This chapter describes and discusses the design of the study, the data and the methods used for the empirical analysis in the following chapters. I begin with a delineation of the case selection strategy that allows me to test the propositions from Chapter 2. Following this, I describe the measures, concepts and data that are used to analyse the policy of the social democratic parties under review and the labour market reforms implemented in terms of the recommodification framework outlined in the previous chapter. Next, I present the statistical methods and the data sources applied to study the postulated electoral effects.

Case selection and strategy of comparison

The empirical analysis of the electoral consequences of Third Way welfare state reforms in the following chapters is based on a comparative analysis of four Western European countries. The rationale of the case selection is to choose cases that make it possible to test the propositions from Chapter 2 empirically. I apply theory testing case studies to examine causal claims (George & Bennett 2005: 75), and the recommended intentional selection is based on the research objective and strategy (King et al. 1994: 139f ). More concretely, selecting typical cases on the variables of interest is the suggested method for this type of study (Gerring 2007: 91ff ).

The case selection comprised three stages. First, I had to choose a country in which a large social democratic party did not undergo an encompassing Third Way transformation during the 1990s, unlike many other social democratic parties at the time. This was the case in Sweden, where the reforms were more moderate and often temporary (e.g. Lindbom 2001). Sweden is used as a ‘control case’ where the key explanatory concept – the welfare state reforms under Third Way agendas – takes a different value (King et al. 1994: 200ff ); that is, the reforms are more moderate and do not represent an outright break with existing policy principles. Another reason is that Swedish social democracy has always advocated active labour market policies and did not have to introduce these policies in a way that conflicted with previous policy principles. The choice of Sweden also controls for the decline of class voting as a general explanation for the electoral decline of social democracy.