Formal long-term athlete development programmes emerged at the turn of the century and, despite some fierce criticisms, have evolved significantly since their inception. The first generation of athletes to grow up with these systems are now coming of age. The purpose of this thesis was to track a population of adolescent school-level swimmers between the ages of 12 and 18 years over an 8-year period so as to assess their performance progression as they matured under these athlete development programmes. The first study aimed to track the performances of the sub-elite athletes at an annual international school championship and to compare their progression with those of both junior elite and elite-level swimmers. In addition to narrowing the gender gap, the records of the sub-elite swimmers have continued to improve. In contrast, both of these factors remained relatively stable for junior elite and elite-level swimmers over the same period. Swimming affords athletes the possibility of within-sport specialisation. This almost unique aspect of swimming led to the two investigations of the second study. Firstly, the paired stroke combinations preferred by swimmers were determined using Cohen’s Kappa tests in a cross-sectional design. Secondly, the stability in the event selection of each swimmer during their adolescent years was explored longitudinally. Both males (33.9±5.8%) and females (36.9±6.5%) preferred to swim the 50 and 100 m freestyle events together over any other paired stroke combination. The majority of swimmers preferred to specialise in specific stroke techniques over distance specialisms with breaststroke being the only stroke in which swimmers of both sexes chose to specialise early. Most notable was that females specialised earlier than males. Studies three (males, n = 446) and four (females, n = 514) utilised mixed linear modelling to determine the quadratic functions of the performance progressions of adolescent swimmers (between the ages of 12 and 19 y) in seven individual competition events. Males progressed at more than twice the rate of females (3.5 and 1.7% per year, respectively) in all strokes over this age range. This was likely due to the fact that females reach puberty before males. Thresholds of peak performance occurred between the ages of 18.5±0.1 y (50 m freestyle and the 200 m individual medley) and 19.8±0.1 y (100 m butterfly) for males, but between the wider range of 16.8±0.2 y (200 m individual medley) and 20.6±0.1 y (100 m butterfly) for females. Using an independent sample of Dutch Junior national swimmers (n = 13), the fifth and final study aimed to evaluate the efficacy of the models developed in studies three and four as both target setting and talent identification tools. This was achieved through a mixed-methods approach where quantitative and qualitative data confirmed the applicability of the models for adolescent swimmers of any skill level. This thesis demonstrates that sub-elite swimmers have probably benefitted from first generation athlete development models. Longitudinal modelling of their data provides a valuable platform from which all adolescent swimmers can be compared and used to inform the next generation of bespoke swimming-specific youth development programmes.

This dissertation develops influence diagnostics for crossover models. Mixed linear models and generalised mixed linear models are utilised to investigate continuous and count data from crossover studies, respectively. For both types of models, changes in the maximum likelihood estimates of parameters, particularly in the estimated treatment effect, due to minor perturbations of the observed data, are assessed. The novelty of this dissertation lies in the analytical derivation of influence diagnostics using decompositions of the perturbed mixed models. Consequently, the suggested influence diagnostics, referred to as the delta-beta and variance-ratio influences, provide new findings about how the constructed residuals affect the estimation in terms of different parameters of interest. The delta-beta and variance-ratio influence in three different crossover models are studied in Chapters 5-6, respectively. Chapter 5 analyses the influence of subjects in a two-period continuous crossover model. Possible problems with observation-level perturbations in crossover models are discussed. Chapter 6 extends the approach to higher-order crossover models. Furthermore, not only the individual delta-beta and variance-ratio influences of a subject are derived, but also the joint influences of two subjects from different sequences. Chapters 5-6 show that the delta-beta and variance-ratio influences of a particular parameter are decided by the special linear combination of the constructed residuals. In Chapter 7, explicit delta-beta influence on the estimated treatment effect in the two-period count crossover model is derived. The influence is related to the Pearson residuals of the subject. Graphical tools are developed to visualise information of influence concerning crossover models for both continuous and count data. Illustrative examples are provided in each chapter.

Indiana University-Purdue University Indianapolis (IUPUI) / An analysis of longitudinal data collected about beliefs regarding colorectal cancer (CRC) screenings at three-time points was analyzed to determine whether the beliefs improved from either the Web-Based, Phone-Based, or Web + Phone interventions compared to Usual Care. A mixed linear model adjusting for baseline and controlling for covariates was used to determine the effects of the intervention; Web-Based intervention was the most efficacious in improving beliefs, and phone intervention was also efficacious for several beliefs, compared to usual care.

No description available.

No description available.

Made available in DSpace on 2014-07-29T14:52:04Z (GMT). No. of bitstreams: 1 TESE LUICE gomes.pdf: 3188724 bytes, checksum: c4156187d61efe808b3d021846f674dd (MD5) Previous issue date: 2010-08-31 / The plant genetic resources stored ex situ are considered as a genetic repository, and are raw material for the development of the world agriculture. In rice, despite its high genetic variability, the lack of information of accessions to compose a databank prevents its use to help the choice of genitors for the breeding programs. The Embrapa Rice Core Collection (ERiCC) was developed from 10,000 accessions from Embrapa GeneBank, and it was set up by 550 accessions, divided in three subsets: 1) 94 lines and cultivars from Brazil (LCB); 2) 148 lines and cultivars from abroad (LCI); and 3) 308 traditional varieties (VT), obtained from germplasm collection expeditions in Brazil. This work aimed: 1) to evaluate the extension of genetic variability of 550 accessions from ERiCC by means of agronomic traits characterization using mixed models and multivariate statistics; 2) to perform a comparative analysis of the genetic divergence considering the agronomical and SSR markers characterizations; and 3) to identify the genotypes with higher genetic diversity and with the best agronomic performances, aiming to promote the most efficient use of such germplasm in breeding programs. The agronomic characterization of 550 accessions was performed in nine field experiments, evaluating 18 phenological-agronomic traits. The data were analyzed using the mixed linear and AMMI models. There was wide variation range of genotypical values for most evaluated traits. In different environments, it was observed VT accessions among the high-yielding materials, demonstrating the potential of this group of germplasm, particularly important due to its high genetic variability, to contribute to the development of cultivars regionally adapted. The AMMI approach allowed a good discrimination of ERiCC rice genotypes in relation to the adaptive performance, identifying the accessions CA880078, CA990001, CA870071 (subset VT), and CNA0009113 (LCI) as having good yield and broad adaptation to distinct environments. The comparative analysis of genetic diversity between agronomic and molecular data was performed using the 242 lines and cultivars accessions from ERiCC, which were characterized by 86 fluorescent SSR markers, and five agronomic traits with genotypic values predicted (values without from the effects of interaction genotypes x environment, from a joint analysis of nine experiments. The genetic divergence among accessions was estimated by the average Euclidian distance for phenotypical data, and by the Rogers modified by Wright (RW) genetic distance. The datasets were jointly analyzed by descriptive and multivariate statistics, using correlation analyses from hierarchical grouping of Ward and UPGMA methods. The phenotypical and molecular data showed a broad distribution of dissimilarity indexes, despite they showed different patterns of variation between them. Low molecular distances were associated to low phenotypical distances, however to high molecular distances, occurred a high broad range of phenotypical variation. The correlation between genetical and phenotypical dissimilarities was significant for both lowland and upland accessions, despite with different values (r=0.156 and r=0.409, respectively). Due to the low relation between phenotypical and molecular data, the analysis of genotypes to be used in breeding programs must include both evaluations to a better accession characterization. Considering the high yielding accessions, the higher molecular distances were identified among the accessions from lowland system of cultivation, among which BR IRGA 413 and CNA0005014, BR IRGA 413 and CNA0005853, and CNA0004552 and CNA0005014. Considering the upland accessions, maximum genetic distances were identified in CNA0000482 and CNA0006422, CNA0001006 and CNA0006422, and CNA0001006 and CNA0003490. The molecular analysis was able to identify accessions with reduced genetic relationship, that if used as genitors, will result in a progeny with a high probability to find new allelic combinations. On the other hand, the phenotypical characterization is important to identify accessions not just genetically divergent, but with superior agronomic trait performances for breeding programs. The results of this work will permit to increase the activities related to the characterization of accessions from rice Genebank, giving support of breeding programs to choose the best accessions to obtain new cultivars, with favorable traits, and broad genetic basis. In addition, a continuous program of phenotypical and molecular characterization of germplasm will be able to identify accessions to increase the genetic variability of ERiCC. / Os recursos genéticos vegetais armazenados ex situ são considerados reservatórios de genes e funcionam como matéria-prima para o desenvolvimento da agricultura mundial. Na cultura do arroz, apesar da extensa variabilidade genética existente, a deficiência de informações que integrem dados que possam efetivamente auxiliar na escolha de genótipos importantes para os programas de melhoramento constitui o principal fator que limita a utilização mais ampla dos acessos armazenados nos bancos de germoplasma. A Coleção Nuclear de Arroz da Embrapa (CNAE) representa a variabilidade genética de mais de 10 mil acessos constituintes do Banco Ativo de Germoplasma (BAG) da Embrapa Arroz e Feijão, e é composta por 550 acessos subdivididos em três estratos: 1) 94 Linhagens e Cultivares Brasileiras (LCB), provenientes de programas de melhoramento de instituições brasileiras; 2) 148 Linhagens e Cultivares Introduzidas (LCI), provenientes de programas de melhoramento de outros países; e 3) 308 Variedades Tradicionais (VT), que reúne acessos obtidos por expedições de coleta de germoplasma realizadas em vários estados do Brasil. Este trabalho teve como principais objetivos: 1) avaliar a extensão da variabilidade genética dos 550 acessos pertencentes à CNAE por meio da caracterização agronômica via metodologias de modelos mistos e estatísticas multivariadas; 2) realizar a análise comparativa da divergência genética entre acessos, determinada pela avaliação de caracteres agronômicos e marcadores moleculares SSR; e 3) identificar os genótipos com maior diversidade genética e com melhores atributos agronômicos, a fim de indicar uma melhor utilização destes recursos genéticos em programas de melhoramento. Na caracterização agronômica foram avaliados 550 acessos em experimentos conduzidos em nove locais no Brasil, envolvendo um total de 18 caracteres fenológico-agronômicos. Os dados foram analisados empregando-se a abordagem de modelos lineares mistos e modelo AMMI de análise. Verificou-se grande amplitude de variação dos valores genotípicos para a maioria dos caracteres avaliados. Nos diferentes ambientes, houve ocorrência de genótipos do estrato VT entre os mais produtivos, o que demonstra o potencial deste grupo de germoplasma, particularmente importante por sua grande variabilidade genética, em contribuir para o desenvolvimento de cultivares regionalmente adaptadas. A abordagem AMMI permitiu uma boa discriminação dos genótipos de arroz da CNAE quanto ao seu comportamento adaptativo, identificando os acessos CA880078, CA990001, CA870071 (do estrato VT), e CNA0009113 (LCI) com estabilidade, produtividade satisfatória e ampla adaptação à diferentes ambientes. Para a análise comparativa da diversidade genética entre dados agronômicos e moleculares foram considerados 242 acessos da CNAE, os quais foram caracterizados utilizando-se 86 marcadores SSR fluorescentes, sendo que para os dados agronômicos, foram realizadas análises conjuntas dos experimentos e considerados os valores genotípicos preditos de cinco caracteres (valores livres dos efeitos de interação genótipos x ambientes). A divergência genética entre os acessos foi estimada pelo procedimento de distância Euclidiana média para os dados fenotípicos, e por meio da distância de Rogers modificada por Wright (RW) para os dados moleculares, analisando-se os conjuntos de dados por meio de estatísticas descritivas e multivariadas, empregando-se análises de correlação entre matrizes de dissimilaridade e análises de agrupamento hierárquico de Ward e UPGMA. Os dados fenotípicos e moleculares apresentaram uma ampla distribuição dos índices de dissimilaridade, embora tenham apresentado diferentes padrões dessa variação. Baixas distâncias moleculares estiveram associadas a baixas distâncias baseada nos valores genotípicos, no entanto para elevadas distâncias moleculares houve ocorrência de ampla escala de variação fenotípica. A correlação entre as dissimilaridades genéticas e valores genotípicos foi significativa tanto no conjunto de acessos irrigados quanto no de sequeiro, porém, com diferentes magnitudes (r=0,156 e r=0,409, respectivamente). Devido esta baixa relação entre os dados fenotípicos e moleculares, o estudo de genótipos para fins de uso no melhoramento genético deve incluir ambas avaliações para a melhor caracterização dos acessos. Entre os materiais mais produtivos, as maiores distâncias moleculares foram identificadas entre os genótipos do sistema de cultivo irrigado, dentre eles BR IRGA 413 e CNA0005014, BR IRGA 413 e CNA0005853, e CNA0004552 e CNA0005014. Entre os materiais de sequeiro, máximas distâncias genéticas foram identificadas entre os acessos CNA0000482 e CNA0006422, CNA0001006 e CNA0006422, e CNA0001006 e CNA0003490. A análise molecular permitiu que fossem identificados genótipos de vínculo genético reduzido, que quando utilizados como parentais em cruzamentos, possibilitarão que as progênies obtidas apresentem maiores chances de combinações alélicas inéditas. Por sua vez, a caracterização fenotípica tem papel fundamental na identificação de materiais que além de divergentes, apresentem desempenho agronômico superior para os programas de melhoramento. Os resultados deste trabalho permitirão aumentar eficazmente as atividades relacionadas à caracterização de acessos do Banco Ativo de Germoplasma de arroz, subsidiando os programas de melhoramento na escolha de genótipos a serem utilizados para a obtenção de novas cultivares, com características favoráveis, de ampla base genética. Em adição, um programa contínuo de caracterização fenotípica e molecular de germoplasma permitirá ainda a escolha de acessos para a ampliação da variabilidade genética da CNAE.

當線性模型中包含隨機效果項時，若將之視為固定效果或直接忽略，往往會造成嚴重的推測偏差，故應以混合線性模型為架構。若模式中只包含一個隨機效果項，則模式中有兩個變異數成份，若包含 個隨機效果項，則模式中有 個變異數成份。本論文主要在介紹至少兩個變異數成份時固定效果及隨機效果線性組合的最佳線性不偏推測量（BLUP），及其推測區間之推導與建立。然而BLUP實為變異數比率的函數，若變異數比率未知，而以最大概似法（Maximum Likelihood Method）或殘差最大概似法（Residual Maximum Likelihood Method）估計出變異數比率，再代入BLUP中，則得到的是經驗最佳線性不偏推測量（EBLUP）。至於推測區間則與EBLUP的均方誤有關，本論文先介紹如何求算其漸近不偏估計量，再介紹EBLUP之推測誤差除以 後，其自由度的估算方法，據以建構推測區間。 / When random effects are contained in the model, if they are treated as fixed effects or ignore, then it may result in serious prediction bias. Instead, mixed linear model is to be considered. If there is one source of random effects, then the model has two variance components, while it has variance components, if the model contains random effects. This study primarily presents the derivation of the best linear unbiased predictor (BLUP) of a linear combination of the fixed and random effects, and then the conduction of the prediction interval when the model contains at least two variance components. However, BLUP is a function of variance ratios. If the variance ratios are unknown, we can replace them by their maximum likelihood estimates or residual maximum likelihood estimates, then we can get empirical best linear unbiased predictor (EBLUP). Because prediction interval is relating to the mean squared error (MSE) of EBLUP, so the study first introduces how to get its approximate unbiased estimator, m_a , then introduces how to evaluate the degrees of freedom of the ratio of the prediction error for the EBLUP and m_a ^1/2 , in order to use both of them to establish the prediction interval.