Search

Background: Subjective survival probabilities are affected by individual-specific judgment and vary by factors known to differentiate actual mortality. Objective: The aim of this study is to evaluate whether sociodemographic characteristics, physical and mental health, and lifestyle are incorporated into subjective survival probabilities of Europeans aged 50 or higher. Methods: We use data from Wave 6 of the Survey of Health, Ageing and Retirement in Europe (SHARE) and period life tables from the Human Mortality Database (HMD). For the statistical analysis we employ multinomial logistic regression models. Results: Our results show that common factors drive the self-reported subjective survival probabilities. Certain factors affecting actual mortality are considered when forming subjective survival probabilities: income, education, poor physical and mental health, activities of daily living (ADLs), smoking, physical activity, diet, quality of life, and number of children. Other factors are not considered in a manner consistent with actual mortality patterns: age, gender, marital status, and body weight. The findings regarding cognitive function are inconclusive; whereas some aspects seem to be integrated in subjective survival probabilities (e.g., memory or self-writing skills), others are not (e.g., numeracy or orientation in time). Contribution: The contribution of this study is the grouping of sociodemographic, health, and lifestyle characteristics according to the subjective survival probabilities' direction and consistency with general population mortality and actual mortality patterns. Hence, we assess which traits are incorporated in the formation of subjective survival probabilities among Europeans aged 50 or higher.

Background: In the second half of the 20th century, remarkable marriage changes were seen: a great proportion of never married population, high average age at first marriage, and large variance in first marriage timing. Although it is theoretically possible to separate these three elements, disentangling them analytically remains a challenge. Objective: This study's goal is to answer the following questions: Which of the three effects, nonmarriage, delayed marriage, or expansion, has the most impact on nuptiality changes? How does the most influential factor differ by time periods, birth cohorts, and countries? Methods: To quantify nuptiality changes over time, we define the measure 'expected years ever married' (EYEM). We illustrate the use of EYEM, looking at time trends in 15 countries (six countries for cohort analysis) and decompose these trends into three components: scale (the changes in the proportion of never married - nonmarriage), location (the changes in timing of first marriage - delayed marriage), and variance (the changes in the standard deviation of first marriage age - expansion). We used population counts by sex, age, and marital status from national statistical offices and the United Nations database. Results: Results show that delayed marriage is the most influential factor on period EYEM's changes, while nonmarriage has recently begun to contribute to the change in North and West Europe and Canada. Period and cohort analysis complement each other. Conclusions: This study introduces a new index of nuptiality and decomposes its change into the contribution of three components: scale, location, and variance. The decomposition steps presented here offer an open possibility for more elaborate parametric marriage models.

1. IntroductionOn average, 15% of European children live in one-parent households, most of them headed by women, and this kind of household has a higher risk of poverty than couple-headed households (OECD 2014). Furthermore, there is evidence that under some conditions, nurturing children in single-mother households has negative consequences for the children's cognitive and emotional development that affect them throughout their entire lives (see Sigle-Rushton and McLanahan 2004 for a review). In Southern European countries, described as family-based welfare regimes (Esping-Andersen 1999), state protection for children is among the lowest in the European Union. However, when comparing poverty ratios between children nurtured in single-parent families and those in two parent households, family-based regimes show smaller differences than other welfare regimes. In the mid-1990s the risk of poverty among children residing in single-parent families was as much as twice as high as the risk of poverty among children residing in two-parent households, while this ratio was much higher in the other Western nations (Luxembourg Income Study, Key Figures. Data from mid-1990's)It seems paradoxical that the relative standard of living of children in single-parent households is better where state social investment is lower. However, this ceases to be such a paradox when we take into consideration the composition effect of single parenthood in Southern Europe. Previous studies with data from the 1990s or early 2000s showed a positive relationship between single motherhood and educational level in Southern Europe (Flaquer, Almeda, and Navarro 2006; Garriga 2010; McLanahan 2004), while this relationship was negative in Northern Europe (McLanahan 2004; Kennedy and Thomson 2010). Furthermore, a high percentage of single mothers in Southern Europe live with their parents in order to balance work and family life or escape from poverty (Trevino 2006; Flaquer and Garriga 2009).The purpose of our study is to examine whether such a composition effect still exists in Southern Europe, given the changing social context characterized by: a) a dramatic change in labor market conditions, b) huge changes in the ethnic composition of the younger population as a result of massive immigration waves, and c) deep-seated changes in family values affecting divorce, cohabitation, and out-of-wedlock patterns. Given that all these changes have not happened with equal intensity among all southern countries, we have selected Spain and Italy as two cases representing two different trends (Moreno and Mari-Klose 2013). Our results confirm that the relationship between educational level and single parenthood has been clearly inverted in Spain, where we have observed a positive association between less-educated women and single motherhood, where the risk of becoming a single mother is now higher among the less-educated. Similar changes, though lower in intensity, are happening in Italy among younger women and in the wealthiest regions. All in all, these changes seem more related to long-term trends than to any effect derived from the economic crisis or migration.2. The Second Demographic Transition and educational variation in single parenthoodSingle parenthood arises from widowhood, out-of-wedlock birth, divorce, and remarriage. Single parenthood is an option for very young girls who get pregnant out of wedlock, most of them from working-class or underclass backgrounds (Kearney and Levine 2012). This source of single parenthood is more frequent in the American continent than in Western Europe, with the exception of the United Kingdom and Ireland, and is very rare in Southern Europe (Part et al. 2013). Widowhood is another source of single parenthood, and for a long time it was the main cause of single parenthood in societies where divorce was forbidden or tightly restricted by law, such as in pre-democratic Spain (Flaquer 1998). Widowhood is more frequent among the lower social strata, given their shorter life expectancy. …

Cette thèse vise un apport conceptuel et empirique au sujet du rapport à l’alcool chez les hommes et les femmes âgés de 55 ans et plus provenant de la population générale du Canada. Les deux objectifs généraux sont de mieux comprendre l’influence du contexte social/environnemental sur la consommation d’alcool des adultes âgés et de mieux comprendre la relation entre la consommation d’alcool et la santé. La thèse est composée de deux articles empiriques associés à chacun des objectifs. Plus précisément, le premier article tente d’évaluer l’impact des changements associés à l’alcool depuis les années 1990 dans le contexte canadien sur la consommation d’alcool (effet période) en comparant des adultes âgés de 55 à 74 ans provenant de deux enquêtes populationnelles canadiennes : CADS de 1994 (Canada’s Alcohol and other Drugs Survey) et GENACIS de 2004 (GENder, Alcohol, and Culture: an International Study). Il vérifie également si le contexte a un effet significatif sur les différences observées en tenant compte de l’âge, de la perception subjective de la santé et de la composition sociodémographique des échantillons. Les résultats montrent que les taux de buveurs, de consommation excessive, de buveurs de bière et de vin sont plus élevés en 2004, dans un contexte social/environnemental plus favorable à l’alcool. De plus, les profils de consommation d’alcool ne diffèrent pas entre 1994 et 2004 (volume, fréquence, quantité, quantité maximale), signifiant probablement l’influence des changements physiques qui accompagnent le vieillissement normal sur la modulation de la consommation d’alcool avec l’âge. Le deuxième article vise à dégager une typologie de buveurs et de non-buveurs chez les adultes âgés canadiens, à évaluer comment cette typologie est associée à l’état de santé ainsi qu’à comparer le rôle de l’alcool à celui d’un ensemble d’autres déterminants de la santé dans l’association à la santé (ici santé autoperçue). Les résultats issus de l’enquête GENACIS de 2004 montrent que les buveurs se perçoivent en meilleure santé que les non-buveurs, peu importe le profil de consommation d’alcool rapporté. Il est probable que l’état de santé autoperçue dicte le statut de buveur ou de non-buveur et non pas l’inverse. Les résultats montrent aussi que les caractéristiques démographiques, psychosociales et les comportements liés à la santé des adultes âgés sont davantage associés à la perception de leur santé que leur consommation (ou non-consommation) d’alcool seule. Les résultats de ce second article indiquent notamment que les bienfaits de la consommation d’alcool se définissent probablement par la biologie, mais pas seulement par la biologie. Il est recommandé de poursuivre l’étude du rapport à l’alcool chez les adultes âgés qui peut varier d’une cohorte et d’une période à l’autre. D’un point de vue de santé publique, la conceptualisation d’un modèle de santé qui cible les adultes âgés implique une meilleure compréhension de l’impact de la consommation d’alcool et d’autres déterminants de la santé sur l’état de santé. Les contributions et implications pratiques de la thèse ainsi que les principales limites et pistes de futures recherches sont présentées dans la discussion., The aim of the thesis is to contribute to a better conceptual and empirical understanding of alcohol consumption among men and women aged 55 years and older from the general Canadian population. The two general objectives pursued are to gain a better understanding of the potential influence of the social/environmental context on drinking habits among older adults and to achieve a clearer discernment of the relation between alcohol consumption and health. The thesis comprises two empirical articles relating to each respective objective. More specifically, the first article verified whether contextual changes associated to alcohol and unique to the decade spanning the 1990s may have had an impact on alcohol consumption (period effect) by comparing two subsamples of older adults from two Canadian surveys: the 1994 CADS (Canada’s Alcohol and other Drugs Survey) and 2004 GENACIS (GENder, Alcohol, and Culture: an International Study). The article also assessed if the context had a significant effect on the differences observed when age, subjective perception of health and the sociodemographic composition of the samples were taken into consideration. Results first show that, in a social/environmental context that is more favourable to drinking in general, the rates of current drinkers, binge drinking, beer and wine drinkers are significantly higher in 2004. In addition, drinking profiles were not significantly different between 1994 and 2004 (volume, frequency, quantity, maximum quantity), pointing at the potential impact of normal physiological changes due to aging on drinking profiles. The second article aimed to build a drinking typology (including drinkers and non-drinkers) among older Canadian adults, to assess how the drinking typology was related to health status and to compare the role of alcohol with other determinants of health (demographic, psychosocial and health-related factors) in explaining health status. To do so, the subjective perception of health (subjective health) was used as a proxy for actual health status. Results from the 2004 GENACIS survey indicated that drinkers perceived themselves as being in better physical health than nondrinkers, regardless of the drinking pattern. Health status may explain drinking status and not the opposite. Results also showed that the demographic, psychosocial and health-related characteristics of older adults contributed significantly more to the variation in subjective health than the drinking status alone. Health benefits of drinking are explained by biological factors, but not exclusively by biological factors. In conclusion, alcohol research needs to emphasize the study of drinking among older adults, which can vary depending on the cohort and period under study. From a public health perspective, health-predicting models focusing on older adults need to be built on a better understanding of the impact of alcohol and other determinants of health on health status. Main contributions, practical implications, limitations of the thesis and recommendations for future research are presented in the discussion.