Vieillir immigré est devenu une situation commune à un nombre grandissant de personnes en France. En 2015, près d’1,5 million de personnes immigrées étaient âgées de 60 ans et plus. La constante augmentation du nombre d’immigrés âgés constitue une question sociale et politique en France, ce qui a été confirmé tant par les récents travaux de l’Assemblée nationale que par l’avancée des connaissances scientifiques. Le nombre de travaux produits sur des territoires différents, là où les migrants âgés sont plus ou moins présents démographiquement reste encore aujourd’hui restreint. Face à ce constat, l’objectif de cette thèse par articles est de proposer une lecture territoriale et gérontologique de la vieillesse en migration. Deux terrains d’étude ont été retenus dans le cadre du projet doctoral (premièrement, les villes d’Angers et de Cholet en Maine-et-Loire et deuxièmement, la Ville de Montreuil en Seine-Saint-Denis). Les terrains de recherche ont été l’occasion d’analyser les conditions de vieillissement des personnes immigrées dans des contextes spatiaux différents (analyse des données territoriales et du discours des acteurs médico-sociaux, associatifs et politiques intervenants auprès des immigrés vieillissants). Dans la continuité, une analyse des récits biographiques d’un échantillon de migrants âgés résidant en Maine-et-Loire a été effectuée pou mettre en exergue les enjeux sociaux et spatiaux structurant les expériences de vieillissement des immigrés. Les résultats de recherche produits témoignent de la nécessité d’appréhender l’hétérogénéité de ces publics et d’étudier leurs besoins en regard de leurs parcours et des espaces qui structurent leur (fin de) vie. / In France, getting old as an immigrant became a condition shared by a growing number of people. In 2015, about 1.5million of immigrants are 60years old or more. In France, the constant increase in the number of elderly immigrants represents a social and political issue, a fact confirmed by the recent works of the national assembly as well as by the development of scientific research. Nonetheless, the diversity of studies dealing with different territories inhabited by elderly immigrants remains limited. Consequently, this thesis (structured around several scientific articles) aims at proposing a representation of the “migrating elderly”, articulating a territorial and gerontological approach. Two different territories were taken into consideration: Angers andCholet (Maine-et-Loire), and Montreuil (Seine-Saint-Denis).This selection provided the opportunity to analyse the aging conditions of migrating population which evolve according to different spatial contexts. In this study, we considered territorial data, medico-social accounts, as well as the work of associations and political interventions to elderly immigrants.In addition, biographical facts and stories from as ample of elderly migrants in Maine-et-Loire were analysed to highlight the political and social issues that structure the aging conditions of immigrants.The result of this work proves that it is still essentia lto take into account the heterogeneity of these populations. It also underlines the importance of understanding their needs in terms of their own life journeys (and the end of it) which is inextricable of the territory they occupy.

This thesis explored cognitive, psychosocial, and physical domains of wellbeing to find out their contribution to successful ageing in 70-year old individuals. Discovering groups with different patterns of wellbeing and their correlates may be informative about what constitutes success in old age. The objectives were to find out whether distinct groups within and across domains of wellbeing existed, and to find out the variables associated with the resulting groups. Using a cross-sectional design on the Lothian Birth Cohort 1936 (LBC1936, maximum n = 1091), which is a group of community-dwelling 70 year-olds, latent class analysis (LCA) was used to explore possible patterns of ageing in domains of cognitive, psychosocial, and physical function. Demographic, personality, and lifestyle variables that were not used in the LCA were used to characterise the resulting groups. The first study investigated cognitive ability. Individuals were grouped according to their scores on general cognitive ability (g), memory, and speed. I accepted a 3-group solution, including High- (n = 749, 69%), Average- (n = 303, 28%), and Low- (n = 39, 4%) cognition groups. Results indicated the presence of a strong dimension: people who did well on one component also did well on others, and failed to show any indication of uneven patterns of scores. In the second study on psychosocial wellbeing individuals were grouped according to their scores on physical function, quality of life, and emotional wellbeing. A 5-group solution was accepted. High (n =515, 42.7%), Average (n = 417, 38.3%), and Poor (n = 37, 3.4%) Wellbeing groups were identified; however, contrasting patterns of wellbeing across components were noticed in the two other groups: one group scored relatively highly on physical function, but low on emotional wellbeing (High Function/ Low Spirits, n = 60, 5.5%), while another group showed low physical function but relatively high emotional wellbeing (Low Function/High Spirits, n = 62, 5.7%). The next study investigated the physical fitness domain: groups were determined on physical fitness, lack of inflammation, and lack of morbidity. Two groups, High Physical Fitness (n = 757, 73.3%) and Low Physical Fitness (n = 291, 26.7%) were identified, which, like the cognitive domain, also indicated a continuous pattern of wellbeing. In the final study individuals were grouped according to their scores on all variables reflecting cognitive, psychosocial, and physical function. I identified 3 groups showing high or uneven patterns of wellbeing. The majority of individuals fell in the High Wellbeing group (n = 712, 65.3%). The two other groups contained either individuals scoring high on cognitive measures but poorly on psychosocial and physical measures (the Low Bio-Psychosocial group, n = 158, 14.5%), or individuals scoring low on cognitive measures but highly on psychosocial and physical measures (the Low Cognition group, n = 221, 20.3%). Intelligence, personality and health behaviours showed salient differences amongst the groups in all studies. Overall, high childhood cognitive ability, low scores on Neuroticism, and avoiding smoking were associated with high wellbeing. Overall, results demonstrated that although wellbeing in old age is primarily dimensional, there is evidence of groups showing uneven patterns of function, indicating that individuals could show relatively successful patterns in some areas of wellbeing despite relatively poor functioning in other areas. Awareness of the importance of lifelong intelligence and personality traits and health practices to later-life wellbeing amongst health-care professionals and policymakers may help address risk-prevention, and improve compliance and patient-practitioner relationships to reduce health inequalities.

To investigate whether multilingualism can be used to predict working memory in older adults this study has used a categorical variable as well as a continuous variable to measure multilingualism. This was done while also accounting for possible confounding variables such as intelligence. Furthermore, subjective memory was investigated regarding how it can be affected by multilingualism and if subjective memory is an accurate measure of objective memory. A positive correlation between the Ospan score and composite language score was found while correlations could not be found for subjective memory. However, according to the hierarchical multiple regressions that were done, composite language score was not a significant predictor of Ospan score when demographics and intelligence were accounted for, only intelligence could predict both Span and Ospan scores. In regards to subjective memory, it was found that composite language scores could predict subjective memory and that those who spoke four languages rated themselves worse than monolinguals. Working memory tasks could not, however, predict subjective memory. The conclusions that were made were that multilingualism does not have an effect on working memory capacity but that composite language score and quadrilingualism have an effect on subjective memory.

Older adults are increasingly using social networking sites such as Facebook to maintain social relationships. A growing body of literature indicates that frequent communication with family members and close friends influences successful ageing. Consequently, the purpose of this study is to establish the nature of the relationship between the impact of Facebook communication and successful ageing on older adults living in the rural municipality of Hagfors. This study defines an older adult as a person aged 55 years and above. Self-reported life-satisfaction, social engagement and social trust are used as indicators of successful ageing. Predictions about Facebook communication and successful ageing are tested by theories of socio-emotional selectivity and social capital. This is based on 15 semi-structured interviews conducted with older adults living in the municipality of Hagfors. The results indicate that having small social connections and frequently exchanging privately composed communication with family members and close friends promotes successful ageing. Facebook is found to be a potential source of social capital.

Quality of life is one of the main concerns in long-term care amongst ageing populations in many countries. This problem is historically unique and increases the demand for research material. This thesis looks at how different societies promote or fail to promote successful ageing of long-term care of older people, and considers how countries may learn from one another in their search for solutions. The three countries studied (England, the Netherlands and Taiwan) correspond to Esping-Anderson's three types of welfare system (Esping-Anderson, 1990). In addition, the Asian-European dimension has been employed as it is a neglected one. Data sources included conceptual, empirical and statistical documents on long-term care of older people. Moreover, this research used identical qualitative cross-national research methods on three levels in each country: national, county and municipal. A total of 142 interviews were carried out in 2004. This aim of this study was broadly to rank the three welfare systems where there were clear differences but to qualify this by pointing out the complexities and difficulties of mixed economy comparisons. The overall conclusion is that the Netherlands provides higher quality care to older people, thus confirming Esping-Andersen's finding about the superiority of social democratic systems. In reviewing current policies and research in needs and successful ageing, this qualitative comparative study has focused on needs, social inclusion, power and autonomy, care resources as well as partnership as crucial concepts in care systems and discovered good practice in each and lessons to be learnt.