The "Ageing and Inequality" exemplar involves simulation analyses of the structure of social inequality as a function of long-running changes in socio-economic and socio-demographic distributions. "Social inequality" can refer to lots of different things, but we analyse it through a variety of measures of the distribution of economic assets, such as income and occupational position.

The substantive contribution of this exemplar is to try to estimate forward in time the likely profile of social inequality if current major social trends continue, and/or if they are modified in certain specific ways. The "ageing population" is an important example of an ongoing social trend: over time, this demographic change is expected to impact upon many important drivers of the structure of social inequality, such as  employment patterns and opportunities; housing ownership and wealth; and patterns of inherited wealth. A pertinent example of a social trend that could be modified concerns the size of the higher education sector – it is instructive to compare a predicted profile of social inequality under a variety of scenarios for the expansion or contraction of the education system.

The graph shows an early example of an output from this examplar – it shows predicted levels of ‘social mobility’ (measured in terms of father-child correlation in occupational position) under two different scenarios (‘a’ and ‘b’, representing no change, or an major increase, in the higher educational sector), and according to models with four different ways of measuring educational attainment (indicated as ‘educ4’, ‘educ2a’, ‘educ2b’ and ‘isced’).

There are quite a few simulation analysis projects which attempt similar applications concerned with social structure and inequality. However what makes the ageing and inequality application relevant to the infrastructral provisions of NeISS is that there are many different plausible permutations to the models that might be used, and accordingly it is interesting to try out similar simulation models lots of times over with minor variations in the specification. Permutations involve things like the variable definitions used for key measures, the underlying data, and the simulation rules. The support that NeISS offers for modifying and re-running analyses (and documenting each stage) is ideal for this style of research.

The tool for specifying simulation job runs is not yet built, but a broad illustration of the sort of scenarios involved can be summarised as: