This project aims to lay the groundwork for a deeper understanding of the relation between voters and elected representatives in the age of social media. Social media platforms like Twitter carry the potential to transform democracies as they allow for direct contact between politicians and voters and thus for more personalized contact between representatives and represented. Academic work on this topic, however, has so far remained highly descriptive and has focussed on the frequency or volume of social media usage by politicians and not so much on the content of their messages. As a result, we lack knowledge about the information politicians portray in social media activity, as well as what the effect of this communication is, e.g. how voters perceive it. The two studies presented in this proposal intend to target this important research gap and to analyse the micro-processes of the content of online communication and how voters perceive it. Study 1 (‘What Do Politicians Tweet When? New Quantitative Insight Into the Strategic Use of Personality Traits and Policy Positions’) is an observational study that investigates the tweets of a sample of Swiss and German politicians with the help of machine-learning to find out more about the motivations of politicians to tweet. We focus on Twitter as direct and individualized behaviour is best captured with this platform. Following key ideas in the representation and personalisation literature, we focus specifically on the distinction between personal or policy focussed content and investigate whether the content of the tweets depends on variables such as the incentive to cultivate a personal vote, professionalization of parties and parliament, or career stage in comparative perspective - information that is already available from the SNF project on “Parliamentary careers in comparison” . Study 2 (‘What do voters want: Policy or Person’) is a two country survey including an experiment that investigates how voters use social media for political purposes and which information they seek in these platforms. A short survey experiment focuses on how feelings of representation and voting intentions of voters are affected by the content of the twitter message that politicians share via social media and whether voters rather appreciate personal appeals about the politicians or information about policy positions and activities communicated via Twitter. Using an innovative survey design, we first extract the demographic personal profile and policy preferences of the participants. We then use this information to present subjects with candidate profiles and Tweets from politicians that are (dis)similar to them on these key dimensions. As such we contribute to the question whether voters appreciate politicians similar or dissimilar to their own personal profile (desire for descriptive representation) and feel better represented by politicians who are like them or who act on their behalf (desire for substantive representation). The conduct of these studies in Switzerland and Germany allows for a first insight on variation on relevant influencing variables such as the election system, the importance of parties and the professionalization of the parliamentary system. Importantly, this project forms the base for a larger research agenda that we aim to develop in the next years and where we want to further explore the potential of social media in fostering closer linkages between voters and politicians.