As discussed earlier, the three countries selected for the study were chosen as illustrations of different types of welfare state regimes, and because of contrasting attitudes and policies towards the employment of disabled people. A variety of different research methods was used, in a strategy of ‘triangulation’ as advocated by Denzin (1978), and in a comparative context by Simonin (1996). These included secondary analysis of survey data, reviews of published material, and discussions with national experts on policy.

Secondary analysis of survey data

Secondary analysis of data provides a cost-effective strategy for obtaining a broad overview of the outcomes of policy despite its well-known limitations in conceptualising the experience of disabled people (Abberley, 1991, 1992). It was decided to use three large data sets in order to analyse the employment, living standards and social and political participation of disabled people across the three countries and relative to non-disabled people in each country. These are the German Socio-Economic Panel, the British Household Panel Survey and the Swedish Level of Living Survey. The surveys selected were each panel studies, but they have been used cross-sectionally to provide an analysis for 1991. This year was chosen as it is the only one for which data is available for all three surveys. These data sets were chosen in part because of the absence of a suitable harmonised single data set for comparative analysis. The Luxembourg Income Study was considered as a data source, but it proved unsuitable, because a disability indicator was not available for all the countries, and because of a lack of information on economic and social participation, as opposed to income data. The three surveys are, of course, not comparable in the same way as a harmonised data set. Fundamentally, each survey uses a different definition of disability. Definitions of basic indicators such as income, household composition, and employment are all different. More culturally specific concepts such as well-being, health, and social and political life, show, if anything, even more divergence. However, the different definitions are described in order to achieve transparency about what is being compared, and they have functional equivalence in that they are able to provide comparisons of the situation of disabled relative to non-disabled people within each country (see Figure 1).