Differences in mortality between social categories are less marked for women than for men. An individual's mortality risk also varies according to their family situation and labour market status, which are correlated to social category. Once these two factors are taken into account, differences according to professional position disappear for women, whereas they remain, albeit to a lesser extent, for men. Level of education remains an important factor, but while it is predominant in the explanation of women's mortality, it plays in combination with the social status in the case of men. When analysing social differentials of women's mortality, it is more appropriate to take into account the level of education, rather than only the social category. The lesser variability of the women's mortality risk is not only observed between social categories. Unemployment and labour market inactivity not due to retirement are thus associated with over-mortality for both women and men, albeit to a lesser extent for women. Major life events (for example, widowhood) also have a smaller impact on the women's mortality. The Permanent Demographic Sample Survey (EDP) allows to study the annual mortality risks for men and women over three periods, ranging from the start of the 1980s to the mid-1990s, taking into account the socio-demographic characteristics of the individuals (age, sex, social category and level of education), their labour market status (employment, unemployment, retirement and labour market inactivity not linked to retirement) and their family situation. Additional factors were brought to the study of mortality risks for the most recent period: length of time in unemployment and the working hours for employed individuals, whether the individuals have a common-law or legal marriage and the number of children they have. The influence of a partner's characteristics was also taken into consideration for those people living with their partner.