ABSTRACT
Adopting this broad definition encompasses several different research designs and, accordingly, several different approaches to statistical modelling. Each design and the issues it poses for statistical modelling are considered in turn. However, note that the definition of longitudinal research signifies a time frame over which data is collected and referenced. We can conceptualise events in time in two ways. First we can observe the states subjects are in at particular points in the time period, for example, each year we might record whether a person is married or not, employed or not and so on. Following on from this we can monitor transitions between states, that is, moving from being employed to non-employed, for example. Alternatively, or in addition, we can record durations, that is, how long a person is married, employed and so on. Statistical models for durations are considered in Chapter 13.
