Rethinking Representation through the Eyes of Big Data-based Microtargeting

Representation, according to political scientist Pitkin (1967: 8), aims to "make something absent present (again)". However, it is not self-evident how processes of representation produce presence and how the present relate to the absent. While Pitkin assumed that the people, interests and concernts to be represented is a known quantity which can be transmitted in the form of information flows and aggregated in the form of preferences, recent representation research shows that the representation concept itself is contradictory and problematic. If Rosanvallon (2006) is correct to assume that democratic elections and the elected representatives cannot actually reflect the complexity of the electorate, the question arises as to how the relationship between the represented and representatives can be taken more appropriately. Seward (2006) has proposed replacing the traditional assumption of a mapping of the electorate with the concept of a "representative claim" that is produced as part of the process of representation. The interaction between the people represented and their representatives thus does consist in a unidirectional flow of information but presents itself as a creative and performative relationship. The roles of the actors involved are creative because claims about the concerns, interests and priorities of the voters must be actively produced: "The 'interests' of a constituency have to be 'read in' more than ‘read off’; it is an active, creative process, not the passive process of receiving political figures, parties, lobby groups, social ensues "(Saward 2006: 310). Especially the US has been experimenting in recent years with the use of big data in election campaigns. The public narratives of election campaigns since 2008 have emphasized the transformative nature of big data-assisted procedures. The intuition of experienced campaign leaders has been replaced with real-time experiments conducted by behavioral analysts testing different statements, messages and presentations (Bimber 2014). Large datasets enable the "micro-targeting" of individuals and their behavior. Using learning algorithms, the future electoral behavior is expected to be more predictable and controllable. Microtargeting aims to know the voters in detail and understand their behavior; if possible better than the voters themselves. On the basis of constantly updated databases, election campaigns aim to mobilize the individual voters’ political emotions, help them to locate themselves in the political space and identify with the respective candidates. Microtargeting, as one might say provocatively, makes voters and their candidates move closer together and their communication more intimate. Without ever having exchanged a single word (Tufekci 2014), an allegedly intimate, albeit highly asymmetrical, relationship develops between them. With big data and microtargeting, the impossibility of proper representation in democracy seems to have found a new technocratic approach that strives to close the gap between the voters and their representative. As my paper will show, the political technique of microtargeting is situated between data analysis and manipulation. The open question is whether and to what extent the voters accept and support these new representative claims.