Operationalization of the New Class Schema

Abstract

We argued in Chapter 2 that our choice of Britain, Germany, Sweden and Switzerland would provide us with a broad range of different institutional settings in the labour market, the welfare state and industrial relations. Besides theoretical concerns, this selection of countries is also motivated by more practical matters of data availability. For these four nations, individual datasets are available that are both sufficiently large and include detailed information about employment, income and education. For Britain, data is taken from the British Household Panel Survey (BHPS), year 1999, for Germany from the German Socio-Economic Panel (GSOEP), year 2000, for Sweden from the Level-of-Living Survey (LNU), year 2000, and for Switzerland from the Swiss Household Panel (SHP), year 1999. If not otherwise stated, we will work with the cross-sectional weight variables included in the surveys to improve the representativeness of the British, German and Swiss samples. There has been no weighting of the Swedish LNU. For the British and German data, weighting is almost inevitable as the two data sets are not devised to be representative for the entire population but over-represent different subpopulations: Wales and Scotland in the case of the BHPS, foreign citizens in the case of the GSOEP. With respect to Switzerland, weighting is needed to adjust for the heavy under-representation of foreign nationals in the sample.