An analysis of survey data on membership, pillarisation, voting behaviour and attitudes. Belgian society is traditionally portrayed as heavily pillarised, i.e. having a system of exclusive linkages between voluntary associations and political parties, resulting in the formation of a catholic, a socialist and a liberal 'pillar' within society. Recently, several authors have questioned the validity oft his model. Our survey of the Flemish population, however, shows that pillarisation is an enduring feature of Flemish society. Membership of voluntary associations, trade unions and health insurance organisations remains ideologically motivated, and shows a high degree ofconsistency in this respect. Integration into a 'pillar' exerts a strong influence on voting behaviour, although this effect weakens in younger generations. Pillars also have significant, but weaker effects on attitudes like individualism, trust and solidarity. In the social capital research tradition, these differential effects of membership are often neglected. Although there are signs that pillarisation weakens in Fiemish society, the system certainly has not disappeared.

The 1995 simultaneous election of three legislative assemblies (Senate, lower Chamber and regional councils) offered the Belgian voters an opportunity to split their ballot between three different parties. An analysis of Flemish individual level survey data shows that 76.1% cast a straight ticket vote, white 20.9 % split their tickets between two and 3 % between three different parties. Ticketsplitting occurs most frequently amongst voters who mention the personality of individual politicians or the issues as a reason to support a party. In addition, the likehood of ticket-splitting increases amongst the higher educated and the non-partisans. No support was found for the hypothesis that ticket-splitters in a multi-party system tend to lean towards one party on some issues and towards the other party on others. Instead, ticket-splitters generally take a position in between two parties. No evidence was found of a differential issue impact across elections, in this sense that regional issues are more critical to regional elections and federal issues to federal elections.

During the postwar period, the Flemish press-scene changed fundamentally. Alongside further commercialization and concentration, a process of structural depoliticization or depillarization took place: (financial) links betweenparties and trade unions on the one hand and newspapers on the other disappeared. This article examines the impact ofthese structural transformations on the newspapers' content. We emphasize marks of (de)pillarization in Flemish newspapers during cabinet formations. In 1958, the press took undeniably sides in the battle between the pillars: information about the formation of the new cabinet formed the background for these fights. In 1981 most attention went to the cabinet formation itself. The pillarization ofthe content was however on a more latent level not neglectable. Compared to 1958 and 1981 the old alliances between press and ideological institutions were far less visible in the content of 1995's newspapers. Apparently the depillarization ofthe Flemish press-content is an ongoing, longlasting process.

The ultimate factor which determines the election offemale MP's (23%) in the 7 Belgian elections of 1999 is the position which women are granted on the list ofcandidates of a political party. 74% of the women won their seat through a "favourable place" on the list. The floating of votes and hence of seats between parties are accountable for 23% of the directly elected women. Only in the case were women are great vote- attractors they can swift the useful order of the list: however their number is extremely limited, only 3%. The succession where effective candidates are being replaced accounts for 15% of the final amount of female MP's. As important is the combination of double mandates and cooptation.

For many journalists, politicians and political scientists ministerial cabinets in Belgium equal political power. Moreover they argue that this power is, if not illegitimate, at least problematic, and so ministerial cabinets have become one of the most criticised institutions in the Belgian political system. Yet, the lack of empirical data on this controversial topic is striking, certainly when compared to the vast academic attention given to other political agents. There is an urgent need for empirical substance to the debate. This incited us to set up an extensive political-sociological study on the Belgian and Flemish ministerial cabinets. In this article we present the first stage of our study: a classic insight in the (socio-demographic) composition of ministerial cabinets and - to a lesser extent - in the work environment offered by ministerial cabinets of the former government (1995-1999). A first descriptive analysis seems to underpin the commonly held idea of cabinets as networking, loyal, and flexible brain trusts.

This article argues that both hard and soft institutions shape the strategies and - to a lesser extent - the preferences of political actors in the European Union. First of all, it discusses the institutional perspective in political science and presents an institutional model of decision-making. Secondly the institutional argument is illustrated by a detailed account of the decision-makingprocess with respect to the Packaging Directive. An analysis of the actor configuration, the interactions between the involved actors and an overview of the decision-making process itself, all show to what extent actor's strategies and preferences are constrained and empowerd by the European institutional context.