The study analyse how welfare regimes influence public support for welfare state policy. Regardless whether people "get what they want" or "want what they get", scholars expect a strong connection between welfare regimes and cross-national differences in support for welfare state policy. Until recently such regime differences in public attitudes were taken for granted. However, in the micro-level foundation of welfare state theory, little attention has been paid to this question, partly because of inadequate data. Comparative survey analyses have exploited the modules on social inequality (1987, 1992 and 1999) and on the role of government(1985, 1990 and 1996) conducted within the International Social Survey Program (ISSP). Others have used data from the International Social Justice Project (ISJP)(1991), the World Value Studies (1981, 1990, 1995 and 1999) and a couple of Eurobarometer surveys. The empirical results have been confusing - not to say depressing. Some studies do not find the expected pattern at all. Thus Gelissen (2000:298), and Bean & Papadakis (1998:30, 231) reject the assumption of regime clusters in attitudes altogether. Most studies, however, using alternative dependent variables, or limiting the analyses to welfare states that comes close to the regime ideal types, do find the expected clusters but fail to confirm expected variations between classes, or between public and private employees. That is, they do not confirm the expected processes of opinion formation that should produce the variations at the aggregate level. We therefore devote this project to the search for the missing link between regimes and policy attitudes, which we expect to find somewhere else. Previous studies tend simply to conclude that the regime theory is wrong without explaining 1) why it fails or 2) how we can account for the observed cross-national differences. As to the former question one could suggest that 1) the feedback from class mobilisation has lost importance; 2) short-term self-interest has lost importance via-a-vis values; 3) more reflexive people could impede the simple reproduction of justice principle. Still, this does not provide any explanation of cross-national differences. Inspired by early welfare state literature we suggest that cross-national differences in perception of poor and unemployed could be the (intervening) variable that explains the connection between welfare regime and aggregate variations in attitudes. At this point, there are problems with interpretation of data in previous studies. "Provide job for all", "provide basic income" and "redistribution" primarily measure attitudes to policies targeted to low-income workers, not to the overall welfare state. These attitudes may coincide in Liberal regimes but not in corporatist or universal regimes. Coughlin (1980) concluded that the public is most in favour of redistribution to old-people, followed by sick and disabled, needy families with children, unemployed and finally people on social assistance (1980; see also Petterson 1995). Thus, a pro-welfare Scandinavian might agree that it is not the responsibility of the government to provide job for all. With a point of departure in general regime theory the project will thus seek to develop and test an underlying micro-level theory that explain public attitudes to different aspects of the welfare state. First, in regime theory, it is a classic that means testing leads to a negative perception of receivers (themversus us logic), but this is rarely specified or tested empirically. Inspired my micro-level theory of deservingness criteria (Cook, 1979; De Swann, 1988; Will, 1993; Oorschot 2000) and the single cross-national study that has been conducted (Oorschot & Halman 2000) we will try to specify how different regimes can generate different perception of poor and unemployed. The project builds on available cross-national data sets and a new national sample. The ISSP 1992 and 1999 survey on social inequality and the Eurobarometer survey on poverty from 2001 (56.1) will be central but data will also be taken from the other ISSP surveys, the World Value Study and a number of national studies.