Many epidemiological studies assess the effects of time-dependent exposures, where both the exposure status and its intensity vary over time. The analysis of such studies poses the challenge of modelling the association between complex time-dependent drug exposure and the risk, especially given the uncertainty about the etiological relevance of doses taken in different time periods. / To address this challenge, I developed a flexible method for modelling cumulative effects of time-varying exposures, weighted by recency, represented by time-dependent covariates in the Cox proportional hazards model. The function that assigns weights to doses taken in the past is estimated using cubic regression splines. Models with different number of knots and constraints are estimated. Bootstrap techniques are used to obtain pointwise confidence bands around the weight functions, accounting for both the sampling variation of the regression coefficients, and the uncertainty at the model selection stage, i.e. the additional variance due to a posteriori selection of the number of knots. / To assess the method in simulations, I had to develop and validate a novel algorithm to generate event times conditional on time-dependent covariates and compared it with the algorithms available in the literature. The proposed algorithm extends a previously proposed permutational algorithm to include a rejection sampler. While all the algorithms generated data sets that, once analyzed, provided virtually unbiased estimates with comparable variances, the algorithm that I proposed reduced the computational time by more than 50 per cent relative to alternative methods. I used simulations to systematically investigate the properties of the weighted cumulative dose method. Six different weight functions were considered. Simulations showed that in most situations, the proposed method was able to capture the shape of the true weight functions and to produce estimates of the magnitude of the exposure effect on the risk that were close to those used to generate the data. I finally illustrated the use of the weighted cumulative dose modelling by reassessing the association between the use of selected benzodiazepines and fall-related injuries, using administrative data on a cohort of elderly who initiated their use of benzodiazepines between 1990 and 2004.

Submitted by SÉRGIO RICARDO LIMA NEGRO (limanegro@gmail.com) on 2017-12-18T23:04:47Z No. of bitstreams: 1 Tese Sérgio Ricardo Lima Negro final.pdf: 15461052 bytes, checksum: 652aaf7a76daf5705197093369cbb997 (MD5) / Rejected by Cristina Alexandra de Godoy null (cristina@adm.feis.unesp.br), reason: Solicitamos que realize correções no arquivo submetido seguindo as orientações abaixo: 1. Verificar com o João (joao@adm.feis.unesp.br) a normalização do seu trabalho conforme as normas Abnt, por exemplo: título acima das figuras, retirar pontuação entre o número e o título da seção e subseção, alinhamento do sumário, alterar a capa e etc., seguindo as orientações da biblioteca. 2. Inserir a imagem escaneada da folha de aprovação no local adequado. 3. Inserir a ficha catalográfica elaborada pela biblioteca após a normalização do seu trabalho. Agradecemos a compreensão. on 2017-12-19T11:06:29Z (GMT) / Submitted by SÉRGIO RICARDO LIMA NEGRO (limanegro@gmail.com) on 2017-12-20T13:58:16Z No. of bitstreams: 1 tese definitiva corrigida.pdf: 12662965 bytes, checksum: 993de583d00cdfbd324c80a85757183f (MD5) / Approved for entry into archive by Cristina Alexandra de Godoy null (cristina@adm.feis.unesp.br) on 2017-12-20T18:28:52Z (GMT) No. of bitstreams: 1 negro_srl_dr_ilha.pdf: 12662965 bytes, checksum: 993de583d00cdfbd324c80a85757183f (MD5) / Made available in DSpace on 2017-12-20T18:28:52Z (GMT). No. of bitstreams: 1 negro_srl_dr_ilha.pdf: 12662965 bytes, checksum: 993de583d00cdfbd324c80a85757183f (MD5) Previous issue date: 2017-08-18 / A variabilidade espacial dos atributos físicos do solo é importante indicador de manejo localizado nas áreas agrosilvopastoris. No ano agrícola de 2009/10 em Selvíria (MS), analisaram-se os componentes de produção da soja e atributos físicos da relação massa/volume de um Latossolo Vermelho Distroférrico em plantio direto, com objetivo de encontrar correlações lineares e espaciais entre eles. Foi instalada uma malha geoestatística para a coleta dos dados, totalizando 99 pontos amostrais numa área de 10 ha. Os atributos do solo, nas camadas de 0,00-0,10 e 0,10-0,20 m foram: a densidade do solo, DS (métodos do anel volumétrico e do torrão parafinado), densidade de partículas do solo, DP (métodos do balão volumétrico e do balão volumétrico modificado) e a porosidade total do solo, PT, utilizando os valores de DS e DP dos diferentes métodos de determinação, calculada pela fórmula PT= (1-DS/DP). Os componentes de produção da soja foram: número de vagens por planta (NVP), número de grãos por vagem (NGV), massa de cem grãos (MCG), massa de grãos por planta (MGP), população de plantas (POP), altura de plantas (ALT) e produtividade de grãos obtida (PGO). Alguns dos componentes de produção da soja e dos atributos físicos do solo revelaram dependência espacial, possibilitando mapear a área de produção. Assim, os alcances geoestatísticos recomendados para futuras pesquisas deverão estar entre 273 e 526,5 m. Espacialmente, foi possível estimar a PGO pela co-krigagem com a MGP; com a DS (método do anel volumétrico) de 0,00-0,10 m; com a PT, calculada pela relação entre a DS (método anel volumétrico)/DP (método do balão volumétrico) de 0,10-0,20 m; e com a PT, calculada pela DS (método do anel volumétrico)/DP (método do balão volumétrico modificado) de 0,00-0,10 m. A MGP pode ser estimada pela co-krigagem com a DS, quando determinada pelo método do anel volumétrico, na camada de 0,00-0,10 m. Portanto, foi possível estimar a variabilidade espacial da PGO e da MGP e mapear a área, a fim de propor estratégias de manejo visando aumentar a produtividade da soja. / The soil physical attributes spatial variability is an important indicator of localized management practices in the agrosilvopastoral areas. In the 2009/2010 agricultural year in Selvíria (MS), soybean yield components and physical attributes of the mass/volume ratio of a dystroferric Red Latosol (Typic Acrustox) under a no-tillage system were analyzed in order to find linear and spatial correlations between them. It was installed a geostatistical grid to collect the data, totalizing 99 sampling points in an area of 10 ha. The soil attributes in the 0.00-0.10 and 0.10-0.20 m layers were: soil bulk density, BD (volumetric ring and paraffin sealed clod methods), soil particle density, PD (volumetric flask and modified volumetric flask methods) and total soil porosity, TP, calculated by the resulting ratio between BD and PD, using the formula TP = (1-BD/PD). The components of soybean yield were: number of pods per plant, number of grains per pod, mass of 100 grains, grain mass per plant, plant population, plant height and soybean yield. Some of the determined soybean yield components and some of the physical soil attributes revealed spatial dependence, making it possible to map the yield area. Thus, the recommended geostatistical ranges for further researches thus shall be between 273 and 526.5 m. Spatially, it was possible to estimate the soybean yield by co-kriging it with the grain mass per plant; and with the soil bulk density (volumetric ring method) of 0.00-0.10 m; and with the total soil porosity, calculated by the ratio between soil bulk density (volumetric ring method)/soil particle density (volumetric flask method) of 0.10-0.20 m; and with the total soil porosity, calculated by the soil bulk density (volumetric ring method)/soil particle density (modified volumetric flask method) of 0.00-0.10 m. The grain mass per plant could be estimated by co-kriging it with the soil bulk density (volumetric ring method) of 0.00-0.10 m. Therefore, it was possible to estimate the spatial variability of soybean yield and of grain mass per plant in order to map the area to propose management strategies aiming to increase soybean yield.

Submitted by Irene Nascimento (irene.kessia@ufpe.br) on 2016-08-22T18:54:13Z No. of bitstreams: 2 license_rdf: 1232 bytes, checksum: 66e71c371cc565284e70f40736c94386 (MD5) DissertacaoCD_LuizFelipeLiborio.pdf: 3777707 bytes, checksum: a7937e9fc359d5b560c37ea3abf5e74c (MD5) / Made available in DSpace on 2016-08-22T18:54:13Z (GMT). No. of bitstreams: 2 license_rdf: 1232 bytes, checksum: 66e71c371cc565284e70f40736c94386 (MD5) DissertacaoCD_LuizFelipeLiborio.pdf: 3777707 bytes, checksum: a7937e9fc359d5b560c37ea3abf5e74c (MD5) Previous issue date: 2014-08-27 / Capes / Atualmente, é cada vez mais exigido como critério competitivo a aderência do processo de software de uma organização a um modelo ou norma de qualidade. Alcançar altos níveis de aderência é uma tarefa complexa para as organizações, pois a atividade de avaliar essa aderência exige cuidado ao ser realizada. Com isso, busca-se cada vez mais o apoio de ferramentas de avaliação computadorizadas, cujo objetivo é agilizar e tornar mais confiável o processo de avaliação dos processos de software. Ainda assim, um dos fatores que tendem a dificultar a utilização e propagação do uso de ferramentas é a atualização das normas de qualidade e dos métodos de avaliação associados, que visam se adequar constantemente às boas práticas do mercado. Ora, se uma ferramenta é criada baseada na norma de qualidade e no seu método de avaliação, qualquer mudança em algum dos dois componentes significa mudanças na ferramenta, gerando custos de manutenção e de distribuição da nova versão da ferramenta. Este trabalho busca criar uma engine de geração que seja sensível a esses mudanças, sem que seu processo de desenvolvimento necessite ser executado novamente. Para isto, utiliza o processo MDD (Model-Driven Development) para atender a esses requisitos, dado que o seu objetivo é transformar elementos mais abstratos em elementos mais concretos utilizando uma sucessão de transformações, através de modelos. O objetivo deste trabalho é, então, aplicar as regras de MDD aos modelos de qualidade e métodos de avaliação. Isto torna possível a criação de ferramentas para avaliação baseada nos modelos necessários à avaliação. Ou seja, através de um modelo de processo e de um metamodelo da norma de qualidade, além do método de avaliação, é possível criar de forma automática via MDD uma ferramenta que possa ser atualizada sob demanda, sem que seu código-fonte precise ser alterado manualmente. / Nowadays, it is increasingly required as a competitive criteria the adherence of an organization to a software model or quality standard process. Achieving high levels of adherence is a complex task for organizations, because the activity of evaluating this adherence requires care to be performed. Thus, it is increasing the nedd of computerized evaluation tools whose goal is to speed up and make more reliable the process of evaluating software processes.Still, one of the factors that tend to hamper the use and propagation of tool use is the change of the quality standards and assessment methods, which aim to adapt constantly to best practices in the market. So, if a tool is created based on quality standard and its evaluation method, any change in any of the two components mean changes in the tool, generating costs of maintaining and distributing the new version of the tool. This work seeks to develop a tool that is sensitive to that changes, but without the need of performing its development process again. For that, it uses MDD (Model- Driven Development) processes meet these requirements, since their goal is to transform more abstract elements into more concrete elements using a succession of transformations, through models. So, the goal of this work is to apply the rules of MDD to quality models and assessment methods. It makes possible to create tools based on models. In other words, through a process model and a metamodel quality standard, besides the evaluation method, it is possible to create automatically via MDD a tool that can be updated on demand, without its source code need to be changed manually.

The presented doctoral thesis “Fly ash impact in forest ecosystems in Northeastern Germany - an assessment and regionalization approach” intends to (a) test if the field assessment of ferrimagnetic susceptibility can be used as cost efficient method to get information on fly ash deposition impacted chemical site properties. (b) develop a regionalization approach to bridge the gap from plot-wise assessed data to spatial management information. The thesis is a follow-up of extensive research activities by the Institute for Soil Science and Site Ecology on industrial deposition in Dübener Heide and Upper Lusatian region which started in the early 1960ies and were intensified from the middle of the 1990ies on. A central topic of these research activities was the assessment of the impact of fly ash deposition on chemical soil properties. A major challenge was to transfer the assessed chemical characteristics from plot to region and to aggregate the measured values to provide an information basis, which can be used for a site potential and risk oriented forest management. This challenge was picked up by the joint research project “ENFORCHANGE” (FKZ 0330634 K, German Federal Ministry of Education and Research). The presented thesis was carried out in the frame of this project during the period 2005 - 2009. The thesis was conceived as cumulative work, which includes ten papers in total. Five articles are published in peer-reviewed journals (ISI listed, 1 paper still in revision), and five are part of books or conference proceedings. • Chapter 1 “Introduction” gives an overview on the motivation, idea and structure of the thesis. • In chapter 2 “Aims and Scope of the presented work” information on the background and frame of the study within the project ENFORCHANGE is given. • Chapter 3 “Background and State of the Art” deals with the history of fly ash deposition in the model region Dübener Heide. • Chapter 4 “Material and Methods” gives information on fly ash and presents the spatial assessment design and the hereon based approaches for up-scaling and correlation of magnetic susceptibility with selected chemical characteristics. • Chapter 5 “Results” presents results of the spatial modeling and linear regression based approach to use ferrimagnetic susceptibility for predicting the contents of selected base cations, selected acid and heavy metal cations and Black Carbon. • Chapter 6 “Discussion and Conclusions” compares the assumptions and findings in the different articles, discusses contradictory findings and open questions and provides a comprehensive evaluation of the outcomes. Final conclusions are drawn and an outlook is given. A key finding of the thesis is that the industrial complex Bitterfeld was the most important source of fly ash deposited in the model region Dübener Heide. The power plant Zschornewitz plays only a minor role contrary to the research hypothesis formulated in ENFORCHANGE. Related to the targets of the thesis, spatial variation of magnetic susceptibility was predicted with high precision by a multiple linear regression model. A slightly differing set of model parameters  according to their explanatory value for three selected depth levels  improved the prediction quality. The selection of the parameters supported understanding the major drivers for magnetic particle deposition, storage, and vertical displacement in the forest soils. Humus layer (depth level 6-10 cm), horizontal distance to Bitterfeld and soil type (Podzol, semi-terrestrial sites) were the most important variables. These variables point to a slowed-down humus dynamic, which causes the accumulation of fly ash in the humus layer. In depth level 11 – 15 cm, variables such as “aspect” gain in importance, which describe the exposure against the major wind direction and thus indicate the probability and of deposition. For the mineral horizon (depth level 21-25 cm), exposition and especially stand properties are most important. The latter gives evidence for the intensity of deposition caused by surface roughness. Therefore, the variables “coniferous” and “mixed” stands were highly relevant for the model. Variable correlations between mass susceptibility and selected base cations, acid cations and heavy metals have been found. When using a linear regression model, a prediction of Ca and Mg and of Mn was possible. The model performance was lower for Fe, Al, Cd and Black Carbon. A possible reason was the use of different plot types: the assessment of magnetic susceptibility and chemical soil properties was well harmonized at the ENFORCHANGE plots considering the sampling material and sampling location. A comparable harmonization could not be achieved at a number of monitoring plots, which were included into the analysis to broaden the data base. Comparing the results from the linear regression model based prediction with the results achieved by multiple regression based spatial modeling lead to the conclusion that the mul¬tiple regression approach is more promising: by using other model parameters such as orographic, climatic or stand parameters together with magnetic susceptibility, the prediction quality of the deposed agents could be improved and small scale variations in nutrient potentials and risks driven by fly ash deposition could be better recognized and made available for forest management decisions. / Die vorgelegte Doktorarbeit “Fly ash impact in forest ecosystems in Northeastern Germany - an assessment and regionalization approach” (Flugascheeinträge in Waldökosysteme in Nordostdeutschland  ein Erfassungs- und Regionalisierungsansatz) verfolgte die Ziele (a) zu testen, ob sich die Erfassung der ferrimagnetischen Suszeptibilität eignet, um kosteneffizient quantitative und / oder qualitative Informationen zu den eingetragenen Flugaschemengen und den in der Folge veränderten bodenchemischen Potenzialen zu erheben (b) zu testen, ob der Indikator „ferrimagnetische Suszeptibilität“ genutzt werden kann, um Informationen über Flugascheeinträge von der punktbezogenen Erfassung auf einen regionalen Maßstab hoch zu skalieren. Grundlage dieser Zielstellungen sind Forschungsarbeiten zu der Frage der langfristigen Wirksamkeit und ökologischen Bedeutung von Industrieexhalationen auf Waldökosysteme, die am Institut bereits in den 1960ziger Jahren begonnen wurden und verstärkt seit Mitte der 1990ziger Jahre fortgeführt wurden. Auf ihrer Basis wurde die Herausforderung eines kostengünstigen und flächenbezogenen Erhebungsansatzes identifiziert und formuliert. Die vorgelegte Arbeit ordnete sich in diese Forschungsarbeiten ein und führte sie im Rahmen des Verbundforschungsvorhabens ENFORCHANGE ((FKZ: 0330634 K, Bundesministerium für Bildung und Forschung) von 2005 - 2009 fort. Die Doktorarbeit ist als kumulative Arbeit angelegt, im Rahmen derer insgesamt 10 Publikationen zusammengefasst wurden. Davon sind 5 in internationalen Journalen bereits publiziert, akzeptiert oder in einem Fall in Begutachtung; 5 weitere Publikationen wurden ergänzend und auf speziellere Themen bezogen in Proceedings oder Buchbeiträgen publiziert. Die Arbeit gliedert sich in 5 Abschnitte: • Kapitel 1 (Einleitung) gibt einen kurzen Überblick zur Motivation und Struktur der Doktorarbeit. • In Kapitel 2 (Ziele und Rahmen der Arbeit) wird der Arbeitsansatz im Rahmen des Verbundforschungsvorhabens ENFORCHANGE vorstellt. • Kapitel 3 umfasst eine Auswertung von Veröffentlichungen zur Geschichte und den ökologischen Auswirkungen der Flugascheeinträge am Beispiel der Modellregion Dübener Heide. • In Kapitel 4 wird der methodische Ansatz der Arbeit vorgestellt, der von einem Vortest zur Eignung der Erfassung der magnetischen Suszeptibilität über die Ableitung eines flächigen Erhebungsansatzes bis hin zur Frage der Modellbildung und Korrelation mit chemischen Kenngrößen reicht. • Kapitel 5 beinhaltet die Ergebnisse der räumlichen Modellbildung und der Korrelation der magnetischen Suszeptibilität mit ausgewählten Basen-, Säure- und Schwermetallkationen sowie mit Schwarzem Kohlenstoff. • Kapitel 6 diskutiert, vergleicht und bewertet die Ergebnisse der den Veröffentlichungen zugrunde liegenden Studien und zieht ein abschließendes Resumé. Ein Schlüsselergebnis der vorgelegten Arbeit belegt, dass entgegen der ursprünglichen Arbeitshypothese des Projektverbundes ENFORCHANGE nicht das mehr als 100 Jahre alte Kraftwerk Zschornewitz die wesentliche Quelle für die Flugascheeinträge in der Modellregion Dübener Heide war, sondern der räumlich entfernter gelegene, aber deutlich größere Industriekomplex Bitterfeld. Bezogen auf die Zielsetzung der vorgelegten Arbeit, konnte mithilfe multipler Regressionsverfahren und auf Basis von Feldaufnahmen der ferrimagnetischen Suszeptibilität in einem regelmäßigen Stichprobenraster ein hoch auflösendes räumliches Modell gebildet werden. Unter Berücksichtigung weiterer Modellparameter, die schrittweise hinsichtlich ihres Erklärungswertes ausgewählt wurden, konnten mikrotopographische und vegetationsbedingte Informationen genutzt werden, um die räumliche Variabilität des magnetischen Signals differenziert darzustellen. Damit ergibt sich eine Planungsgrundlage, die die bisher genutzte, auf Waldschadensansprachen basierende Stratifizierung in Zonen unterschiedlicher Eintragsintensität mit Bezug zur Planungseinheit deutlich detaillierter untersetzt. Der Versuch, auf Flugascheeintragsmengen, respektive -vorräte zu schließen ließ sich hingegen auf Basis der verfügbaren Daten nicht umsetzen. Die Korrelationsbeziehungen der von Volumen- in den Massenbezug umgerechneten Suszeptibilität mit Basen-, Säure und Schwermetallkationen sowie Schwarzem Kohlenstoff fielen heterogen aus. Eine gute Vorhersage auf Basis eines linearen Regressionsmodells konnte für Ca, Mg und Mn getroffen werden, wohingegen die Modellqualität für Fe, Al sowie Cd und Schwarzen Kohlenstoff deutlich schlechter zu beurteilen war. Dies ergab sich zum einen aus der verfügbaren Datenbasis, die keine durchgängige Harmonisierung für die Erhebungen der Suszeptibilität und der chemischen Kennwerte erlaubte. Zum anderen geht diese Erkenntnis mit Ergebnissen aus der Regionalisierung einher, die einen Einbezug weiterer Modellparameter und die Nutzung multipler anstelle linearer Regressionsmodelle nahe legt.

Syftet med denna enkätstudie är att undersöka hur mellanstadielärare upplever användningen av formativa bedömningsmetoder av elevers skrivande, men även att undersöka vilka för- och nackdelar som dessa metoder kan medföra. Lärarnas svar tyder på att det finns en varierad tillämpning av lärarrespons, kamratbedömning och självbedömning där fördelarna med respektive metod kan leda till att lärandet exempelvis blir tydligt och synligt samtidigt som eleverna blir aktiva och får möjlighet till att träna ansvar i sitt lärande. Däremot framförs även nackdelar som visar på formativ bedömning inte alltid har den avsedda effekten. Slutsatsen blir därför att det bör ske en reflektion över betydelsen av de didaktiska frågorna när, hur och varför bedömningsmetoderna tillämpas i undervisningen då lärarnas svar tyder på att eleverna inte alltid förstår bedömningen och därför inte kan använda den för att utveckla sitt skrivande.

No description available.

В работе были проанализированы основные подходы, применяемые для оценки кредитоспособности в сфере малого и среднего бизнеса, даны рекомендации по ее совершенствованию. / The paper analyzes the main approaches used to assess the creditworthiness in the sphere of small and medium-sized businesses, gives recommendations for its improvement.

In South Africa, legislation dictates that employees must be developed resulting in the need to identify candidate's development capability. As such, it is essential to be able to measure the learning potential of candidates. In practice, cognitive ability has been linked to the complex process of job performance. This study aimed to determine whether learning potential can be used as a predictor of job performance. The Learning Potential Computerised Adaptive Test (LPCAT) was utilised to measure learning potential and job performance was measured through a promotion ratio (taking job knowledge and tenure into account). The moderator variables were investigated. A sample of 135 technical employees was drawn. The data was statistical manipulated reporting various significant relationships confirming the internal reliability of the LPCAT and indicating a strong significant relationship between learning potential and job performance. Variables such as learning potential, job grade and tenure are significant when predicting job performance. / Industrial and Organisational Psychology / MA (Industrial and Organisational Psychology)

The purpose of this study was to investigate the challenges affecting teachers’ classroom assessment practices and to explore how these challenges influence effective teaching and learning. The study was qualitative in nature and employed an instrumental case study approach. Semi-structured interviews, observations, and document analyses were used in the investigation. The study utilized the conceptual framework based on the model suggested by Hargreaves, Earl and Schmidt (2002). The model comprised four perspectives – technological, cultural, political and postmodern, with the intention of accounting for teachers’ assessment practices. Teachers from four different schools in North West Province were interviewed and observed. The data collected through the observations and interviews allowed me to map out the different challenges encountered by teachers in their classroom assessment practices. I also analysed documents that teachers used in conducting assessment. Document analysis was used to triangulate the information collected through observations and interviews. Textual data was analysed using content analysis. The teachers’ narratives varied according to their teaching experience and background within the diverse contexts of their particular school environments. Major challenges that emanated from this study were policy interpretation, overcrowding, support, parental involvement, moderation mechanisms (internal and external), assessment planning, implementation and communication as well as lack of resources. In order to address these challenges, teachers relied on cluster meetings, their colleagues and mostly their personal experiences. The study also revealed that teachers’ understanding and practices of classroom assessment are influenced by their social and educational context. / Curriculum and Instructional Studies / D. Ed. (Curriculum Studies)

Performance in Lesotho at senior level (COSC) has been poor. One of the subjects which is performed poorly is Biology 5090. This prompted the researcher to examine factors responsible for the poor performance in Biology 5090. The study also aimed at finding ways in which teachers envisage to improve Biology 5090 performance. Data was collected from three schools through semi-structured interviews and document analysis. The study also adopted the General System Theory (GST) as the theoretical framework. Data was analyzed in line with qualitative content analysis, and from the perspective of the GST. The data analysis scheme was developed and presented (appendix P). The results of the study showed that Biology 5090 performance was affected by factors such as human resources, and the transformation process. The study also suggested ways in which Biology 5090 performance could be improved in Lesotho. These include: in-serving training, team work and others. / Science and Technology Education / M. Ed. (Natural Science Education)