/ 25 August 2005

Surfing our way to segregation?

Geography isn’t history, yet. Far from ushering in the death of distance, the internet is making us more anxious to live alongside like-minded people, reinforcing rather than reducing social divides.

This theory appears in a study of websites that publish information about our neighbourhoods. The Joseph Rowntree Foundation, a UK social policy charity, warns that sites providing such data may promote ”online marginalisation” — a virtual segregation of deprived areas leading to real segregation on the ground.

Local information sites were some of the web’s most exciting innovations. The Joseph Rowntree study calls them Ibnis, for internet-based neighbourhood information systems. The foundation found at least 33 different Ibnis sites in the English speaking world. They range from the trailblazing Up My Street (www.upmystreet.com), which dates from 1998, to the British government’s official ”geodemographic” site (www.neighbourhood.statistics.gov.uk), to the highly subjective Crap Towns (www.idler.co.uk/crap/)

These sites draw on official and unofficial data such as crime statistics and house prices to build location-specific pictures of what life is like in a particular postcode or town. Some also draw conclusions about the sort of people likely to live there, so-called ”geodemographic data”.

Geodemographics has long been available to marketing specialists and scientists. In the United States, the science of assigning characteristics to Zip codes goes back to the 1970s. Credit scoring agencies use sophisticated geodemographic databases such as Acorn and Mosaic to score people according to where they live.

This apparently blunt instrument works because people with similar demographics, lifestyles and values tend to cluster together. ”From the point of view of marketing professionals, knowledge of where someone lives is a particularly powerful predictor of all manner of consumption practices, values, tastes, preferences and so on,” the Rowntree authors say.

Most commentators treat the availability of such information as good news. The Joseph Rowntree report is ambivalent, setting out optimistic and pessimistic scenarios. The optimistic prospect is that more information will help create better engagement, whether through house-buying or local politics. The pessimistic prospect is that it will lead to more informed social sorting.

The Rowntree authors worry that neighbourhood information websites will bring this specialist language into the public domain and that people will start taking it seriously. An area’s Mosaic classification may soon rank with house prices and Ofsted reports as a subject of dinner-party conversation — for the web-savvy middle class, at least. So far, most debate about geodemographics is about the consequence of being unfairly misrepresented — the nightmare of unwittingly moving into a ”red-lined” address with a terrible credit history. The Rowntree report warns that even when data is accurate it is dangerous, leading to ”ongoing processes of inter-neighbourhood segregation and intra-neighbourhood homogenisation”.

Such sorting is further promoted as more data about schools’ performance becomes electronically available. In theory, the government already knows the postcodes of all schoolchildren. It would not be difficult, the Rowntree authors say, to post the Acorn or Mosaic data of children who attend a particular school along with its league table scores.

”The dilemma … is whether it is considered acceptable that consumers could look up a school in order to discover the kind of pupils that go there.”

It is on such issues that the ambivalence about online information is at its most acute: ”there is a real awareness of the potential social risks involved in making particular types of information widely available”.

The study found that geodemographic sites receive surprisingly few complaints — in one case only 27 from 175 000 visitors. (In the US, communities are more vociferous about being ”informationally misrepresented”.) The Rowntree study warns that UK geodemographic sites do not seem to be set up to handle large volumes of complaints. It suggests that sites find ways to allow ”bottom up” contributions.

The study is the latest of many alarms that have been sounded about the consequences of posting on the web data that was previously available to specialists: hospital death rates, for example. Theories about the internet’s polarising effect are not new, either. Cass Sunstein, a Chicago professor of jurisprudence, warned in his 2001 book Republic.com of the risk to democracy from the web creating communities of individuals engaging only with others who think the same way.

The Rowntree researchers accept that they cannot reverse the trend towards more data being publicly available. However they hold out another outcome — that the sorting of information by neighbourhoods is itself a passing phase.

Thanks to the march of technology and the widespread use of devices that leave electronic traces about individuals, geodemographics will become unnecessary. The consequences of that trend for social policy will keep researchers busy for a while. – Guardian Unlimited Â