In this paper three ways of measuring and defining change in cross-sectional designs are presented. The first measure of change concerns structural stability: continuity in the organisation of a set of variables across time. The second is level stability: the comparison of means in different groups differing on age or socio-cultural dimensions. The third is based on the changing parameter model (Firebaugh, 1997). In this model, it is the interaction between ageing effects or socio-cultural factors (period and cohort effects) and another variable that is assessed. These three approaches of change are illustrated on the basis of beliefs about the self, life and society assessed in representative populations of elders in two regions of Switzerland (Geneva and Central Valais) in 1979 ({N} = 1608) and 1994 ({N} = 2101).