ABSTRACT
There is no doubt about the heuristic potential of either prospective or retrospective longitudinal data. Indeed, such data make it possible to:
• analyse the duration of social phenomena; • highlight differences or changes, between one period and another, in
the values of one or more variables; • identify sleeper effects, that is, connections between events and transitions
that are widely separated in time because they took place in very different periods, as in the relation between childhood, adulthood and old age (Elder, 1985; Hakim, 1987). For example, the experience of old age has much to do with hardship in the adult years and one’s responses to it: the same event or transition followed by different adaptations can lead to very different trajectories (Elder and Liker, 1982; Negri, 1990). Also caring at a young age has a significant effect on earnings and risk of poverty in later life as young carers are often absent from school and fail to gain even the basic qualifications (Olsen, 1996; Payne, 2001);
• describe subjects’ intra-individual and inter-individual changes over time and monitor the magnitude and patterns of these changes;
• explain the changes in terms of certain other characteristics (these characteristics can be stable, such as gender, or unstable, that is, timevarying, such as income) (van der Kamp and Bijleveld, 1998: 3).
