La polyconsommation d’alcool et de cannabis est rapportée par un grand nombre de jeunes adultes canadiens (Flight, 2007). Les recherches épidémiologiques suggèrent que le statut de polyconsommateur est associé à certains comportements à risque, dont la consommation excessive d’alcool et la conduite d’un véhicule automobile sous l’influence de l’alcool (Jones et al. 2001; Mohler-Kuo, et al. 2003; Shillington & Clapp, 2006). Les études qui soutiennent le risque accru de comportements à risque pour les polyconsommateurs se focalisent sur l’effet des substances. En rupture avec cette approche, ce mémoire présente une étude situationnelle de la polyconsommation en examinant l’effet du statut de polyconsommateur et, pour ceux-ci l’effet de la consommation simultanée d’alcool et de cannabis, en situant l’action dans son contexte de survenu et en examinant la contribution du contexte. La probabilité d’avoir conduit une voiture sous l’influence de l’alcool et d’avoir consommé excessivement de l’alcool sera examinée auprès d’étudiants universitaires. La contribution respective des substances, des situations et de l’expérience de la vie universitaire sera examinée. La méthodologie employée repose sur la construction de modèles de régression logistique multiniveaux, à la fois chez l’ensemble des buveurs (10 747 occasions, nichées dans 4396 buveurs) et dans le sous-échantillon des polyconsommateurs (2311 occasions de consommation d’alcool, nichées dans 880 polyconsommateurs). Les données sont issues de l’Enquête sur les campus canadiens (2004), menée auprès d’un échantillon représentatif de 6282 étudiants issus de 40 universités. Le statut de polyconsommateur est associé à la consommation excessive d’alcool, mais pas à la conduite d’une voiture suite à la consommation. Cependant, la consommation simultanée d’alcool et de cannabis n’est pas associée à un risque plus élevé de consommer excessivement de l’alcool, et est négativement associée à la conduite d’une voiture après la consommation. Plusieurs caractéristiques situationnelles sont associées aux deux comportements à l’étude et diminuent la force d’association entre ces comportements et le statut de polyconsommateur. / Alcohol and cannabis polydrug use is a common consumption pattern among Canadian young adults (Flight, 2007). Moreover, epidemiological studies suggest that alcohol and cannabis use is associated with other risky behaviours such as binge drinking and driving a vehicle under the influence of both substances (Jones et al. 2001; Mohler-Kuo et al. 2003; Shillington & Clapp, 2006). Studies which have reported an increased risk for polydrug users focus on substance use. The present study propose a situational approach, in which the effect of poly drug use, as well as the effect of simultaneous use on these risky behaviours are investigated while considering the context in which the consumption takes place. The probability of driving under the influence of alcohol, and the probability of binge drinking are examined among a sample of university students. The respective contribution of substance use, university life experience and drinking situations is examined. The methodology used to perform this study is based on multilevel logistic regression models, among drinkers (10 747 occasions, nested within 4396 drinkers) as well as among the polydrug users subsample (2311 drinking occasions, nested within 880 individuals). The data are drawn from the Canadian Campus Survey conducted in 2004 with a random nationally representative sample of 6282 students from 40 universities. Polydrug status is positively associated with episodic binge drinking, but not with driving under the influence of alcohol. However, cannabis use is not associated with binge drinking, among polydrug users. Moreover, for this specific population, cannabis smoking is negatively associated with driving under the influence of alcohol. Several situational characteristics are associated with either one or both studied behaviours and partially explain the relationship between those behaviours and polydrug use.

Cette recherche étudie l'évolution de l'appui à la souveraineté du Québec entre 1976 et aujourd'hui, en considérant les diverses formulations de la question utilisées par les acteurs politiques et par les sondeurs. Cette question est abordée en faisant une analyse longitudinale multiniveaux de plus de 800 sondages. Une telle approche suggère que les caractéristiques des sondages sont nichées à l'intérieur d'unités contextuelles plus larges, des unités de temps. Cela permet d'une part d'observer quels sont les facteurs liés à la mesure qui influencent l'appui à la souveraineté du Québec et ensuite de voir comment ces mêmes effets varient à travers le temps et comment ils sont influencés par certains événements jugés importants. Il ressort de ces analyses que les propositions séparation, indépendance, souveraineté et souveraineté-association/partenariat entraînent généralement des appuis moyens significativement différents les uns des autres. Il apparaît aussi que certains événements ont eu une influence significative sur l'évolution de l'appui à ces diverses propositions. Enfin, il a été tenté de voir si les appuis propres à ces diverses propositions évoluaient de manière parallèle, s'ils réagissaient similairement aux différents événements considérés dans les analyses ou si au contraire ces appuis pouvaient parfois évoluer de manière différente. Les résultats à cet égard sont intéressants mais non concluants. / The purpose of this research was to study the evolution of the support for Quebec sovereignty from 1976 to 2007, taking into account the many wordings used by political actors and pollsters to refer to the sovereignty project. The methodology used to perform this study is a multilevel and longitudinal analysis of more than 800 polls conducted. Such an approach assumes that measurement-related factors are nested within broader contextual units, in this case time units. Hence, one can find out in what ways measurement-related factors (i.e. question-wording) influence support for Quebec sovereignty on one hand, and how time-related factors (i.e. events and time itself) affect these same measurementrelated factors effects on the other hand. We found that the many constitutional proposals, separation, independence, sovereignty and sovereignty-association/partnership generally do have significantly different mean effects on the support for Quebec sovereignty. We then found that some events cause significant changes in the evolution of sovereignty support. We finally tried to find out if changes in support for these different proposals were similar; in other words, whether these important events affect support for sovereignty in the same way for each proposal or not. In this regard, interesting but yet not conclusive results were found.

This thesis consists of three essays that address various aspects of the economics of maternal health care. The first two essays examine the determinants of utilization of maternal health care services in low-income countries, while the third essay examines the determinants of utilization of prenatal ultrasonography in Canada. The first essay examines the influence of prenatal attendance (as well as a wide array of observed individual-, household- and community-level characteristics) on a woman’s decision to give birth at a health facility or at home for thirty-two low-income countries (across Asia, Sub-Saharan Africa and Latin America). This empirical investigation employs the Demographic and Health Surveys (DHS) data and a two-level random intercept model. The results show that prenatal attendance has a substantial influence on the use of facility delivery in all three geographical regions. Women having four prenatal visits were 7.3 times more likely to deliver at a health facility than those with no prenatal care. The second essay addresses two related questions: what factors determine a woman’s decision to seek prenatal care; and are those the same factors that determine the frequency of care? This investigation also utilizes Demographic and Health Surveys (DHS) data for thirty-two low-income countries (across Asia, Sub-Saharan Africa and Latin America) and applies a two-part and multi-level model to that data. The results suggest that, though a wide range of factors influence both decisions, that influence varies in magnitude across the two decisions, as well as across the three geographical regions. The third essay examines the influence of various socioeconomic and demographic factors on the frequency of prenatal ultrasounds in Canada, while controlling for maternal risk profiles. This investigation utilizes data from the Maternity Experience Survey (MES) of the Canadian Perinatal Surveillance System and employs a count data regression model (the Poisson distribution) to estimate the effect of various factors on the number of prenatal ultrasounds. The results of this investigation suggest that, even after controlling for maternal risk factors, the type of health-care provider, province of prenatal care, and timings of first ultrasound are the strongest predictors of number of ultrasounds.

This thesis consists of three essays that address various aspects of the economics of maternal health care. The first two essays examine the determinants of utilization of maternal health care services in low-income countries, while the third essay examines the determinants of utilization of prenatal ultrasonography in Canada. The first essay examines the influence of prenatal attendance (as well as a wide array of observed individual-, household- and community-level characteristics) on a woman’s decision to give birth at a health facility or at home for thirty-two low-income countries (across Asia, Sub-Saharan Africa and Latin America). This empirical investigation employs the Demographic and Health Surveys (DHS) data and a two-level random intercept model. The results show that prenatal attendance has a substantial influence on the use of facility delivery in all three geographical regions. Women having four prenatal visits were 7.3 times more likely to deliver at a health facility than those with no prenatal care. The second essay addresses two related questions: what factors determine a woman’s decision to seek prenatal care; and are those the same factors that determine the frequency of care? This investigation also utilizes Demographic and Health Surveys (DHS) data for thirty-two low-income countries (across Asia, Sub-Saharan Africa and Latin America) and applies a two-part and multi-level model to that data. The results suggest that, though a wide range of factors influence both decisions, that influence varies in magnitude across the two decisions, as well as across the three geographical regions. The third essay examines the influence of various socioeconomic and demographic factors on the frequency of prenatal ultrasounds in Canada, while controlling for maternal risk profiles. This investigation utilizes data from the Maternity Experience Survey (MES) of the Canadian Perinatal Surveillance System and employs a count data regression model (the Poisson distribution) to estimate the effect of various factors on the number of prenatal ultrasounds. The results of this investigation suggest that, even after controlling for maternal risk factors, the type of health-care provider, province of prenatal care, and timings of first ultrasound are the strongest predictors of number of ultrasounds.

This thesis presents a program of research exploring the multilevel organisational structure of fleet safety management. The aim of this research was to investigate three current fleet safety initiatives, and individual and contextual factors influencing safe driving behaviour in a work vehicle. Three studies were conducted to achieve this aim. This research utilised a sample of employees from a range of Queensland Government agencies.----- Study one evaluated three current fleet safety initiatives within the Queensland Government. From a sample of fleet co-ordinators (N=24) and drivers (N=88), this study established the extent to which specific psychological processes underlying the fleet safety initiatives were adopted, and the attitude change associated with their use. This study found mixed support for the Hypotheses, with the influence of the fleet safety initiatives on fleet co-ordinators' and drivers' attitude change being consistent with processes associated with the persuasive communication framework, and behaviour management. However, the study found no support for the behavioural management processes hypothesised to underlie the incentive scheme (CPP). The findings of the study suggested that while fleet safety initiatives can have an influence on fleet co-ordinator and driver attitude change, their impact depends on the extent to which safety issues are viewed as relevant, and the extent to which there is reinforcement within the organisational environment to support these safety initiatives. Therefore, the findings from this study, combined with existing research into the impact of safety climate, suggest the workplace context needs to be taken into account. For this reason, study two investigated the role of perceptions of the safety climate, in addition to individual attributes, as predictors of self-reported crash involvement.----- Study two applied a framework incorporating driver attributes, including attitudes towards traffic safety and self-efficacy, and drivers' perceptions of the safety climate, as predictors of self-reported crashes in a work vehicle. Within this framework, drivers' perception of the safety climate, and their individual attributes were conceptualised as antecedents of driving performance, and driver safety motivation and knowledge mediated the relationship between these factors and self-reported crashes. A total of 385 drivers participated in this study, which found motivation to drive safely mediated the relationship between driver attributes and self-reported crashes. The initial analysis did not find a significant relationship between safety climate and safety motivation. However, posthoc analyses exploring this non-significant relationship found managerial safety values could be distinguished from other facets of the safety climate construct. Subsequently, the results indicated managerial safety values predicted safety motivation, when drivers perceived a strong safety climate. This study provided a more thorough understanding of the variables predicting driver behaviour at an individual level of analysis. However, a shortcoming is the study did not consider the various influences impacting on drivers' safety perceptions, and individual attributes within the context of the work environment.----- Study three extended on the framework established in study two, and investigated the contribution of leader attributes to the prediction of drivers' safety perceptions, and individual attributes. The leader attribute measures, specifically, perceptions of the safety climate, motivation, knowledge, and work overload were collected from a sample of fleet co-ordinators (N=52) and supervisors (N=88). Through multi-level analyses, both supervisors and fleet co-ordinators were shown to influence the safety perceptions and individual attributes of individuals who drive work vehicles. Support was found for positive relationships between supervisor safety knowledge, and the individual attributes. However, there was a large amount of variation due to group membership unaccounted for by supervisor safety knowledge and the safety performance factors investigated within the supervisor groups. These findings suggested supervisors may not be interacting with drivers in relation to fleet safety matters, but that other factors associated with work group membership are having an impact on drivers' safety perceptions. In comparison, there was a small amount of variation accounted for by fleet co-ordinator group membership. However, the results suggested the fleet co-ordinator leader attributes accounted for a high percentage of this variation in group membership. Support was found for a positive relationship between fleet co-ordinator safety perceptions, and driver safety perceptions. Other results found fleet co-ordinators were engaging in higher workloads to enhance the safety perceptions, and attitudes towards traffic safety of drivers within their groups.----- Overall, these studies establish a multilevel organisational process of effect, whereby individual and leader attributes, and organisational initiatives all play a role in influencing the safety performance of work-related drivers. The results also indicated an unclear structure in the management of fleet safety, as perceived by drivers, and through the roles and responsibilities of supervisors and fleet co-ordinators. The implications of these results for the management of fleet safety are discussed.

This thesis examined the intra-individual relationship between state affect and task performance, with an emphasis on a) the moderating effects of trait affect and task difficulty; and b) the mediating effects of cognitive and affective regulation. Theory and empirical research from the emotion, motivation, and personality literatures was integrated to develop a multilevel model of states, traits, and situational factors as predictors of task performance. Data from five studies were analysed using single- and multi-level techniques to test the hypothesised model of relationships. The findings are reported within three manuscripts, which comprise the body of this thesis. Manuscript 1 presents validation evidence for the psychometric instruments used to measure the self-regulatory components of the model. Three studies (N = 758) were conducted to examine the nomological network of cognitive and affective regulation. In parallel, two new self-report scales were developed to operationalise these constructs within a repeated measures paradigm. The three studies demonstrated that the new cognitive and affective regulation scales were uniquely associated with other self-regulatory, personality, affective and achievement variables at the intra-individual and inter-individual levels. Study 1 provided evidence for the unidimensionality, internal consistency, and construct validity of each scale. Study 2 replicated and extended construct validity evidence using a different sample and performance domain. Study 3 established the utility of each scale for assessing intra-individual variability in cognitive and affective regulation, and their ability to predict performance within individuals. In sum, the three studies suggested that the new measures of cognitive and affective regulation were psychometrically adequate for use in model testing. Manuscripts 2 and 3 tested the intra-individual relationship between state affect and task performance, with a focus on the moderating effects of trait affect and task difficulty (Manuscript 2), and the mediating effects of cognitive and affective regulation (Manuscript 3). Each manuscript analysed different portions of data from two laboratory experiments (N = 182). In each experiment, participants performed multiple trials of an air-traffic control simulation that varied in task difficulty at the inter-individual (Study 4) or intra-individual (Study 5) level. Trait positive and negative affect were measured before the task, whereas state positive and negative affect, cognitive and affective regulation, and task performance were measured at repeated intervals over practice. In Manuscript 2, hierarchical linear modelling demonstrated that state positive affect was positively related, whereas state negative affect was negatively related, to task performance at the intra-individual level of analysis. As hypothesised, the strength of these affect-performance relationships was significantly moderated by trait affect and task difficulty. In both studies, the positive intra-individual relationship between state positive affect and performance was stronger for individuals with high (versus low) trait positive affect, particularly when task difficulty was high (versus low). In contrast, the negative intra-individual relationship between state negative affect and performance only emerged for individuals with low (versus high) trait negative affect, regardless of the level of task difficulty. In Study 4, the intra-individual relationship between state negative affect and task performance was also more pronounced when task difficulty was high (versus low). In Manuscript 3, multilevel multiple-mediation modelling demonstrated that cognitive regulation significantly mediated the intra-individual relationship between state positive affect and performance in both studies, alongside the intra-individual relationship between state negative affect and performance in Study 4. Unexpectedly, affective regulation failed to mediate either of these affect-performance relationships. However, state positive affect was positively related to affective regulation in both studies, whereas state negative affect was positively related to affective regulation in Study 5. Overall, this thesis makes theoretical, empirical and methodological contributions to understanding how affect relates to performance at the intra-individual level of analysis; and for whom, when, and why these relationships emerge. Within a multilevel framework, it integrates interdisciplinary perspectives to identify the affective determinants of performance from two levels of analysis. The results demonstrate that state positive and negative affect can differentially predict intra-individual variability in task performance, and that these effects may a) depend on trait affect and task difficulty; and b) be partially explained by cognitive regulation. These findings emphasise the importance of adopting a multilevel, repeated measures paradigm to examine how affective states, traits, and task demands interactively predict task performance. There is scope for extending this research further by investigating a broader range of moderating and mediating constructs. Practical implications and directions for future research are discussed.

This thesis examined the intra-individual relationship between state affect and task performance, with an emphasis on a) the moderating effects of trait affect and task difficulty; and b) the mediating effects of cognitive and affective regulation. Theory and empirical research from the emotion, motivation, and personality literatures was integrated to develop a multilevel model of states, traits, and situational factors as predictors of task performance. Data from five studies were analysed using single- and multi-level techniques to test the hypothesised model of relationships. The findings are reported within three manuscripts, which comprise the body of this thesis. Manuscript 1 presents validation evidence for the psychometric instruments used to measure the self-regulatory components of the model. Three studies (N = 758) were conducted to examine the nomological network of cognitive and affective regulation. In parallel, two new self-report scales were developed to operationalise these constructs within a repeated measures paradigm. The three studies demonstrated that the new cognitive and affective regulation scales were uniquely associated with other self-regulatory, personality, affective and achievement variables at the intra-individual and inter-individual levels. Study 1 provided evidence for the unidimensionality, internal consistency, and construct validity of each scale. Study 2 replicated and extended construct validity evidence using a different sample and performance domain. Study 3 established the utility of each scale for assessing intra-individual variability in cognitive and affective regulation, and their ability to predict performance within individuals. In sum, the three studies suggested that the new measures of cognitive and affective regulation were psychometrically adequate for use in model testing. Manuscripts 2 and 3 tested the intra-individual relationship between state affect and task performance, with a focus on the moderating effects of trait affect and task difficulty (Manuscript 2), and the mediating effects of cognitive and affective regulation (Manuscript 3). Each manuscript analysed different portions of data from two laboratory experiments (N = 182). In each experiment, participants performed multiple trials of an air-traffic control simulation that varied in task difficulty at the inter-individual (Study 4) or intra-individual (Study 5) level. Trait positive and negative affect were measured before the task, whereas state positive and negative affect, cognitive and affective regulation, and task performance were measured at repeated intervals over practice. In Manuscript 2, hierarchical linear modelling demonstrated that state positive affect was positively related, whereas state negative affect was negatively related, to task performance at the intra-individual level of analysis. As hypothesised, the strength of these affect-performance relationships was significantly moderated by trait affect and task difficulty. In both studies, the positive intra-individual relationship between state positive affect and performance was stronger for individuals with high (versus low) trait positive affect, particularly when task difficulty was high (versus low). In contrast, the negative intra-individual relationship between state negative affect and performance only emerged for individuals with low (versus high) trait negative affect, regardless of the level of task difficulty. In Study 4, the intra-individual relationship between state negative affect and task performance was also more pronounced when task difficulty was high (versus low). In Manuscript 3, multilevel multiple-mediation modelling demonstrated that cognitive regulation significantly mediated the intra-individual relationship between state positive affect and performance in both studies, alongside the intra-individual relationship between state negative affect and performance in Study 4. Unexpectedly, affective regulation failed to mediate either of these affect-performance relationships. However, state positive affect was positively related to affective regulation in both studies, whereas state negative affect was positively related to affective regulation in Study 5. Overall, this thesis makes theoretical, empirical and methodological contributions to understanding how affect relates to performance at the intra-individual level of analysis; and for whom, when, and why these relationships emerge. Within a multilevel framework, it integrates interdisciplinary perspectives to identify the affective determinants of performance from two levels of analysis. The results demonstrate that state positive and negative affect can differentially predict intra-individual variability in task performance, and that these effects may a) depend on trait affect and task difficulty; and b) be partially explained by cognitive regulation. These findings emphasise the importance of adopting a multilevel, repeated measures paradigm to examine how affective states, traits, and task demands interactively predict task performance. There is scope for extending this research further by investigating a broader range of moderating and mediating constructs. Practical implications and directions for future research are discussed.

No description available.

Introdução-O aleitamento materno promove vantagens para a criança, mulher e sociedade. Entretanto as taxas de amamentação, principalmente o aleitamento materno exclusivo estão abaixo dos níveis desejados em diversos países. A identificação dos determinantes da interrupção precoce e a quantificação da interação destes fatores é crucial para estabelecer, avaliar e promover políticas públicas em prol das práticas de amamentação. Objetivo-Estimar os efeitos dos fatores associados à prevalência do aleitamento materno exclusivo, predominante e continuado no primeiro e segundo anos de vida em países da América Latina e Caribe no período de 2000 a 2016. Métodos-Foram utilizados microdados coletados no projeto Monitoring and Evaluation to Assess and Use Results Demographic and Health Survey. Para complementar as informações foram recolhidos dados agregados do World Bank. Os indicadores analisados foram aleitamento materno exclusivo (AME), aleitamento materno predominante (AMP) e aleitamento materno continuado (AMC) no primeiro e segundo anos de vida; os valores foram expressos em prevalência e estratificados por quintos de riqueza. Selecionou-se o último inquérito disponível de cada país da América Latina e Caribe com dados entre os anos de 2000 e 2016. As variáveis foram selecionadas segundo disponibilidade e enquadramento no modelo conceitual proposto por ROLLINS et al. (2016). O modelo utilizado no estudo foi elegido após análise com Regressão de Poisson e, posteriormente, foi realizada análise logística multinível de efeitos mistos para quantificar o efeito de cada determinante do efeito fixo ajustado pelo efeito randômico. Resultados-A variação das prevalências dos indicadores foi de 7,4% a 60,3% no AME, 13,9% a 71% no AMP, 35,0% a 86,6%no AMC no primeiro ano e 18,9% a 55,0% no AMC no segundo ano de vida. Bolívia e Guatemala se destacam por apresentar as maiores taxas de AME, 60,3% e 54,2%, respectivamente. Os países apresentaram maior prevalência dos indicadores nos quintos inferiores de renda, destacando-se as associações do padrão de aumento da magnitude do efeito negativo à medida que o quinto de renda aumenta. O AME e AMP se associaram diretamente à \"amamentação na primeira hora\", \"estado conjugal\" e \"queria o último filho; \"idade materna\" e \"status de trabalho\" se associaram inversamente aos indicadores AMC no 1º e 2º anos de vida. Na análise de efeitos mistos, os indicadores AME e AMP se associaram diretamente a \"amamentação na primeira hora\" e \"estado conjugal\" e inversamente às variáveis \"local do parto\" e \"status de trabalho\". A escolaridade materna apresentou associação inversa com os indicadores AMP, AMC no 1º e 2º anos de vida e o quinto de riqueza apresentou associação inversa com todos os indicadores estudados. O AME apresenta maior variância explicada pelo nível individual, enquanto o AMC no 2º ano de vida sofre grande influência pelo nível contextual (8,99e-15% e 69,7%, respectivamente). Conclusão-As variáveis relacionadas ao indivíduo foram as que melhor explicaram a variância do modelo para o AME, AMP e AMC no 1º ano de vida. As variáveis relacionadas ao contexto explicam maior variabilidade da prevalência do AMC no 2º ano de vida, exibindo a maior interferência do PIB e de influências de âmbito nacional. / Introduction-Breastfeeding promotes benefits for the child, woman and society. However, breastfeeding rates, especially exclusive breastfeeding, are below desired levels in many countries. The identification of the determinants of the interruption precociated and the quantification of interaction factors are crucial to establish, evaluate and promote public policies in breastfeeding practices. Objective-Estimate the effects of factors associated with the prevalence of exclusive, predominant and continued breastfeeding in the first and second years of life in Latin American and Caribbean countries from 2000 to 2016.Methods- We used microdata collected in the project Monitoring and Evaluation to Assess and Use Results Demographic and Health Survey. In addition to the information, aggregated World Bank data was collected. The indicators analyzed were exclusive breastfeeding (EB), predominant breastfeeding (PB) and continuous breastfeeding (CB) in the first and second years of life; the values were expressed in prevalence and stratified by quintiles of wealth. The last available survey of each Latin American and Caribbean country with data between the years 2000 and 2016 was selected. The variables were selected according to availability and framing in the conceptual model proposed by ROLLINS et. al. (2016). The model used in the study was chosen after analysis with Poisson Regression and later, multi-level mixed-effects logistic analysis was performed to quantify the effect of each determinant of the fixed effect adjusted by the random effect. Results- The prevalence of indicators ranged from 7.4% to 60.3% EB, 13.9% to 71% in PB, 35.0% to 86.6% in CB in the first year, and 18.9% to 55.0% in CB in the second year of life. Bolivia and Guatemala stand out for having the highest rates of EB, 60.3% and 54.2%, respectively. The countries presented a higher prevalence of indicators in the lower quintiles of income, especially the associations of the pattern of increase of the magnitude of the negative effect as the fifth of income increases. EB and PB were directly associated with \"first-time breastfeeding\", \"marital status\" and \"wanted the last child; \"maternal age\" and \"work status\" were inversely associated with the CB indicators in the 1st and 2nd years of life. In the mixed effects analysis, the EB and PB indicators were directly associated with \"first-hour breastfeeding\" and \"marital status\" and inversely to the variables \"place of birth\" and \"work status\".The maternal schooling presented an inverse association with the PB, CB indicators in the 1st and 2nd years of life, and the fifth of the wealth had an inverse association with all the indicators studied. The EB shows greater variance explained by the individual level, while the CB in the second year of life suffers great influence at the contextual level (8.99e-15% and 69.7%, respectively). Conclusion- The variables related to the individual were the ones that best explained the variance of the model for EB, PB and CB in the 1st year of life. The variables related to the context explain a greater variability of the prevalence of CB in the second year of life, showing the greater interference of GDP and national influences.

Science education becomes more important for future national development globally in high-technology-based society. In reaction to the trend, the International Association for the Evaluation of Educational Achievement (IEA) has conducted achievement tests in science along with mathematics, called TIMSS every four years. In TIMSS 2003, while Korea was a higher-performing country, South Africa was ranked in the lower-performing countries. Korea features homogenous demography, centralized curriculum, and competitive educational zeal while South Africa is characterized by multicultural demography with various languages, and previously segregated schools based on races. The current research, which is a secondary analysis of TIMSS 2003 data, aimed at explaining the differences and similarities by identifying factors most likely to influence science achievement in the two countries. A conceptual research framework was built on the comprehensive literature review which involved mainly school effectiveness research and factors related to science achievement. The conceptual framework consists of multi-levels, viz., student, classroom, school, and context, and three key concepts, namely time on task, opportunity to learn, and quality. Two research questions were formulated to reach the goal of the research and the first question is: To what extent does TIMSS 2003 reflect factors related to effective science education? Data from the student, teacher and school questionnaires were included in conjunction with the achievement data and analysed by means of factor, reliability and correlation analyses. The factors found to influence science achievement in three levels are as follows: at the student level, books at home, attitudes towards science, time on task; at the classroom level, time scheduled for science and teacher interaction; at the school level, school size, community size, and student background. The second research question is: To what extent do the factors derived from the analysis explain the differences in the achievement of Korean and South African students? To answer this question, the current research used multilevel modeling techniques to deconstruct the total variance in achievement into within- and between-classroom/school level. The strongest predictor is attitudes towards science in both countries at the student level. Student background in Korea and safety in school in South Africa is the strongest predictor of science achievement at the classroom/school level. Furthermore, educational resources such as books at home and educational level of father are significant in Korea while language, teacher qualification, physical resources, and educational leadership are significant in South Africa. For Korea, 93% of total variance in science achievement occurred at the student level while only 7% was attributable to the classroom/school level. For South Africa, 41% of the total variance was assigned at the student level and 59% at the class/school level. From this comparative study, it was recommended that development of student-centred teaching practices to address negative attitudes to science in Korea be considered as opposed to basic issues such as improving teachers’ subject knowledge, developing language skills, and fostering a culture of learning to improve science performance in South Africa. / Thesis (PhD)--University of Pretoria, 2011. / Science, Mathematics and Technology Education / unrestricted