Cette thèse a un double objectif. En premier lieu, elle présente les méthodes ayant permis de construire deux bases de données historiques relatives aux départements français. La première met à disposition les tables de mortalité départementales sur la période 1901-2014. La seconde permet de disposer des distributions départementales de revenu sur la période 1960-2014. En second lieu, cette thèse présente les travaux issus de l'utilisation conjointe de ces deux bases de données et d'autres statistiques: ils concernent aussi bien les dynamiques longues des inégalités spatiales que certains évènements historiques. Ainsi, l'analyse de la répartition spatiale de la population depuis le milieu du 19eme siècle permet de comprendre à la fois la dynamique induite par l'exode rural, mais aussi par les nouvelles tendances des migrations d'aujourd'hui, différentes selon les âges. L'analyse des inégalités de mortalité depuis 200 ans montre quant à elle que les inégalités ont largement baissé depuis la fin du 19ème siècle, alors que la géographie de la surmortalité a profondément changé. Enfin, l'analyse des inégalités spatiales de revenus révèle une baisse continue des inégalités depuis les années 1920, baisse qui n'intervient que depuis 1950 si l'on introduit les inégalités de mortalité dans un indicateur synthétique de bien-être. La thèse se conclut par l'analyse des migrations internes durant la seconde guerre mondiale : leur caractère à la fois massif et à destination de la zone libre témoigne aussi bien de l'impact qu'a eu cet évènement sur la démographie française que de la formidable quête de la liberté des français de l'époque, peu entravée par la ligne de démarcation. / This thesis has a dual purpose. First, it presents the methods used to build two new historical databases relating to departments. The first database provides the departmental lifetables for the period 1901-20-14. The second database provides the departmental distributions of income over the period 1960-2014. Second, this thesis presents the first work resulting from the joint use of these two databases and other statistics: they concern both the dynamics of spatial inequalities and some specific historical events. Thus, the analysis of the spatial distribution of the population since the middle of the 19th century allows to understand the dynamics induced by the rural exodus, but also by the new trends of today's migrations. The analysis of mortality inequalities over the last 200 years shows that inequalities have fallen dramatically since the end of the 19th century, while the geography of excess mortality has changed. Finally, the analysis of spatial income inequalities reveals a continuous decline since the 1920s. This decline occurred only since 1950 spatial inequalities are observed using a synthetic indicator of welfare, combining both mortality inequalities and income inequalities. The thesis ends with the analysis of internal migrations during the Second World War: these migrations were massive, and clearly oriented towards the free zone. These results testify both to the impact of this event on French demography, and to the quest for freedom of the French of that time, little hampered by the demarcation line.

Made available in DSpace on 2014-12-17T15:43:48Z (GMT). No. of bitstreams: 1 GenycleaSO_DISSERT.pdf: 2009007 bytes, checksum: 19a6d4fcd41f06f499ef6b4431291aa7 (MD5) Previous issue date: 2012-12-17 / In Brazil, despite the decline in infant mortality in recent decades it still has high rates going against recommended by WHO. Being the largest percentage of infant mortality rate composed of neonatal deaths. Objective: A study was conducted to analyze the spatial distribution of neonatal mortality and its correlation with the biological, socioeconomic and maternal and child health care in the Brazilian states in the period from 2006 to 2010. Method: The study made thematic maps and correlation (LISA) for verification of spatial dependence and multiple linear regression models. Results: Was found that there is no spatial autocorrelation for neonatal mortality in the Brazilian states (R = 0.002, p = 0.48). Most of variables were correlated (r> 0.3, p <0.05) with neonatal mortality, forming clusters in the North and Northeast, with the highest rates of teenage mothers, low household income per capita, lower prenatal appointments and beds of Neonatal Intensive Care Unit. The number of Neonatal UCI beds remained independent effect after regression analysis. Conclusion: The study concludes that regional inequalities in living conditions and especially the access to maternal and child health services contribute to the unequal distribution of neonatal mortality in Brazil / No Brasil, apesar do decl?nio da mortalidade infantil nas ?ltimas d?cadas esta ainda possui altas taxas indo de encontro ao preconizado pela OMS. Sendo o maior percentual da taxa de mortalidade infantil composto pelos ?bitos neonatais. Objetivo: Realizou-se um estudo para analisar a distribui??o espacial da mortalidade neonatal e sua correla??o com os fatores biol?gicos, socioecon?micos e de aten??o ? sa?de materno-infantil nos estados brasileiros, no per?odo de 2006 a 2010. Desenho: ecol?gico, utilizando os estados brasileiros e o Distrito Federal como unidades de an?lise. M?todo: Foram constru?dos mapas tem?ticos e de correla??o (LISA) para verifica??o de depend?ncia espacial e modelos de regress?o linear m?ltipla. Resultados: Verificou-se que n?o h? autocorrela??o espacial para mortalidade neonatal nos estados brasileiros (I =0,002; p=0,48). A maioria das vari?veis estavam correlacionadas (r >0,3, p<0,05) com a mortalidade neonatal, formando clusters em estados do Norte e Nordeste, com maiores taxas de m?es adolescentes, renda domiciliar per capta baixa, menor realiza??o de consultas de pr?-natal e de leitos de UTI Neonatal. O n?mero de leitos de UTI Neonatal manteve efeito independente ap?s a an?lise de regress?o. Conclus?o: As desigualdades regionais das condi??es de vida e principalmente de acesso aos servi?os de sa?de materno-infantil contribuem para a distribui??o desigual da mortalidade neonatal no Brasil