No description available.

The purpose of this study was to estimate the level of statistical power demonstrated in recent dissertations in the field of educational leadership. Power tables provided in Cohen's (1988) Statistical Power Analysis for the Behavioral Sciences were used to determine the power of the statistical tests conducted in dissertations selected from five universities in Tennessee. The meta-analytic approach was used to summarize and synthesize the findings. The population of this study consisted of all dissertations successfully defended by doctoral students majoring in educational leadership/administration at East Tennessee State University, the University of Tennessee at Knoxville, Tennessee State University, the University of Memphis, and Vanderbilt University from January 1, 1996 through December 31, 1998. Dissertations were included if statistical significance testing was used, if the reported tests were referenced in associated power tables from Cohen's (1988) Statistical Power Analysis for the Behavioral Sciences, and if sample sizes were reported in the study. Eighty out of 221 reviewed dissertations were analyzed and statistical power was calculated for each of the 2629 significance tests. The mean statistical power level was calculated for each dissertation. The mean power was .34 to detect small effects, .79 to detect medium effects, and .94 to detect large effects with the dissertation as the unit of analysis. The mean power level across all significance tests was .29 to detect small effects, .75 to detect medium effects, and .93 to detect large effects. These results demonstrated the highest statistical power levels for detecting large and medium effects. The statistical power estimates were quite low when a small effect size was assumed. Researchers had a very low probability of finding true significant differences when looking for small effects. Though the degree of statistical power demonstrated in analyzed dissertations was satisfactory for large and medium effect sizes, neither power level nor Type II error was mentioned in any of the 80 dissertations that were analyzed. Therefore, it is hard to determine whether these dissertations were undertaken with consideration of Type II error or the level of statistical power. The mean sample size used for the 2,629 significance tests was 2.5 times the mean optimal sample size, although most significance tests used samples that were much smaller than optimal sample size. It is recommended that doctoral students in educational leadership receive additional training on the importance of statistical power and the process for estimating appropriate sample size.

No description available.

This simulation study focused on determining the effect of unequal sample sizes on statistical power of SIBTEST and Mantel-Haenszel procedures for detection of DIF of moderate and large magnitudes. Item parameters were estimated by, and generated with the 2PLM using WinGen2 (Han, 2006). MULTISIM was used to simulate ability estimates and to generate response data that were analyzed by SIBTEST. The SIBTEST procedure with regression correction was used to calculate the DIF statistics, namely the DIF effect size and the statistical significance of the bias. The older SIBTEST was used to calculate the DIF statistics for the M-H procedure. SAS provided the environment in which the ability parameters were simulated; response data generated and DIF analyses conducted. Test items were observed to determine if a priori manipulated items demonstrated DIF. The study results indicated that with unequal samples in any ratio, M-H had better Type I error rate control than SIBTEST. The results also indicated that not only the ratios, but also the sample size and the magnitude of DIF influenced the behavior of SIBTEST and M-H with regard to their error rate behavior. With small samples and moderate DIF magnitude, Type II errors were committed by both M-H and SIBTEST when the reference to focal group sample size ratio was 1:.10 due to low observed statistical power and inflated Type I error rates. / Ph. D.

This dissertation illustrates how to detect the rater centrality effect in a simulation study that approximates data collected in large scale performance assessment settings. It addresses three research questions that: (1) which of several centrality-detection indices are most sensitive to the difference between effect raters and non-effect raters; (2) how accurate (and inaccurate), in terms of Type I error rate and statistical power, each centrality-detection index is in flagging effect raters; and (3) how the features of the data collection design (i.e., the independent variables including the level of centrality strength, the double-scoring rate, and the number of raters and ratees) influence the accuracy of rater classifications by these centrality-detection indices. The results reveal that the measure-residual correlation, the expected-residual correlation, and the standardized deviation of assigned scores perform better than the point-measure correlation. The mean-square fit statistics, traditionally viewed as potential indicators of rater centrality, perform poorly in terms of differentiating central raters from normal raters. Along with the rater slope index, the mean-square fit statistics did not appear to be sensitive to the rater centrality effect. All of these indices provided reasonable protection against Type I errors when all responses were double scored, and that higher statistical power was achieved when responses were 100% double scored in comparison to only 10% being double scored. With a consideration on balancing both Type I error and statistical power, I recommend the measure-residual correlation and the expected-residual correlation for detecting the centrality effect. I suggest using the point-measure correlation only when responses are 100% double scored. The four parameters evaluated in the experimental simulations had different impact on the accuracy of rater classification. The results show that improving the classification accuracy for non-effect raters may come at a cost of reducing the classification accuracy for effect raters. Some simple guidelines for the expected impact of classification accuracy when a higher-order interaction exists summarized from the analyses offer a glimpse of the "pros" and "cons" in adjusting the magnitude of the parameters when we evaluate the impact of the four experimental parameters on the outcomes of rater classification. / Ph. D.

O fenótipo de caracteres complexos é o produto final da inter-relação entre genes, vias ontogenéticas e efeitos ambientais. A variação desses fatores não apenas influencia o fenótipo final, mas também como caracteres covariam e evoluem de forma integrada. A seleção natural pode influenciar a integração entre caracteres, levando a mudança de padrões de correlação ao longo do tempo. Assim, uma visão integrativa e dinâmica de fenótipos complexos é essencial para a compreensão da história evolutiva destas estruturas. Na tese atual investiguei a integração morfológica de caracteres cranianos em Caniformes sob duas perspectivas. Em uma primeira abordagem investiguei o padrão de integração morfológica das espécies a partir da comparação das variâncias e covariâncias dos caracteres. Os resultados evidenciam dois principais pontos. O primeiro é que houve uma considerável estabilidade na covariância entre caracteres ao longo de toda a história evolutiva de Carnivora, sugerindo a manutenção dos padrões de desenvolvimento no grupo como um todo. O segundo ponto é que, apesar desta estabilidade, espécies da família Canidae apresentam modificações em sua integração morfológica que os tornam mais similares entre si e mais dissimilares com os demais Carnivora. Essas diferenças estão relacionadas principalmente com caracteres da região facial, que apresentaram maior flexibilidade evolutiva, maiores correlações entre caracteres, e contêm uma maior proporção da variância em Canidae que nos demais Carnivora. Em uma segunda abordagem investiguei as propriedades estatísticas de dois testes baseados na teoria de genética quantitativa: o teste de regressão de autovalores e o teste de correlação de componentes principais (PCs). Estes testes avaliam a proporcionalidade entre padrões de covariância genética e entre-espécies como forma de testar a hipótese nula de deriva genética. Os resultados mostram que o uso de contrastes filogenéticos independentes (PIC) reduz erros do tipo I inflados, principalmente no caso do teste de correlação. Quando PIC são utilizados, o teste de correlação apresenta taxas de erro tipo I nominais para todos os números de espécies. Entretanto, a flutuação do número efetivo populacional (Ne) infla o erro tipo I deste testes. O teste de regressão, apesar de apresentar erro do tipo I inadequado para número de espécies baixo, é robusto a flutuações de Ne. A redução do número de PCs reduz o erro do tipo I a valores nominais à custa de uma redução no poder do teste. O poder de ambos os testes é similar nos diversos cenários avaliados, com uma leve tendência de maior poder para o teste de correlação em números amostrais mais baixos. Adicionalmente, as famílias de Caniformes foram utilizadas como estudo de caso para ambos testes. Os testes foram realizados com métodos paramétricos e não-paramétricos (simulações) e com e sem PIC. Houve rejeição de deriva para quase todas as famílias, com exceção de Mephitidae e Ursidae. Os testes de regressão baseados em simulações se mostraram consistentes com e sem o uso de PIC, apresentando intervalos de confiança menores que os testes paramétricos. Os resultados da presente tese abrem diversas possibilidades de investigação futura, tanto do ponto de vista empírico (em relação a modificações de Canidae e dos processos evolutivos deste grupo e de Ursidae e Mephitidae), assim como metodológicos (aprofundamento das investigações sobre as propriedades dos métodos para investigações macroevolutivas baseados em genética quantitativa) / The phenotype of complex characters is the end-product of the interrelations between genes, ontogenetic pathways and environmental effects. The variation in these factors influences not only the final phenotype, but also how characters covary and evolve in an integrated way. Natural selection can influence the interaction among characters, leading to changes in the patterns of integration. Therefore, a integrative and dynamic view of complex phenotypes is essential to the understanding of the evolutionary history of such structures. In the present thesis I investigated the morphological integration of cranial characters in Caniform species in two perspectives. In the first approach I investigated the pattern of morphological integration of the species through the comparison of character variances and covariances. The results of this investigation highlighted two points. The first is that there is considerable stability in the covariance among characters along the evolutionary history of Carnivora, suggesting the maintenance of ontogenetic pathways in the group. The second is that, despite this stability, Canidae species show changes in their morphological integration that make them more similar among each other and more different from the rest of Carnivora. These changes are related mainly to characters from the facial region, which showed a greater evolutionary flexibility, greater correlation among characters, and concentrate a greater proportion of the variance in Canidae than in the rest of Carnivora. In a second approach I evaluated the statistical properties of tests based on quantitative genetics theory: the test of regression of eigenvalues and the test of correlation of principal components (PCs). These tests investigate the proportionality between patterns of genetic and between-species covariance as a way to test the null hypothesis of genetic drift. The results show that the use of phylogenetic independent contrasts (PIC) reduces the inflated type I error, especially in the case of the correlation test. When PIC are employed, the correlation test shows nominal type I error rates for all species sample sizes. However, the oscillation of the effective population size (Ne) inflates type I error rates of these tests. The regression test, despite showing inadequate type I error rates at small species sample sizes, is robust to the oscillation of Ne. The reduction of the number of PCs reduces type I error rates to nominal values at the expense of statistical power. The power of both tests is similar under different scenarios evaluated, with a slight tendency of the correlation test to perform better at small number of species. Additionally, the Caniform families were used as case studies for both tests. Tests were performed using parametric and non-parametric (simulations) techniques, with and without PIC. The drift hypothesis was rejected for almost all families, with the exception of Mephitidae and Ursidae. The regression tests based on simulations were consistent with and without the use of PIC, showing narrower confidence intervals than the ones for parametric tests. The results of the present thesis open a wide range of future investigation opportunities, both from the empirical (relative to the differences in Canidae patterns of morphological integration or the evolutionary processes underlying Ursidae and Mephitidae diversification) and methodological (further investigations of the properties of the quantitative genetics-based tests for macroevolution) points of view

O fenótipo de caracteres complexos é o produto final da inter-relação entre genes, vias ontogenéticas e efeitos ambientais. A variação desses fatores não apenas influencia o fenótipo final, mas também como caracteres covariam e evoluem de forma integrada. A seleção natural pode influenciar a integração entre caracteres, levando a mudança de padrões de correlação ao longo do tempo. Assim, uma visão integrativa e dinâmica de fenótipos complexos é essencial para a compreensão da história evolutiva destas estruturas. Na tese atual investiguei a integração morfológica de caracteres cranianos em Caniformes sob duas perspectivas. Em uma primeira abordagem investiguei o padrão de integração morfológica das espécies a partir da comparação das variâncias e covariâncias dos caracteres. Os resultados evidenciam dois principais pontos. O primeiro é que houve uma considerável estabilidade na covariância entre caracteres ao longo de toda a história evolutiva de Carnivora, sugerindo a manutenção dos padrões de desenvolvimento no grupo como um todo. O segundo ponto é que, apesar desta estabilidade, espécies da família Canidae apresentam modificações em sua integração morfológica que os tornam mais similares entre si e mais dissimilares com os demais Carnivora. Essas diferenças estão relacionadas principalmente com caracteres da região facial, que apresentaram maior flexibilidade evolutiva, maiores correlações entre caracteres, e contêm uma maior proporção da variância em Canidae que nos demais Carnivora. Em uma segunda abordagem investiguei as propriedades estatísticas de dois testes baseados na teoria de genética quantitativa: o teste de regressão de autovalores e o teste de correlação de componentes principais (PCs). Estes testes avaliam a proporcionalidade entre padrões de covariância genética e entre-espécies como forma de testar a hipótese nula de deriva genética. Os resultados mostram que o uso de contrastes filogenéticos independentes (PIC) reduz erros do tipo I inflados, principalmente no caso do teste de correlação. Quando PIC são utilizados, o teste de correlação apresenta taxas de erro tipo I nominais para todos os números de espécies. Entretanto, a flutuação do número efetivo populacional (Ne) infla o erro tipo I deste testes. O teste de regressão, apesar de apresentar erro do tipo I inadequado para número de espécies baixo, é robusto a flutuações de Ne. A redução do número de PCs reduz o erro do tipo I a valores nominais à custa de uma redução no poder do teste. O poder de ambos os testes é similar nos diversos cenários avaliados, com uma leve tendência de maior poder para o teste de correlação em números amostrais mais baixos. Adicionalmente, as famílias de Caniformes foram utilizadas como estudo de caso para ambos testes. Os testes foram realizados com métodos paramétricos e não-paramétricos (simulações) e com e sem PIC. Houve rejeição de deriva para quase todas as famílias, com exceção de Mephitidae e Ursidae. Os testes de regressão baseados em simulações se mostraram consistentes com e sem o uso de PIC, apresentando intervalos de confiança menores que os testes paramétricos. Os resultados da presente tese abrem diversas possibilidades de investigação futura, tanto do ponto de vista empírico (em relação a modificações de Canidae e dos processos evolutivos deste grupo e de Ursidae e Mephitidae), assim como metodológicos (aprofundamento das investigações sobre as propriedades dos métodos para investigações macroevolutivas baseados em genética quantitativa) / The phenotype of complex characters is the end-product of the interrelations between genes, ontogenetic pathways and environmental effects. The variation in these factors influences not only the final phenotype, but also how characters covary and evolve in an integrated way. Natural selection can influence the interaction among characters, leading to changes in the patterns of integration. Therefore, a integrative and dynamic view of complex phenotypes is essential to the understanding of the evolutionary history of such structures. In the present thesis I investigated the morphological integration of cranial characters in Caniform species in two perspectives. In the first approach I investigated the pattern of morphological integration of the species through the comparison of character variances and covariances. The results of this investigation highlighted two points. The first is that there is considerable stability in the covariance among characters along the evolutionary history of Carnivora, suggesting the maintenance of ontogenetic pathways in the group. The second is that, despite this stability, Canidae species show changes in their morphological integration that make them more similar among each other and more different from the rest of Carnivora. These changes are related mainly to characters from the facial region, which showed a greater evolutionary flexibility, greater correlation among characters, and concentrate a greater proportion of the variance in Canidae than in the rest of Carnivora. In a second approach I evaluated the statistical properties of tests based on quantitative genetics theory: the test of regression of eigenvalues and the test of correlation of principal components (PCs). These tests investigate the proportionality between patterns of genetic and between-species covariance as a way to test the null hypothesis of genetic drift. The results show that the use of phylogenetic independent contrasts (PIC) reduces the inflated type I error, especially in the case of the correlation test. When PIC are employed, the correlation test shows nominal type I error rates for all species sample sizes. However, the oscillation of the effective population size (Ne) inflates type I error rates of these tests. The regression test, despite showing inadequate type I error rates at small species sample sizes, is robust to the oscillation of Ne. The reduction of the number of PCs reduces type I error rates to nominal values at the expense of statistical power. The power of both tests is similar under different scenarios evaluated, with a slight tendency of the correlation test to perform better at small number of species. Additionally, the Caniform families were used as case studies for both tests. Tests were performed using parametric and non-parametric (simulations) techniques, with and without PIC. The drift hypothesis was rejected for almost all families, with the exception of Mephitidae and Ursidae. The regression tests based on simulations were consistent with and without the use of PIC, showing narrower confidence intervals than the ones for parametric tests. The results of the present thesis open a wide range of future investigation opportunities, both from the empirical (relative to the differences in Canidae patterns of morphological integration or the evolutionary processes underlying Ursidae and Mephitidae diversification) and methodological (further investigations of the properties of the quantitative genetics-based tests for macroevolution) points of view

The objectives of this thesis relate to questions needed to be addressed in the context of genetic monitoring for implementing the Convention on Biological Diversity for the gene level. Genetic monitoring is quantifying temporal changes in population genetic metrics. Specific goals of this thesis include i) synthesizing existing information relevant to genetic monitoring of Swedish species, ii) providing a genetic baseline for the Swedish moose, iii) evaluating the relative performance of nuclear versus organelle genetic markers for detecting population divergence, iv) actually monitoring the genetic composition, structure, level of variation, and effective population size (Ne) and assessing the relation between Ne and the actual number of individuals for an unexploited brown trout population. The concept of conservation genetic monitoring is defined and Swedish priority species for such monitoring are identified; they include highly exploited organisms such as moose, salmonid fishes, Norway spruce, Atlantic cod, and Atlantic herring. Results indicate that the Swedish moose might be more genetically divergent than previously anticipated and appears to be divided into at least three different subpopulations, representing a southern, a central, and a northern population. The relative efficiency of nuclear and organelle markers depends on the relationship between the degree of genetic differentiation at the two types of markers. In turn, this relates to how far the divergence process has progressed. For the monitored brown trout population no indication of systematic change of population structure or allele frequencies was observed over 30 years. Significant genetic drift was found, though, translating into an overall Ne-estimate of ~75. The actual number of adult fish (NC) was assessed as ~600, corresponding to an Ne/NC ratio of 0.13. In spite of the relatively small effective population size monitoring did not reveal loss of genetic variation.

La spectrométrie de masse fait partie des technologies haut débit et offre à ce titre un regard inédit, à une échelle nouvelle, sur les protéines contenues dans divers échantillons biologiques. Les études biomédicales utilisant cette technologie sont de plus en plus nombreuses et visent à détecter de nouveaux biomarqueurs de différents processus biologiques, notamment de processus pathologiques à l'origine de cancers. Cette utilisation comme outil de criblage pose des questions quant à la capacité même des expériences de spectrométrie de masse dans cette détection. La puissance statistique traduit cette capacité et rappelle que les études doivent être calibrées pour offrir des garanties suffisantes de succès. Toutefois, cette exploration de la puissance statistique en spectrométrie de masse n'a pas encore été réalisée. L'objet de cette thèse est précisément l'étude des déterminants de la puissance pour la détection de biomarqueurs en spectrométrie de masse. Une revue de la littérature a été réalisée, reprenant l'ensemble des étapes nécessaires du traitement du signal, afin de bien comprendre les techniques utilisées. Les méthodes statistiques disponibles pour l'analyse du signal ainsi traité sont revues et mises en perspective. Les situations de tests multiples, qui émergent notamment de ces données de spectrométrie de masse, suggèrent une redéfinition de la puissance, détaillée par la suite. La puissance statistique dépend du plan d'expérience. La taille d'échantillon, la répartition entre groupes étudiés et l'effet différentiel ont été investigués, par l'intermédiaire de simulations d'expériences de spectrométrie de masse. On retrouve ainsi les résultats classiques de la puissance, faisant notamment ressortir le besoin crucial d'augmenter la tailles des études pour détecter des biomarqueurs, particulièrement lorsque ceux-ci présentent un faible effet différentiel. Au delà de ces déterminants classiques de la puissance, des déterminants propres à la spectrométrie de masse apparaissent. Une chute importante de puissance est mise en évidence, due à l'erreur de mesure des technologies de spectrométrie de masse. Une synergie péjorative existe de plus entre erreur de mesure et procédure de contrôle du risque de première espèce de type FDR. D'autre part, les méthodes de détection des pics, par leurs imperfections (faux pics et pics manqués), induisent un contrôle suboptimal de ce risque de première espèce, conduisant à une autre chute de puissance. Ce travail de thèse met ainsi en évidence trois niveaux d'intervention possibles pour améliorer la puissance des études : la meilleure calibration des plans d'expérience, la minimisation de l'erreur de mesure et l'amélioration des algorithmes de prétraitement. La technologie même de spectrométrie de masse ne pourra conduire de façon fiable à la détection de nouveaux biomarqueurs qu'au prix d'un travail à ces trois niveaux. / Mass-spectrometry (MS) belongs to the high-throughput technologies and therefore offers an originalperspective on proteins contained in various biological samples, at a new scale. Biomedicalstudies using this technology are increasingly frequent. They aim at detecting new biomarkersof different biological processes, especially pathological processes leading to cancer. This use asa screening tool asks questions regarding the very detection effectiveness of MS experiments.Statistical power is the direct translation of this effectiveness and reminds us that calibratedstudies are required to offer sufficient guarantees of success. However, this exploration of statisticalpower in mass-spectrometry has not been performed yet. The theme of this work is preciselythe study of power determinants for the detection of biomarkers in MS studies.A literature review was performed, summarizing all necessary pretreatment steps of thesignal analysis, in order to understand the utilized techniques. Available statistical methods forthe analysis of this pretreated signal are also reviewed and put into perspective. Multiple testingsettings arising from MS data suggest a power redefinition. This power redefinition is detailed.Statistical power depends on the study design. Sample sizes, group repartition and the differentialeffect were investigated through MS experiment simulations. Classical results of statisticalpower are acknowledged, with an emphasis on the crucial need to increase sample sizes forbiomarker detection, especially when these markers show low differential effects.Beyond these classical power determinants, mass-spectrometry specific determinants appear.An important power drop is experienced when taking into account the high measurement variabilityencountered in mass-spectrometry. A detrimental synergy exists between measurementvariability and type 1 error control procedures (e.g. FDR). Furtheremore, the imperfections ofpeak detection methods (false and missed peaks) induce a sub-optimal control of this type 1error, leading to another power drop.This work shows three possible intervention levels if we want to improve power in MS studies: a better study design, measurement variability minimisation and pretreatment algorithmsimprovements. Only a work at these three levels can guarantee reliable biomarker detections inthese studies.

Effect sizes are the currency of psychological research. They quantify the results of a study to answer the research question and are used to calculate statistical power. The interpretation of effect sizes—when is an effect small, medium, or large?—has been guided by the recommendations Jacob Cohen gave in his pioneering writings starting in 1962: Either compare an effect with the effects found in past research or use certain conventional benchmarks. The present analysis shows that neither of these recommendations is currently applicable. From past publications without pre-registration, 900 effects were randomly drawn and compared with 93 effects from publications with pre-registration, revealing a large difference: Effects from the former (median r = 0.36) were much larger than effects from the latter (median r = 0.16). That is, certain biases, such as publication bias or questionable research practices, have caused a dramatic inflation in published effects, making it difficult to compare an actual effect with the real population effects (as these are unknown). In addition, there were very large differences in the mean effects between psychological sub-disciplines and between different study designs, making it impossible to apply any global benchmarks. Many more pre-registered studies are needed in the future to derive a reliable picture of real population effects.