The debate of what motivates voters’ attitudes to politics has in a long time had a central place in the international election studies, and is this day still just as relevant. The assumption that self-interest motivates political attitudes has become a widely held belief in the public debate. But the greatest surprise is that the assumption likewise stands unchallenged in extensive parts of the scientific literature, despite the fact that a systematic analysis of the explanation strength of self-interest doesn’t exist (e.g. Pierson 1994, 2001b). The ambition for this master dissertation is to try and make up for this lack of research and is the first known study to systematically analyse, to what extent are Danish voters’ attitudes to the welfare state motivated by short-term narrow self-interest? Our research design differentiates from earlier studies because it focuses on different types of variations. That is variations over time, across policy variations and individual-level variations. We use quantitative methods. The data consist of different Danish questionnaires in cross-sectional design and panel data design with the primarily use of the “Danish Election Study 2005”. In explaining the research question we use a combination of simple and advanced statistics (logistic regression). Attitudes to the welfare state are multidimensional and therefore we have chosen to split the analysis in a general level and a specific level. The dependent variables are therefore both general questions and specific questions to the welfare state. The independent variables are the usual suspects (e.g. gender, age, children living at home, etc.). When we look at variations over time. We find a few indicators on self-interest, but it is not the general trend. One indicator for self-interest is that from 2001 to 2005 general support for more tax cuts is reduced, which probably is due to a tax reduction from 2004. In 2005 we see that voters’ endorsement towards the welfare state is the highest ever seen. We find the same pattern on the voters’ political agenda. This could speak in favour of the self-interest hypothesis. But we don’t find stabile attitudes on taxes and welfare through time, but instead great fluctuations from one year to another. Questions on socio-economic and political questions based on values also dominate the agenda. This instead speaks highly towards the hypothesis that voters’ instead are discourse-dependent and perceived problem-dependent. When we look across variations between public policy areas and responsibilities we don’t find self-interest. Instead a “logic of appropriateness” seems to exist with a consensus that specific public policy areas should get more subsidies from the government than others. There seems to be a pattern that some groups in society are more deserving to public support than others. The most prioritised are the elderly and the sick, subsequently the children and young people, unemployed and last the social security recipients. The variations seems to be explained by what seems fair and just and which groups voters’ can identify with - a dichotomy between “them” and “us” (e.g. Oorschot 2000, 2006). At the individual-level variations we have chosen the most likely public policy areas to find self-interest (social services and benefits). If we do not find indications of self-interest here, it should be hard to find anywhere. On the general level more than 85 pct. of the population, who earns less than 700.000 DKK in the household, do not notably differentiate in attitudes towards the welfare state. It is not before the household income level exceeds 900.000 DKK that self-interest shows its face. On questions on the specific level concerning the elderly we only find few indicators of short-term self-interest. E.g. 18-29 years of age has the lowest support for increased funding for nursing homes and old age pension. But all in all the primary explanation on the elderly questions seems to come from a generation perspective, where the 68’generation shows signs of being the most welfare orientated. When we look at the questions on children and young people we see the clearest indicators for self-interest. The different policy areas like day care centres and education are highly prioritised by the affected groups. It can however be discussed if the results could not also be interpreted in terms of values of care. When we look at the questions on unemployment benefits we do not find clear indicators of self-interest from the unemployed groups. Instead we find what seems to be a cleavage between self-employed and higher non-manual employees versus workers and lower non-manual employees. In other words the perceived risk of becoming unemployed. All in all our results concludes that other explanations than self-interest in much greater extent seems to explain voter’s attitudes to the welfare state which challenge the widely held belief of e.g. Pierson (1994). It is therefore more to do with values and norms of socialisation than of narrow self-interest. Voters’ attitudes to the welfare state are therefore more a question of a “logic of appropriateness” than of a “logic of consequentiality”.