three studies presented in this dissertation make a theoretical contribution by establishing media and technology habits as a suitable framework for the study of user engagement and by introducing a novel perspective that emphasizes the repetitive, predictable, and context-dependent nature of media and technology habits. The research makes an important empirical contribution through the use of novel, large-scale, objective behavioral data, enhancing the ecological validity and real-world applicability of its findings. Methodologically, it pioneers the use of context-aware sequential machine learning techniques for the study of media and technology habits. The insights garnered from this research have the potential to inform the design of engaging and ethical online social platforms and mobile technologies, highlighting its practical implications for the billions of users navigating these digital environments on a daily basis.

As residential mobility recursively links individual life courses and the characteristics of places, it is unsurprising that geographers have long sought to understand how people make moving decisions. However, much of our knowledge of residential mobility processes derives from cross-sectional analyses of either mobility decision-making or moving events. Comparatively few studies have linked these separate literatures by analysing how residential (im)mobility decisions unfold over time within particular biographical, household and spatio-temporal contexts. This is problematic, as life course theories suggest that people frequently do not act in accordance with their underlying moving desires. To evaluate the extent to which residential (im)mobility is volitional or the product of constraints therefore requires a longitudinal approach linking moving desires to subsequent moving behaviour. This thesis develops this longitudinal perspective through four linked empirical studies, which each use British Household Panel Survey data to analyse how the life course context affects the expression and realisation of moving desires. The first study investigates how people make moving decisions in different ways in response to different motivations, triggers and life events. The second study harnesses the concept of ‘linked lives', exploring the extent to which the likelihood of realising a desire to move is dependent upon the desires of a person's partner. The third study analyses the biographical dimension of mobility decision-making, investigating how the long-term trajectories of life course careers are associated with particular mobility biographies. The final empirical chapter develops these insights, exploring the duration and abandonment of moving desires. Taken together, these studies test and extend conceptual models of mobility decision-making by empirically engaging with neglected facets of life course theories. Fundamentally, the thesis uncovers how aggregate mobility patterns are produced by the interactions between individual choices and multi-scalar constraints.

Orientadores: Luiz Carlos de Freitas, Dalton Francisco de Andrade / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Educação / Made available in DSpace on 2018-08-24T22:44:15Z (GMT). No. of bitstreams: 1 Dalben_Adilson_D.pdf: 5011742 bytes, checksum: e9c6413b4e6fb98c276dcbdecd13440b (MD5) Previous issue date: 2014 / Resumo: Esta pesquisa é um estudo sobre a eficácia e equidade escolar que tem ganhado atenção especial nos países que usam as avaliações em larga escala a serviço da gestão do sistema educativo. No Brasil, que desde a década de 1990 colocou a avaliação educacional como recurso central em suas políticas educacionais, mas coletando dados seccionais, que são muito frágeis para essa finalidade. Essa fragilidade decorre da alta associação que os fatores extraescolares, sobretudo o nível socioeconômico do aluno, têm sobre as medidas de proficiência. Diante disso, foram usados dados longitudinais e a análise foi feita por meio de modelos lineares hierárquicos. Esta pesquisa teve como objetivo principal desenvolver um modelo estatístico capaz de identificar tais fatores para a realidade brasileira, considerando que a aprendizagem é um processo complexo, isto é, ela é influenciada simultaneamente por múltiplos fatores. Foram desenvolvidos modelos de valor agregado que não só identificam tais variáveis, como também caracterizam sua influência em alunos com distintas proficiências no início de cada período de escolarização. A base de dados utilizada nesses modelos foi fornecida pelo Projeto GERES, que, no período de 2005 a 2008, coletou dados dos mesmos alunos de 1ª a 4ª séries de uma amostra de 312 escolas em cinco grandes cidades brasileiras. Foram medidas as proficiências em Leitura e Matemática de 35.538 alunos e coletadas informações de contexto desses alunos, seus familiares, professores, diretores e escola. Após a redução do grande número de informações disponibilizadas pelo Projeto GERES, feita por meio da Análise Fatorial Exploratória (AFE), as variáveis resultantes foram reorganizadas em três arquivos usados para análise em modelos lineares hierárquicos de três níveis. Os resultados encontrados evidenciam uma significativa instabilidade nos efeitos que as variáveis têm sobre a proficiência, tanto em leitura quanto em matemática. Ao final da pesquisa, são encontrados alguns fatores que influenciam positivamente e negativamente a proficiência em Leitura e Matemática e outros que afetam especificamente cada uma dessas áreas, indicando que podem colaborar para o aumento da eficácia e da equidade das escolas. No entanto, constatam-se também algumas variáveis que têm comportamentos incoerentes com o esperado e outras com comportamentos opostos nas duas áreas. Assim, dos achados das pesquisas, comprova-se que, com base nos dados utilizados, procedimentos metodológicos e modelos estatísticos adotados, os modelos de valor agregado melhoram a confiabilidade das análises em comparação aos modelos que usam dados seccionais, mas ainda são inviáveis como ferramentas para a gestão do sistema educativo, sobretudo para o uso meritocrático de seus resultados. Dessa forma, esta pesquisa corrobora os achados de outras realizadas no âmbito internacional e permite afirmar que a qualidade da modelagem estatística depende da qualidade dos dados que busca modelar, podendo gerar distorções, estabelecer relações inesperadas ou levar a conclusões equivocadas. Em contrapartida, trata-se de recursos que podem ser usados no sistema educativo, fornecendo dados importantes para a orientação das políticas públicas numa perspectiva de avaliação formativa, com vistas ao melhoramento da qualidade de ensino oferecido pelas escolas e à melhor formação dos profissionais docentes e não-docentes que nelas trabalham / Abstract: This research is a study on school effectiveness and equality in Brazil, adding up to a number of other researches that have drawn special attention in countries that use large-scale evaluations at the service of the education system management. In the Brazil has regarded the educational evaluation as a central resource in national education policies, but using cross-sectional data, which are far more fragile for such purpose. This fragility has derived from the great influence that extra-school factors, particularly the students¿ socioeconomic status, exerts on proficiency measures. Longitudinal data was used in the analyses with hierarchical linear models. The main objective of this research was to develop a statistical model to identify such factors in the Brazilian reality, considering that learning is a complex process, i.e. it is simultaneously influenced by multiple factors. Value-added models were developed not only to identify such variables, but also to characterize their influence on students showing different proficiencies at the beginning of every school term. The data base used in those models was provided by the GERES Project, which collected data of the same students from the 1st to the 4th grade from a sample of 312 schools in five Brazilian cities from 2005 to 2008. Proficiencies of 35,538 students were measured, and information about these students¿ context, family, teachers, principals and school were gathered. After the reduction of the great amount of information made available by the GERES Project by means of Exploratory Factor Analysis (EFA), the resulting variables were reorganized in three files used for analysis in three-level hierarchical linear models. The results evidenced significant instability in the effects that the variables have on proficiency both in Reading and in Mathematics. At the end of the research, some factors that influence Reading and Mathematics proficiency either positively or negatively, as well as other factors that specifically affect one of those areas, were found, thus indicating that they may contribute to increased school effectiveness and equality. However, some variables whose behavior was inconsistent with the one expected, and others with opposite behaviors in the two areas were also found. Therefore, from the research findings, based on the data used, the methodological procedures and the statistical models adopted, it has been evidenced that value-added models improve the analysis reliability in comparison with models that use cross-sectional data, but they are still impracticable as tools for education system management, particularly for meritocratic use of their results. Hence, this research has corroborated the findings of other studies carried out over the world and has enabled us to state that the quality of the statistical modeling depends on the quality of data that it attempts to model, and it may generate distortions, establish unexpected relationships or lead to misleading conclusions. On the other hand, these resources may be used in the education system by providing important data for guiding public policies in a educative evaluation perspective, aiming at improving the quality of teaching offered by schools, teachers and other professionals that work in the school setting / Doutorado / Ensino e Práticas Culturais / Doutor em Educação

No description available.