Shrinkage empirical likelihood estimator in longitudinal analysis with time-dependent covariates-application to modeling the health of Filipino children

The method of generalized estimating equations (GEE) is a popular tool for analysing longitudinal (panel) data. Often, the covariates collected are time-dependent in nature, for example, age, relapse status, monthly income. When using GEE to analyse longitudinal data with time-dependent covariates, crucial assumptions about the covariates are necessary for valid inferences to be drawn. When those assumptions do not hold or cannot be verified, Pepe and Anderson (1994, Communications in Statistics, Simulations and Computation 23, 939–951) advocated using an independence working correlation assumption in the GEE model as a robust approach. However, using GEE with the independence correlation assumption may lead to significant efficiency loss (Fitzmaurice, 1995, Biometrics 51, 309–317). In this article, we propose a method that extracts additional information from the estimating equations that are excluded by the independence assumption. The method always includes the estimating equations under the independence assumption and the contribution from the remaining estimating equations is weighted according to the likelihood of each equation being a consistent estimating equation and the information it carries. We apply the method to a longitudinal study of the health of a group of Filipino children.

Tractable nonparametric bayesian inference in poisson processes with Gaussian process intensities

The inhomogeneous Poisson process is a point process that has varying intensity across its domain (usually time or space). For nonparametric Bayesian modeling, the Gaussian process is a useful way to place a prior distribution on this intensity. The combination of a Poisson process and GP is known as a Gaussian Cox process, or doubly-stochastic Poisson process. Likelihood-based inference in these models requires an intractable integral over an infinite-dimensional random function. In this paper we present the first approach to Gaussian Cox processes in which it is possible to perform inference without introducing approximations or finitedimensional proxy distributions. We call our method the Sigmoidal Gaussian Cox Process, which uses a generative model for Poisson data to enable tractable inference via Markov chain Monte Carlo. We compare our methods to competing methods on synthetic data and apply it to several real-world data sets. Copyright 2009.

Tractable nonparametric Bayesian inference in Poisson processes with Gaussian process intensities

The inhomogeneous Poisson process is a point process that has varying intensity across its domain (usually time or space). For nonparametric Bayesian modeling, the Gaussian process is a useful way to place a prior distribution on this intensity. The combination of a Poisson process and GP is known as a Gaussian Cox process, or doubly-stochastic Poisson process. Likelihood-based inference in these models requires an intractable integral over an infinite-dimensional random function. In this paper we present the first approach to Gaussian Cox processes in which it is possible to perform inference without introducing approximations or finite-dimensional proxy distributions. We call our method the Sigmoidal Gaussian Cox Process, which uses a generative model for Poisson data to enable tractable inference via Markov chain Monte Carlo. We compare our methods to competing methods on synthetic data and apply it to several real-world data sets.

Individual and social determinants of multiple chronic disease behavioural risk factors in Canadian children and adolescents

Contexte: Les facteurs de risque comportementaux, notamment l’inactivité physique, le comportement sédentaire, le tabagisme, la consommation d’alcool et le surpoids sont les principales causes modifiables de maladies chroniques telles que le cancer, les maladies cardiovasculaires et le diabète. Ces facteurs de risque se manifestent également de façon concomitante chez l’individu et entraînent des risques accrus de morbidité et de mortalité. Bien que les facteurs de risque comportementaux aient été largement étudiés, la distribution, les patrons d’agrégation et les déterminants de multiples facteurs de risque comportementaux sont peu connus, surtout chez les enfants et les adolescents. Objectifs: Cette thèse vise 1) à décrire la prévalence et les patrons d’agrégation de multiples facteurs de risque comportementaux des maladies chroniques chez les enfants et adolescents canadiens; 2) à explorer les corrélats individuels, sociaux et scolaires de multiples facteurs de risque comportementaux chez les enfants et adolescents canadiens; et 3) à évaluer, selon le modèle conceptuel de l’étude, l’influence longitudinale d’un ensemble de variables distales (c’est-à-dire des variables situées à une distance intermédiaire des comportements à risque) de type individuel (estime de soi, sentiment de réussite), social (relations sociales, comportements des parents/pairs) et scolaire (engagement collectif à la réussite, compréhension des règles), ainsi que de variables ultimes (c’est-à-dire des variables situées à une distance éloignée des comportements à risque) de type individuel (traits de personnalité, caractéristiques démographiques), social (caractéristiques socio-économiques des parents) et scolaire (type d’école, environnement favorable, climat disciplinaire) sur le taux d’occurrence de multiples facteurs de risque comportementaux chez les enfants et adolescents canadiens. Méthodes: Des données transversales (n = 4724) à partir du cycle 4 (2000-2001) de l’Enquête longitudinale nationale sur les enfants et les jeunes (ELNEJ) ont été utilisées pour décrire la prévalence et les patrons d’agrégation de multiples facteurs de risque comportementaux chez les jeunes canadiens âgés de 10-17 ans. L’agrégation des facteurs de risque a été examinée en utilisant une méthode du ratio de cas observés sur les cas attendus. La régression logistique ordinale a été utilisée pour explorer les corrélats de multiples facteurs de risque comportementaux dans un échantillon transversal (n = 1747) de jeunes canadiens âgés de 10-15 ans du cycle 4 (2000-2001) de l’ELNEJ. Des données prospectives (n = 1135) à partir des cycle 4 (2000-2001), cycle 5 (2002-2003) et cycle 6 (2004-2005) de l’ELNEJ ont été utilisées pour évaluer l’influence longitudinale des variables distales et ultimes (tel que décrit ci-haut dans les objectifs) sur le taux d’occurrence de multiples facteurs de risque comportementaux chez les jeunes canadiens âgés de 10-15 ans; cette analyse a été effectuée à l’aide des modèles de Poisson longitudinaux. Résultats: Soixante-cinq pour cent des jeunes canadiens ont rapporté avoir deux ou plus de facteurs de risque comportementaux, comparativement à seulement 10% des jeunes avec aucun facteur de risque. Les facteurs de risque comportementaux se sont agrégés en de multiples combinaisons. Plus précisément, l’occurrence simultanée des cinq facteurs de risque était 120% plus élevée chez les garçons (ratio observé/attendu (O/E) = 2.20, intervalle de confiance (IC) 95%: 1.31-3.09) et 94% plus élevée chez les filles (ratio O/E = 1.94, IC 95%: 1.24-2.64) qu’attendu. L’âge (rapport de cotes (RC) = 1.95, IC 95%: 1.21-3.13), ayant un parent fumeur (RC = 1.49, IC 95%: 1.09-2.03), ayant rapporté que la majorité/tous de ses pairs consommaient du tabac (RC = 7.31, IC 95%: 4.00-13.35) ou buvaient de l’alcool (RC = 3.77, IC 95%: 2.18-6.53), et vivant dans une famille monoparentale (RC = 1.94, IC 95%: 1.31-2.88) ont été positivement associés aux multiples comportements à risque. Les jeunes ayant une forte estime de soi (RC = 0.92, IC 95%: 0.85-0.99) ainsi que les jeunes dont un des parents avait un niveau d’éducation postsecondaire (RC = 0.58, IC 95%: 0.41-0.82) étaient moins susceptibles d’avoir de multiples facteurs de risque comportementaux. Enfin, les variables de type social distal (tabagisme des parents et des pairs, consommation d’alcool par les pairs) (Log du rapport de vraisemblance (LLR) = 187.86, degrés de liberté = 8, P < 0,001) et individuel distal (estime de soi) (LLR = 76.94, degrés de liberté = 4, P < 0,001) ont significativement influencé le taux d’occurrence de multiples facteurs de risque comportementaux. Les variables de type individuel ultime (âge, sexe, anxiété) et social ultime (niveau d’éducation du parent, revenu du ménage, structure de la famille) ont eu une influence moins prononcée sur le taux de cooccurrence des facteurs de risque comportementaux chez les jeunes. Conclusion: Les résultats suggèrent que les interventions de santé publique devraient principalement cibler les déterminants de type individuel distal (tel que l’estime de soi) ainsi que social distal (tels que le tabagisme des parents et des pairs et la consommation d’alcool par les pairs) pour prévenir et/ou réduire l’occurrence de multiples facteurs de risque comportementaux chez les enfants et les adolescents. Cependant, puisque les variables de type distal (telles que les caractéristiques psychosociales des jeunes et comportements des parents/pairs) peuvent être influencées par des variables de type ultime (telles que les caractéristiques démographiques et socioéconomiques), les programmes et politiques de prévention devraient également viser à améliorer les conditions socioéconomiques des jeunes, particulièrement celles des enfants et des adolescents des familles les plus démunies.

Approximation de la distribution a posteriori d'un modèle Gamma-Poisson hiérarchique à effets mixtes

La méthode que nous présentons pour modéliser des données dites de "comptage" ou données de Poisson est basée sur la procédure nommée Modélisation multi-niveau et interactive de la régression de Poisson (PRIMM) développée par Christiansen et Morris (1997). Dans la méthode PRIMM, la régression de Poisson ne comprend que des effets fixes tandis que notre modèle intègre en plus des effets aléatoires. De même que Christiansen et Morris (1997), le modèle étudié consiste à faire de l'inférence basée sur des approximations analytiques des distributions a posteriori des paramètres, évitant ainsi d'utiliser des méthodes computationnelles comme les méthodes de Monte Carlo par chaînes de Markov (MCMC). Les approximations sont basées sur la méthode de Laplace et la théorie asymptotique liée à l'approximation normale pour les lois a posteriori. L'estimation des paramètres de la régression de Poisson est faite par la maximisation de leur densité a posteriori via l'algorithme de Newton-Raphson. Cette étude détermine également les deux premiers moments a posteriori des paramètres de la loi de Poisson dont la distribution a posteriori de chacun d'eux est approximativement une loi gamma. Des applications sur deux exemples de données ont permis de vérifier que ce modèle peut être considéré dans une certaine mesure comme une généralisation de la méthode PRIMM. En effet, le modèle s'applique aussi bien aux données de Poisson non stratifiées qu'aux données stratifiées; et dans ce dernier cas, il comporte non seulement des effets fixes mais aussi des effets aléatoires liés aux strates. Enfin, le modèle est appliqué aux données relatives à plusieurs types d'effets indésirables observés chez les participants d'un essai clinique impliquant un vaccin quadrivalent contre la rougeole, les oreillons, la rub\'eole et la varicelle. La régression de Poisson comprend l'effet fixe correspondant à la variable traitement/contrôle, ainsi que des effets aléatoires liés aux systèmes biologiques du corps humain auxquels sont attribués les effets indésirables considérés.

Longitudinal familial analysis of blood pressure involving parametric (co)variance functions

Abstract Background For analyzing longitudinal familial data we adopted a log-linear form to incorporate heterogeneity in genetic variance components over the time, and additionally a serial correlation term in the genetic effects at different levels of ages. Due to the availability of multiple measures on the same individual, we permitted environmental correlations that may change across time. Results Systolic blood pressure from family members from the first and second cohort was used in the current analysis. Measures of subjects receiving hypertension treatment were set as censored values and they were corrected. An initial check of the variance and covariance functions proposed for analyzing longitudinal familial data, using empirical semi-variogram plots, indicated that the observed trait dispersion pattern follows the assumptions adopted. Conclusion The corrections for censored phenotypes based on ordinary linear models may be an appropriate simple model to correct the data, ensuring that the original variability in the data was retained. In addition, empirical semi-variogram plots are useful for diagnosis of the (co)variance model adopted.

Interhemispheric cerebral blood flow balance during recovery of motor hand function after ischemic stroke - a longitudinal MRI study using arterial spin labeling perfusion

BACKGROUND Unilateral ischemic stroke disrupts the well balanced interactions within bilateral cortical networks. Restitution of interhemispheric balance is thought to contribute to post-stroke recovery. Longitudinal measurements of cerebral blood flow (CBF) changes might act as surrogate marker for this process. OBJECTIVE To quantify longitudinal CBF changes using arterial spin labeling MRI (ASL) and interhemispheric balance within the cortical sensorimotor network and to assess their relationship with motor hand function recovery. METHODS Longitudinal CBF data were acquired in 23 patients at 3 and 9 months after cortical sensorimotor stroke and in 20 healthy controls using pulsed ASL. Recovery of grip force and manual dexterity was assessed with tasks requiring power and precision grips. Voxel-based analysis was performed to identify areas of significant CBF change. Region-of-interest analyses were used to quantify the interhemispheric balance across nodes of the cortical sensorimotor network. RESULTS Dexterity was more affected, and recovered at a slower pace than grip force. In patients with successful recovery of dexterous hand function, CBF decreased over time in the contralesional supplementary motor area, paralimbic anterior cingulate cortex and superior precuneus, and interhemispheric balance returned to healthy control levels. In contrast, patients with poor recovery presented with sustained hypoperfusion in the sensorimotor cortices encompassing the ischemic tissue, and CBF remained lateralized to the contralesional hemisphere. CONCLUSIONS Sustained perfusion imbalance within the cortical sensorimotor network, as measured with task-unrelated ASL, is associated with poor recovery of dexterous hand function after stroke. CBF at rest might be used to monitor recovery and gain prognostic information.

La coordinación motriz en la Adolescencia y su relación con el IMC, hábitos de práctica y motivación en E.F: Estudio transversal y longitudinal

En las últimas décadas, ha aumentado el interés de la investigación sobre el desarrollo de la coordinación motriz en la adolescencia por ser una etapa sensible, crítica y crucial para la adquisición de hábitos y conductas saludables de vida. Estos estudios han mostrado que la adquisición de unos niveles óptimos de coordinación y competencia motriz van a ser determinantes para el bienestar del adolescente y van a estar relacionados e influidos por otras dimensiones del desarrollo de la persona. Recientes investigaciones han sacado a la luz datos alarmantes sobre el aumento de problemas de coordinación motriz en la población infantil y adolescente (Cantell, Smyth y Ahonen, 1994; Gómez, 2004; Ruiz, Graupera, Gutiérrez y Miyahara, 2003; Sudgen y Chambers, 2005) donde abrocharse los botones de una camisa o correr de forma armónica puede ser todo un mundo lleno de dificultades y consecuencias sobre otras dimensiones del desarrollo (Ramón-Otero y Ruiz, 2015). Estos problemas han sido tratados por investigadores como una “dificultad oculta” (Gómez, Ruiz y Mata, 2006), cuya manifestación está presente en las actividades de la vida cotidiana, en contextos deportivos, en juegos y/o en la clase de Educación Física (Ruiz, 2004). La preocupación por estas dificultades se ha extendido a nivel internacional, creando todo un campo de investigación que estudia el diagnóstico de éstos problemas, conocido bajo las siglas DCD (Developmental Coordination Disorder). El presente estudio se centra en la etapa adolescente, periodo de transición entre la etapa infantil y adulta, caracterizada por numerosos cambios biológicos, cognitivos y socioemocionales (Santrock, 2005), que van a determinar la adaptación con el entorno (Gallahue, Ozmun y Goodway, 2011; Gómez, Ruiz, y Mata, 2006). El propósito principal del estudio es analizar el desarrollo de la coordinación motriz en la etapa adolescente investigando las diferencias de género y de edad en relación con variables psicosociales, los hábitos de práctica y las variables antropométricas. El diseño de la investigación se estructura en dos estudios. El primero de ellos, de carácter transversal, analizó una muestra representativa de 1.966 adolescentes de 1º a 4º de la ESO. El segundo, de naturaleza longitudinal, utilizó un grupo de 89 adolescentes del estudio transversal los cuales fueron estudiados durante 4 años, desde los 12 a los 15 años. Los mismos instrumentos fueron utilizados en ambos estudios: el Test Sportcomp para la evaluación de la coordinación motriz, el test AMPET4 para valorar la motivación de logro para el aprendizaje en Educación Física, el inventario HBSC para conocer los hábitos saludables sobre la práctica de actividad física y, por ultimo, se utilizó un estadiómetro para obtener el peso y la altura y así calcular el índice de masa corporal (IMC). La toma de datos del Estudio Transversal se realizó en 2 cursos académicos (2011/12 - 2012/13), en la cual se requirieron 3 sesiones coincidiendo con la clase de Educación Física. En la primera sesión, se evaluó la coordinación motriz. En la segunda se aplicaron los cuestionarios (AMPET4 y HBSC) y, en la última sesión se midió el peso y la altura en un espacio reservado al estadiómetro. El análisis de datos fue descriptivo y diferencial de cada una de las variables estudiadas: motoras, psicosociales, de hábitos de práctica de actividad física y antropométricas. Asimismo, se llevaron a cabo pruebas de análisis univariante y multivariante, calculando el valor-p y las pruebas de efecto. Respecto al Estudio Longitudinal, la toma de datos se llevó cabo durante 4 años desde el 2011 al 2014. La evaluación de la coordinación motriz se realizó en cada uno de los 4 años. Sin embargo, los 2 cuestionarios y las medidas antropométricas fueron evaluadas en el primer y cuarto año. Los análisis de datos fueron descriptivos y comparativos entre las variables analizadas. En el caso de la coordinación motriz, se realizaron las pruebas de medidas repetidas y, en el caso de las demás variables analizadas, se realizaron Prueba T para muestras relacionadas. Los resultados globales mostraron que el índice motor en el Estudio Transversal fue progresivo en el conjunto de chicos. Sin embargo, en las chicas, el rendimiento se estabiliza a partir de los 13 años. En el caso del Estudio Longitudinal, este índice se estabiliza en los 3 primeros años y a la edad de los 14, es cuando comienzan a acusarse las diferencias de género. En el caso de los hombres el rendimiento mejora y, por el contrario, en las mujeres empeora. En el Estudio Transversal, el análisis de varianza mostraron diferencias en función de la edad [F(7, 1958) = 220.70, p < .001; η2 = .101], del género [F(7, 1958) = 29.76, p < .001; η2 = .044], así como en la interacción entre ambos [F(7, 1958)= 11.90, p < .001; η2 = .018]. Únicamente aparecieron diferencias significativas con la edad en todos los grupos de hombres, excepto entre 14 y 15 años. En el Longitudinal, los contrastes multivariados mostraron que no hubo diferencias sgnificativas en el tiempo [F(3,85) = .05, p = .987, η2= .002] mostrando un nivel de coordinación estable a lo largo de los años, aunque existieron diferencias entre chicos y chicas [F(3,85) = 4.64 p = .005] con un tamaño de efecto destacable (η2 = .141). En cuanto a la motivación de logro para prender en Educación Física, en ambos estudios, los chicos fueron los que obtuvieron puntuaciones más elevadas en todas las dimensiones positivas del test (compromiso de aprendizaje, competencia autopercibida y comparada). Sin embargo, en la dimensión negativa del test, la referida a la ansiedad y al agobio ante el fracaso, fueron las chicas las que puntuaron más alto. En el Estudio Transversal, los resultados mostraron diferencias significativas en todas las dimensiones del AMPET4 en función del nivel de coordinación motriz: compromiso con el aprendizaje [F(2, 1644) = 8.66, p < .001; η2 = .010], competencia autopercibida [F(2, 1644) = 50.94, p < .001; η2 = .048], competencia comparada [F(2, 1644) = 41.56, p < .001, η2 = .020] y ansiedad [F(2, 1644) = 16.67, p < .001, η2 = .058]. En este sentido, los grupos de mejor nivel de coordinación motriz, fueron los que mayor puntuación obtuvieron en las dimensiones positivas y los que menor, en la negativa. En el Estudio Longitudinal, también se encontraron diferencias entre el primer y cuarto año de estudio en todas las dimensiones, excepto en competencia motriz autopercibida. Estas diferencias se tradujeron en una disminución en las 3 variables significativas del primer al cuarto año. Respecto al inventario HBSC, en el Estudio Longitudinal, la prueba T mostró únicamente la existencia de diferencias significativas entre el primer y cuarto año en 2 de los 11 ítems: percepción de la forma física (p = .006) y percepción de la salud (p = .047), los cuales disminuyeron en el intervalo de tiempo del estudio. En el Transversal, las diferencias se observaron en función del género (p < .001) y de la edad (p < .001). Asimismo, se mostraron diferencias significativas en todos los ítems respecto al nivel de coordinación motriz, excepto en 2 de ellos: frecuencia tiempo libre con los amigos fuera del colegio (p = .580) y facilidad para hacer amigos en el centro escolar (p = .098). Por último, en las variables antropométricas, los resultados del Estudio Transversal y Longitudinal coinciden tanto en la estatura como en el peso, apuntando, que en ambos estudios, se produce un aumento progresivo tanto en chicos como en chicas a medida que se avanza en edad. Concretamente en el Transversal, estas diferencias en la edad se encuentran en todos los grupos en ambos géneros, excepto en el conjunto de chicas entre los 14 y los 15 años. Asimismo, ambos estudios coincidieron en que tanto las ganancias en cm y kg, como las puntuaciones medias, fueron mayores en los chicos que en las chicas. Respecto al IMC, los 2 estudios coincidieron en que la evolución es paralela, y tal y como apuntan los resultados del Transversal, no se encontraron diferencias ni en la edad (p = 792) ni en el género (p = 284). No obstante, el Longitudinal apuntó únicamente diferencias significativas entre el primer y cuarto año en el conjunto de los hombres [t(41) = -4.01, p < .001]. Finalmente, y en relación con los niveles de coordinación motriz, hubo diferencias significativas en relación con el IMC (p = .012), mostrando como el grupo de peso normal coincide con puntuaciones óptimas de coordinación motriz. A modo de conclusiones, el presente estudio revela cómo la adquisición de un nivel de coordinación óptimo va a ser fundamental para el desarrollo psicosocial, para el desarrollo de hábitos saludables de práctica y para mantener un IMC dentro de la normalidad para el género y la edad. De esta manera, el desarrollo de la coordinación motriz será un aspecto fundamental para lograr un estado de bienestar físico y mental, y unos hábitos favorables para la práctica de actividad física. ABSTRACT In the past couple of decades, adolescence stage in motor coordination gained significant interest in research especially due to its sensitive and critical importance to achieving a healthy life style. These studies observed how to acquire optimum levels of coordination and motor competence, which proved crucial to the quality of the adolescent stage in addition to being influenced by other dimensions of development for each individual. Recent research shed light to an alarming set of data, which showed increased motor coordination problems in children and adolescents (Cantell, Smyth & Ahonen, 1994; Gómez, 2004; Ruiz, Graupera, Gutierrez & Miyahara, 2003; Sugden & Chambers, 2005). For instance, even to the extent that buttoning a shirt or running in a harmonic form can lead to a whole set of consequences and difficulties on the development stage. Researchers have addressed such problems in various studies such as “dificultad oculta” (Gomez, Ruiz & Mata, 2006), which literally translates as “hidden trouble”. The studies are evidently present in the activities of daily life, sporting contexts, games and/or Physical Education (Ruiz, 2004). Concern about these difficulties spread internationally, creating a whole framework research studying the diagnosis of these problems, known under the acronym DCD (Developmental Coordination Disorder). The study focuses on the adolescent stage, transition period between childhood and adulthood characterized by numerous biological, cognitive and socio-emotional changes (Santrock, 2005), which interestingly determines an individual´s adaptation to the environment (Gallahue, Ozmun & Goodway, 2011; Gomez, Ruiz & Mata, 2006). The main purpose of the study is to analyse the development of motor coordination in the adolescent stage investigating gender differences and age in relation to psychological variables, physical activity habits and anthropometric variables. The research design is structured in two studies. The first (transversal nature), analyses a representative sample of 1,966 adolescents from 1st to 4th of Secondary Education School. The second (longitudinal nature) used a group of 89 teenagers from cross-sectional study, which were studied for four years, from 12 to 15 years. The same instruments were used in both studies, namely; “Sportcomp Test” used to evaluate of motor coordination; “AMPET4 Test” which assesses the motivational achievement of learning Physical Education; “HBSC Inventory” to find out the healthy habits gained from physical activities; And finally a “stadiometer” was used to obtain the weight and height and thus calculate the body mass index (BMI). The data collection of the cross-sectional Study was conducted in two academic years (2011/12 - 2012/13), in which 3 sessions coinciding with the Physical Education level are required. In the first session, motor coordination was evaluated; questionnaires were applied in the second session (AMPET4 and HBSC); and in the last session the weight and height were measured in a reserved space for the “stadiometer”. Notably, data analysis was descriptive and differential in each of the variable studies: motor, psychological, practical and anthropometric habits of physical activity. Thus the tests were conducted in a univariate and multivariate analysis, calculating the p-value and effect tests. Regarding the Longitudinal Study, data collection was carried out during four years from 2011 to 2014 inclusively. The assessment of motor coordination was performed on each of the four years, however, the 2 questionnaires and anthropometric measures were evaluated in the first and fourth year. Data analyses were also descriptive and comparative among the variables that were put to the test. In the case of motor coordination tests, they were done on repeated measures, whilst, in the case of other variables analysed, they were accomplished through T Tests under comparable samples. The overall results showed that the engine Motor Index in Study 1 was progressive in all male gender studies, however in the females the performance remained constant after reaching 13 years of age. For the Longitudinal Study, this index is stabilized in the first 3 years and at the age of 14 is when the gender differences take place. In the case of males, the performance improves, however, in females worsens. The cross-sectional Study, analysis of variance showed differences in terms of age [F(7, 1958) = 220.70, p < .001; η2 = .101], gender [F(7, 1958) = 29.76, p <.001; η2 = .044], as well as their interaction [F(7, 1958) = 11.90, p <.001; η2 = .018]. They only show significant differences in respect to age in the male set sample, in all groups except between 14 and 15 years old. In the Longitudinal, the multivariate contrasts showed no significant differences in time [F(3,85) = 0.05, p = 0.987, η2 = 0.002] showing a stable level of coordination over the years, but if there were differences between both genders [F(3,85) = 4.64, p = .005] it took place with a noteworthy effect size (η2 = .141). In regards, to the Motivational Achievement for learning Physical Education, in both studies the male sample administered obtained higher scores on all the positive dimensions of the test (commitment to learning, self-assessed competence, and comparable competence). However, on the negative assessment side, namely, anxiety and fear of failure, the female sample scored higher than the male one. In Study 1, the multivariate analysis showed significant differences between the psychosocial dimensions and levels of motor coordination with moderate to significant effect [Lambada de Wilks = .931, F(8, 3282) = 14.99; p = <0.001; η2 = .035]. By the same token, the groups with the best level of motor coordination were the highest scoring ones in the positive dimensions, whilst the lower performing ones, performed better in the negative dimension. In the longitudinal study, there is also differences were also found between the first and fourth years of study in all dimensions, except in self-perceived motor competition. These differences resulted in a significant decrease in the 3 variables from first to fourth year. Regarding, the “HBSC Inventory”, the T test in the longitudinal study showed uniquely the existence of significant differences between the first and fourth year in 2 of the 11 items: perception of physical fitness (p = .006) and perceived health (p = 047), which diminished in the interval time of the study. In the Cross-sectional study, the se differences were also observed in gender (p < .001) and age (p < .001). Similarly, they showed significant differences in all items in respect to the motor coordination level, except in 2 of them; frequency of free time with friends outside of school (p = .580) and the ease to make friends at the educational centre (p = 098). And last but not least, the anthropometric variables, both the results of the Transversal and Longitudinal Study matched both height and weight, pointing out that in both studies a gradual increase in both genders, as they grow older. Notably in the Cross-sectional, these differences in age are found in all groups in both genders, except for the set of girls between 14 and 15 years. Thus both studies concluded that both gains in cm and kg and the mean scores were higher amongst males compared to females. Regarding BMI, the 2 studies concluded that the evolution is parallel, and as pointed cross-sectional study there isn’t differences found in age (p = 792) or in gender (p = 284). However, the Longitudinal study uniquely shows significant difference between the first and fourth year for male set sample [t (41) = -4.01, p < .001]. Finally, in relation to levels of motor coordination, there were significant differences in relation to BMI (p = .012), showing how the “normal weight group” matches the optimal scores of motor coordination. In conclusion, this study reveals how the acquisition of an optimal level of coordination is vital for psychological development, to develop and practice healthy habits, and to maintain a BMI within the normal range for age and gender. Therefore, the development of motor coordination is fundamental to achieving a state of physical and mental wellbeing, and preferable habits to pursuing physical activity.

Examination of body image among Canadian adolescents: Relations with physical activity and screen time

Background: Adolescence is a period of life associated with self-perceptions of negative body image. Physical activity levels are low and screen time levels are also high during this stage. These perceptions and behaviours are associated with poor health outcomes, making research on their determinants important. With adolescent populations, certain groups may be at higher risk of body dissatisfaction than others, and body dissatisfaction may influence individual physical activity and screen time levels. Objectives: The objectives of this thesis were to: 1) describe body image among young Canadians, examining possible health inequalities 2) estimate the strength and significance of associations between body satisfaction, physical activity and screen time, and 3) examine the potential etiological role of biological sex. Methods: Objective 1: The 2013/2014 Health Behaviour in School-aged Children study was employed. Sex-stratified Rao-Scott chi-square analyses were conducted to examine associations between socio-demographic factors and body satisfaction. Objective 2: The 2005/2006 and 2013/2014 cross-sectional and 2006 longitudinal HBSC data sets were used. Sex-stratified modified Poisson regressions were conducted and risk estimates and associated confidence intervals obtained. Results: Objective 1: Among males, being older, of East and Southeast Asian ethnicity, and reporting low SES all were associated with body dissatisfaction. Among females, being older, of Arab and West Asian or African ethnicity, being born in Canada, and reporting low SES were all associated with being body dissatisfied. Objective 2: Cross-sectionally, males who reported ‘too fat’ body dissatisfaction were more likely to be physically inactive. Adolescents of both sexes who reported ‘too fat’ body dissatisfaction were more likely to engage in high levels of screen time. Data from the longitudinal component supported the idea that male ‘too fat’ body dissatisfaction temporally leads to physical inactivity, but showed an inverse relationship between body dissatisfaction and screen time. Conclusions: Objective 1: Future prevention efforts in Canada should target subgroups to effectively help those at greatest risk of body dissatisfaction, and ameliorate potential inequalities at the population level. Objective 2: The presence of these relationships may inform future interventions as part of a multi-factorial etiology, in order to increase physical activity and decrease screen time among youth.

A longitudinal study on the lexical sophistication and sentence length in Finnish upper secondary school students’ English compositions.

The aim of the present study was to find out the level of lexical sophistication and the mean length of sentences used in compositions written by Finnish upper secondary school students of English. In addition, the present study investigated the possible relationship between the two abovementioned variables. The study at hand was longitudinal: as data, a set of 50 compositions were collected in 2014 from the same writers both in the first and the final year of upper secondary school, 25 in the first year and 25 in the final year. In the analysis, an internet-based program called VocabProfile was utilized in order to find out the lexical sophistication of the investigated students. To find out the mean length of sentence and the relationship between these two, I used Microsoft Excel. Findings of the present study include a minor decrease in the use of less frequent vocabulary and a slight increase in the use of the two most frequently appearing thousand words of English: both of these changes were 1.99 percentage points. As for the mean length of sentence, it grew by 1.28 words during upper secondary school. As for the relationship between the two variables, no clear correlations could be found. It became, however, relatively clear that the topic of the composition might have an effect on the results. Thus more research is needed to fully see the effect of lexical sophistication and mean length of sentence on one another. In addition, future research would benefit greatly if all investigated students wrote on the same topic.

Project-based organization as adaptation to risk and opportunity in small firms

Recent years have seen a rapid increase in SMEs working collaboratively in inter-organizational projects. But what drives the emergence of such projects, and what types of industries breed them the most? To address these questions, this paper extends the long running literature on the firm and industry antecedents of new venturing and alliance formation to the domain of project-based organization by SMEs. Based on survey data collected among 1,725 small and medium sized organizations and longitudinal industry data, we find an overall pattern that indicates that IOPV participation is primarily determined by a focal SME’s scope of innovative activities, and the munificence, dynamism and complexity of its environment. Unexpectedly, these variables have different effects on whether SMEs are likely to engage in IOPVs, compared to with how many there are in their portfolio at a time. Implications for theory development are discussed.

Alternative modes of entrepreneurial opportunity exploitation : the case on inter-organizational project

While entrepreneurship research has taken firm formation to be the predominant mode of opportunity exploitation, entrepreneurship can take place through many other types of organizational arrangements. In the present article, we consider one such alternative arrangement, namely the formation of inter-organizational projects (IOPs). We propose a multi-level contingency model that suggests that uncertainty both at the level of the firm and at the level of the environment makes the exploitation of opportunities through IOPs more likely. The model is tested by telephone survey data collected amongst a panel of 1725 SMEs and longitudinal industry data. Our findings provide strong support for the industry-level part of the model, but interestingly, only partial support for the firm level part of the model. This indicates that the effects of uncertainty need to be dissected into different levels of analysis to understand the conditions under which alternative modes of opportunity exploitation can be a prominent entrepreneurial alternative to new firm formation.

Alternative modes of entrepreneurial opportunity exploitation

While entrepreneurship research has taken firm formation to be the predominant mode of opportunity exploitation, entrepreneurship can take place through many other types of organizational arrangements. In the present article, we consider one such alternative arrangement, namely the formation of inter-organizational projects (IOPs). We propose a multi-level contingency model that suggests that uncertainty both at the level of the firm and at the level of the environment makes the exploitation of opportunities through IOPs more likely. The model is tested by telephone survey data collected amongst a panel of 1725 SMEs and longitudinal industry data. Our findings provide strong support for the industry-level part of the model, but interestingly, only partial support for the firm level part of the model. This indicates that the effects of uncertainty need to be dissected into different levels of analysis to understand the conditions under which alternative modes of opportunity exploitation can be a prominent entrepreneurial alternative to new firm formation.

Working covariance model selection for generalized estimating equations

We investigate methods for data-based selection of working covariance models in the analysis of correlated data with generalized estimating equations. We study two selection criteria: Gaussian pseudolikelihood and a geodesic distance based on discrepancy between model-sensitive and model-robust regression parameter covariance estimators. The Gaussian pseudolikelihood is found in simulation to be reasonably sensitive for several response distributions and noncanonical mean-variance relations for longitudinal data. Application is also made to a clinical dataset. Assessment of adequacy of both correlation and variance models for longitudinal data should be routine in applications, and we describe open-source software supporting this practice.

Timing of puberty : Genetic regulation

In the general population, the timing of puberty is normally distributed. This variation is determined by genetic and environmental factors, but the exact mechanisms underlying these influences remain elusive. The purpose of this study was to gain insight into genetic regulation of pubertal timing. Contributions of genetic versus environmental factors to the normal variation of pubertal timing were explored in twins. Familial occurrence and inheritance patterns of constitutional delay of growth and puberty, CDGP (a variant of normal pubertal timing), were studied in pedigrees of patients with this condition. To ultimately detect genes involved in the regulation of pubertal timing, genetic loci conferring susceptibility to CDGP were mapped by linkage analysis in the same family cohort. To subdivide the overall phenotypic variance of pubertal timing into genetic and environmental components, genetic modeling based on monozygous twins sharing 100% and dizygous twins sharing 50% of their genes was used in 2309 girls and 1828 boys from the FinnTwin 12-17 study. The timing of puberty was estimated from height growth, i.e. change in the relative height between the age when pubertal growth velocity peaks in the general population and adulthood. This reflects the percentage of adult height achieved at the average peak height velocity age, and thus, pubertal timing. Boys and girls diagnosed with CDGP were gathered through medical records from six pediatric clinics in Finland. First-degree relatives of the probands were invited to participate by letter; altogether, 286 families were recruited. When possible, families were extended to include also second-, third-, or fourth-degree relatives. The timing of puberty in all family members was primarily assessed from longitudinal growth data. Delayed puberty was defined by onset of pubertal growth spurt or peak height velocity taking place 1.5 (relaxed criterion) or 2 SD (strict criterion) beyond the mean. If growth data were unavailable, pubertal timing was based on interviews. In this case, CDGP criteria were set as having undergone pubertal development more than 2 (strict criterion) or 1.5 years (relaxed criterion) later than their peers, or menarche after 15 (strict criterion) or 14 years (relaxed criterion). Familial occurrence of strict CDGP was explored in families of 124 patients (95 males and 29 females) from two clinics in Southern Finland. In linkage analysis, we used relaxed CDGP criteria; 52 families with solely growth data-based CDGP diagnoses were selected from all clinics. Based on twin data, genetic factors explain 86% and 82% of the variance of pubertal timing in girls and boys, respectively. In families, 80% of male and 76% of female probands had affected first-degree relatives, in whom CDGP was 15 times more common than the expected (2.5%). In 74% (17 of 23) of the extended families with only one affected parent, familial patterns were consistent with autosomal dominant inheritance. By using 383 multiallelic markers and subsequently fine-mapping with 25 additional markers, significant linkage for CDGP was detected to the pericentromeric region of chromosome 2, to 2p13-2q13 (multipoint HLOD 4.44, α 0.41). The findings of the large twin study imply that the vast majority of the normal variation of pubertal timing is attributed to genetic effects. Moreover, the high frequency of dominant inheritance patterns and the large number of affected relatives of CDGP patients suggest that genetic factors also markedly contribute to constitutional delay of puberty. Detection of the locus 2p13-2q13 in the pericentromeric region of chromosome 2 associating with CDGP is one step towards unraveling the genes that determine pubertal timing.