The analysis of work by women living as part of a couple is generally based on an assumption of specialisation within the couple. The man is assumed to gain a higher return on the labour market and the woman takes precedence in handling the household's domestic activities. Using this framework, econometric studies have shown the complexity of the decision surrounding whether women work by analysing the link between fertility and labour market participation, the incentive or disincentive effects of the tax policy and the cost of having young children. These findings present a major shortcoming in that they do not take account of the heterogeneity of individual behaviour due to individuals' personal skills and preferences and the way in which these factors are incorporated into their trajectories. The French part of the European Household Panel and new econometric methods are used to refine the longitudinal analysis of the working patterns of women living as part of a couple. This new approach questions the usually accepted influence of certain determinants of female work patterns. If, as expected, the more qualified the women the more they work, then the spouse's income has little effect on the work decision, at least in the short run. It is more his average medium- or long-run wage level that counts. Likewise, although young children temporarily affect women's work patterns, the total number of children seems to give rise to more of a choice between fertility and work over the entire life cycle, even though the costs of caring for the children also count in many women's work decisions.