One fundamental condition for a democracy to work is the possibility for voters to hold their legislators to account. Because voter turnout in European Parliament (EP) elections is so low, however, the task of holding MEPs accountable is passed on to the national parties. These become ‘transmission belts’ between voters and legislators: they instruct the MEPs how to vote and present them threats and rewards to ensure they vote as promised. To study how well this works, this thesis aims at exploring what influences how an MEP votes in situations where the national party and the European Party Group (EPG) hold conflicting policy positions. A new analytical framework is built to generate and preliminarily test hypotheses concerning what influences how an MEP votes. The analytical framework is largely inspired by Strøm (1990), who assumes that legislators are driven by their wishes of getting policy through, reaching high office and being re-elected. Normally, these goals can be reached with the help of the legislator’s party, which is why legislators have an incentive to keep their party happy and vote according to its policy position. In the EP, however, things are complicated by the fact that certain goals can only be reached with the help of legislators’ EPGs. Hypotheses concerning what may drive MEPs to vote either with her national party or with her EPG are tested in the thesis with the help of unique interview data, collected from five Swedish MEPs from parties in government. The thesis’ results show that policy-related incentives are most important to MEPs, and that these incentives are mainly controlled by EPGs. Office- and re-election-related incentives are only ascribed limited importance. There are contextual factors that affect these results: if an issue is perceived as important to Sweden or to the national party or if it receives high media attention, MEPs ascribe less importance to EPG-controlled incentives, although this does not mean that national party-controlled incentives increase in importance. In conclusion, MEPs do not seem to feel pressured by their national parties when deciding how to vote. This leads me to the conclusion that the role of national parties as ‘transmission belts’ between voters and MEPs is defective

Cette thèse analyse la socialisation politique de 92,6% des eurodéputés polonais au Parlement européen pendant leur premier mandat entre 2004 et 2009. Son objectif est de comprendre ce processus en détail dans cette institution par définition, supranationale et pro-européenne - et donc de préciser concrètement quels acteurs politiques (première partie) se conforment à quoi, pourquoi, où, en combien de temps, comment, grâce à qui ou quoi (seconde partie) et jusqu’à quel point (troisième partie). L’étude montre que la socialisation politique pendant les cinq premières années n’est pas un processus très puissant car elle provoque seulement l’ajustement aux spécificités formelles et informelles de la vie quotidienne et, possiblement, l’approfondissement plus ou moins léger des attitudes et du comportement de base (dans un sens pro- ou anti- européen, en fonction de l’orientation). Autrement dit, elle transforme les novices en experts mais pas en natifs. Soit elle n’intervient que partiellement et il serait alors préférable de qualifier ce qui se passe réellement tout simplement d’intégration politique et non de socialisation politique soit elle nécessite plus de cinq ans, en commençant par l’acquisition des spécificités europarlementaires, puisque cette étape-ci prend déjà parfois même tout le mandat. / This thesis analyses the political socialization of 92.6% of the Polish Members of the European Parliament during their first mandate from 2004 to 2009. Its aim is to understand in full this process in this institution which is by definition supranational and pro-European and therefore to specify concretely which political actors (first part) comply with what, why, where, how long, how, thanks to whom or what (second part) and to what extent (third part). The study shows that political socialization in the first five years is not a very powerful process as it causes only the adjustment to everyday formal and informal specificities and, possibly, a more or less weak deepening of initial attitudes and behavior (in a pro or anti-European sense, according to the orientation). In other words, the process turns newcomers into experts without making them natives. It either occurs only in part, in which case it would be preferable to characterize what actually happens simply as political integration and not as political socialization, or it takes more than five years, starting with the acquisition of Euro-parliamentary specificities, as this stage sometimes already takes the whole mandate.