Vriendjespolitiek.net: research into post-demographics

On: June 11, 2008
Print Friendly, PDF & Email
About Erik Borra
Erik Borra is assistant professor in Journalism and New Media at the University of Amsterdam.


Since 1998, and on paper since 1989 (Stemwijzer 2008), general elections in the Netherlands have spawned a variety of so called voting recommendation machines. These systems typically ask the user to answer some questions after which they offer the user a voting recommendation, based on the compatibility between his or her answers and the political parties. The questions are either based on the political party’s programs (Kieskompas 2008), or on its actual voting behavior in Parliament over the past few years (Politix 2008).

Since the birth of online social networking sites lots of people have, quite unconsciously, put their likes and dislikes on public display. They not only show with whom they affiliate, but also what kind of music, movies, food, or even brands they prefer.

We have developed a post-demographic recommendation tool derived from digital life software systems, while at the same time addressing them – based on the aggregated profiles of pals of political party leaders as they appear on the biggest Dutch online social network, Hyves. By providing appropriate visualizations we show both the demographics and the relations of a group of pals, and replicate the existing, arguably anti-participatory democratic, voting recommendation machines. Ultimately the goal is to raise awareness of one’s digital public self – one’s data body (Well.com 1995) – to create conscience of simple yet powerful profiling techniques, and the tools of the surveillance and control society. (Bogard 1996, Deleuze 1992). We intentionally chose to highlight the entertaining quality and lightness of peer-based behavior this society is so immersed in. In addition, this paper introduces and explains the term post-demographics (first coined by Richard Rogers in August 2007) in the context of control society.

An online social network is a web service where people form communities and share interests and activities or explore the interests and activities of others. Every day millions of people use all kinds of social networking websites on a regular basis; social networking seems to have become part of everyday life. Users can make a profile to introduce themselves and make ‘friends’-connections with other users. Most social networking sites are open/public to all web users and are designed to attract advertisers, who can customize adds according to a user’s preferences.

Online social network sites are the newest version of mediated publics; they are environments where people can gather publicly through mediating technology. In a certain way mediated publics are similar to unmediated publics (e.g. parks, malls, the sea shore). Dana Boyd (2007) however stated that four significant features distinguish mediated from unmediated publics: persistence, searchability, replicability, and invisible audiences. As the mediated data are archived in databases and made public on the web, they stay traceable for a long time – the submitted data are persistent. The Critical Art Ensemble described this (often automatic) creation of one’s digital public self as the Data Body (Well.com 1995), a virtual body composed of the files related to an individual. Because one’s digital representation is stored in a database with an associated index, it might be retrieved by a simple web search. Digital data are also easily replicable and tampered with, so you never know what is original. And finally, because lots of data are public, you never know who is reading along. “In mediated publics, not only are lurkers invisible, but persistence, searchability, and replicability introduce audiences that were never present at the time when the expression was created.” (Boyd 2007)

No wonder that online social network sites change interaction [between teenagers], as well as their representation to the outside world (Boyd 2007). Research however (Boyd 2007) made apparent that most users of online social networks are not aware of the various concerns related to them. When concerns do appear in the news, they focus in most cases on the invisible audiences, i.e. privacy (e.g. sexual predators, identity theft, marketing, police investigations, job rejection, governmental control) (Albrechtslund 2008).

In order to understand why these concerns should be addressed we must take a look at a short history of profiling. For the 1890 U.S. Census, Herman Hollerith built a tabulating machine, to automatically crystallize information about every single U.S. citizen. This aggregation and automatic counting of statistical data was quickly adopted for research, business marketing, and planning purposes. The value of demographic relationships was later increased when information about commodities and services was integrated or cross-referenced, by more advanced computers, e.g. with consumer or sales data; consumers could then be sorted into taste groups, ‘clusters’, or ‘profiles’ (Elmer 2004). This discriminatory practice allowed for more specific customer or group targeting – following the shift to a post-Fordist, mass customization and just in time economy. With the networking of demographic databases the limitations of timeliness could be overcome, and real-time, or continuous, tracking and inventory applications were developed, allowing for ingenious feedback loops. This rationalization of relationships between bits of demographic, psychographic and consumer behavioral data can provide condensed pictures or profiles of particular groups and places. This is also called dataveillance, “the systematic use of personal data systems in the investigating and monitoring of the actions or communications of one or more persons” (Clarke 1988). Moreover, the automatic grouping and data mining can lead to population characteristics and relations which were not known, or did not exist, before. These derived and previously non-existent characteristics and relations are post-demographic, as demographic data refer to factual population characteristics. With Vriendjespolitiek.net (1) the post-demographic element becomes apparent when relating users’ preferences to a voting recommendation, thereby attributing political value to his or her preferences. “Information systems increasingly place individual wants and desires into larger, rationalized and easily diagnosable profiles (demographics and psychodemographics). Surveillance in this light cannot be removed from notions of social control or from the potential for certain planned effects.” (Elmer 2004, p23)

Talking about surveillance implies Foucault coming into the picture. Foucault noticed that subjects in populations are involved in political power relations. He discovered connections between institutional (re)organizations and scientific disciplines. According to Foucault, sciences are predominantly organized and shaped by practices, or bodies of text – which he coined discourses. For the 18th and 19th centuries Foucault used a discussion of Bentham’s panopticon to describe these connections between (social) science and institutional organization . “The discourse/practice of the panopticon was a condition for a new form of biopower, a means of controlling masses of people for the development of industrial processes” (Poster 1990). As in Bentham’s panopticon, Foucault described surveillance as a classificatory architecture, an archive in which individuals or bodies are separated and classified by way of files. “[T]he surveillance is based on permanent registration”, at the beginning of the ‘lock-up’ a document for each inhabitant of the panopticon (the prisoner) is made which bears “the name, age, sex of everyone, notwithstanding his condition” (Foucault 1975). Without the recording and storing of the subject’s behavior, surveillance is incomplete. The emerging social sciences thus supplied administrations with the knowledge of record keeping and its evaluation. This aggregated demographic and psychographic data allowed for the control of entire populations through biopower: the “explosion of numerous and diverse techniques for achieving the subjugations of bodies and the control of populations” (Foucault 1976). In Foucault’s so called disciplinary societies,the panopticon was the ideal project of an environment of enclosure in which a body – described by demographic and psychographic characteristics, could be subjugated to power, or molded to a category or norm, by the techniques of biopower. This then gave rise to biopolitics; that mode of organizing, managing, and above all regulating the population, considered as a biological species entity (Thacker 2005).

Deleuze (1992) argued that Foucault’s concept of the disciplinary society was overtaken by the society of control. He conceptualized the way in which modes of data accumulation, storage, and processing are networked into an increasingly dispersed and automated infoscape. These modes with immediate feedback loops boil down to a modulation of coded figures, rather than to the fitting of a body into a mold. Galloway (2004), linking Deleuze to Foucault’s biopolitics, states that control societies deal with the way life may be analyzed and controlled from a distance. Whereas the context of Foucault’s biopolitics are the practices of statistics, demographics, and population control in the 18th and 19th centuries, his theories may also be viewed in the light of post-demographics. In control societies the body becomes obsolete, the molding of the deviant in disciplinary societies could only be enforced by the existence of fixed categories (or norms derived from the ‘normal’ in the category). With the derivation of post-demographic properties from dynamic aggregations the coded figure can no longer be molded to the norm – the norm has ceased to exist, but is constantly tweaked and modulated – granted or denied access. The development of computer-matching or -profiling techniques attempting to attribute general characteristics to individuals for the purpose of discriminating them should therefore be questioned, especially if users of online social networks are proven to be constantly lured into giving more personal information, which is then persistently stored, indexed and publicly searchable.

Lyon (1994) claims that users now “’trigger their own inclusion into systems of surveillance”. They voluntarily put their information on public display and are constantly lured into providing even more information. Although various authors argue that this online exhibitionism might actually empower users (Chun 2006, Albrechtslund 2008) it also makes the profiling of public data possible. Profiling or establishing a category means, however, that sorting individuals and groups is inherently discriminatory.

Vriendjespolitiek.net intents to raise awareness of this potential to categorize and dividuate; as well as the possible conclusions which can be drawn from it. A crawler aggregated the profiles of the pals of all Dutch political party leaders. Data mining techniques and appropriate visualizations were used to match any Hyves user with a political leader, based on the profiles of the party leader’s pals. In a democracy, people’s representatives are elected. In Vriendjespolitiek.net we considered links with a political party leader to be votes for that same person: his ‘pals’ are his electorate. The small world model has it that ‘birds of a feather flock together’; people whose profiles show similar preferences are likely to be pals (Krebs 2005). Our exploratory research indicates that this is also valid for pals of Dutch political party leaders. For instance, Bible readers are likely to be pals of Andre Rouvoet, a deeply devoted Christian political party leader. Data mining the profiles of the political party leaders’ pals, labels their aggregated, combined, and analyzed preferences, which subsequently can be used to provide voting recommendations for a person who is not explicitly known to be a pal of a political party leader. A voting recommendation system based on social network profiles is born.

Since 1998, and on paper since 1989 (Stemwijzer 2008), general elections in the Netherlands have spawned a variety of so called voting recommendation machines. These systems typically ask the user to answer some questions after which they offer the user a voting recommendation, based on the compatibility between his or her answers and the political parties. The questions are either based on the political party’s programs (Kieskompas 2008), or on its actual voting behavior in Parliament over the past few years (Politix 2008).

We thought a voting recommendation system based on social network profiles to be a very good tool to address the concerns of preference exhibitionism (entailing data persistence, searchability, authenticity questions, and lurchers), because in a democracy the ballot should be private and secret. By relating apparently non-political preferences to political forecasts the users of Vriendjespolitiek.net realize that their voluntarily displayed data may start a life of their own, spreading around conclusions they’d rather not have on public display. The power of profiling is visualized by clustering the users’ preferences to politicians. Searching a pal, or comparing a politician to his own ‘electorate’ becomes possible: the user becomes a lurcher. By tweaking one’s preferences to fit a certain ‘political profile’, authenticity cannot be taken for granted any more. Or the other way around: what if politicians tweak their messages onto certain groups? And even: representative democracy or rather a democracy of representation? “It is time to put the demo back into democracy.” (Mayfield 2005).

Although since the presentation we experienced some down time, our system is back online. Go check it out at Vriendjespolitiek.net. You can try for example ‘femkehalsema’, the green party leader, and see that she is compatible with her electorate. When you try ‘jpb’, the nickname of Jan-Peter Balkenende – our prime minister, you will find that he is more compatible with the electorate of Andre Rouvoet. The system is currently in Dutch, as it is targeted at users of the Dutch social networking platform Hyves. You might also take a look at our English paper, detailing the information visualization.


“In a networked society each individual’s data has value.” (Ito 2005)

“We leave data everywhere we go. It’s not just our bank accounts and stock portfolios, or our itemized bills, listing every credit card purchase and telephone call we make. It’s automatic road-toll collection systems, supermarket affinity cards, ATMs and so on. It’s also our lives. Our love letters and friendly chat. Our personal e-mails and SMS messages. Our business plans, strategies and offhand conversations. Our political leanings and positions. And this is just the data we interact with. We all have shadow selves living in the data banks of hundreds of corporations’ information brokers — information about us that is both surprisingly personal and uncannily complete — except for the errors that you can neither see nor correct. What happens to our data happens to ourselves.” (Schneier 2008)


1) ‘Vriendjespolitiek’ might be translated as ‘Politics and Pals’.


  • Albrechtslund, Anders. Online Social Networking as Participatory Surveillance. First Monday, 2008.
    Bogard, William. The Simulation of Surveillance:Hypercontrol in Telematic Societies. Cambridge University Press, 1996.
  • Boyd, Danah. Social Network Sites: Public, Private, or What? Knowledge Tree, 2007.
  • Chun, Wendy. Control and Freedom: Power and Paranoia in the Age of Fiber Optics. MIT Press, 2006.
  • Clarke, Roger A. Information Technology and Dataveillance, Commun. ACM 31, 5, May 1988, 498-512.
  • Deleuze, Gilles. “Postscript on the Societies of Control”, from _OCTOBER_ 59, Winter 1992, MIT Press, Cambridge, MA.
  • Elmer, Greg. Profiling Machines: Mapping the Personal Information Economy. MIT Press, Cambridge, MA, 2004.
  • Foucault, Michel. Discipline & Punish: The Birth of the Prison. Vintage Books, 1975.
  • Foucault, Michel. Histoire de la sexualité, Vol I: La Volonté de savoir. Gallimard, 1976.
  • Galloway, Alexander. Protocol, How Control Exists After Decentralization. MIT Press, 2004.
  • Groot, L.F.M. Een kritische evaluatie van de StemWijzer 2002. Amsterdam: SISWO, 30 January 2003.
  • Ito, Joi. Emergent Democracy, in Extreme Democracy, 2005. http://www.extremedemocracy.com/chapters/Chapter%20One-Ito.pdf, 25 May 2005.
  • Kieskompas, http://kieskompas.nl/, May 25, 2008.
  • Krebs, Valdis. It’s the Conversations, Stupid! The Link between Social Interaction and Political Choice, in Extreme Democracy 2005. http://www.extremedemocracy.com/chapters/Chapter%20Nine-Krebs.pdf
  • Lyon, D. The Electronic Eye: The Rise of Surveillance Society. University of Minnesota Press, 1994.
  • Mayfield, Ross. Social Network Dynamics and Participatory Politics, in Extreme Democracy, 2005,
    http://www.extremedemocracy.com/chapters/Chapter%20Ten-Mayfield.pdf, 25 May 2008.
  • Politix, http://www.politix.nl/nieuwekieswijzer.php, 25 May 2008.
  • Poster, Mark. The Mode of Information. University of Chicago Press. Chicago, 1990.
  • Schneier, Bruce. Our Data, Our Selves. http://www.wired.com/politics/security/commentary/securitymatters/2008/05/securitymatters_0515, 15 May 2008.
  • Stemwijzer, http://www.publiek-politiek.nl/Thema-s/Nationaal/Projecten/StemWijzer, May 25 2008.
  • Thacker, Eugene. The Global Genome: Biotechnology, Politics, and Culture. MIT Press Ltd (USA), 2005.
  • Well.com, November 14 1995, Critical Art Ensemble, Utopian Promises, http://www.well.com/user/hlr/texts/utopiancrit.html, 25 May 2005.

Cross posted at my blog.

Comments are closed.