Web politics: The honeymoon is over

Spam and astroturf are souring the dream

Parallel moves in Canada and the US may signal the end of the honeymoon for web-based political campaigning - or change it beyond recognition.

Politicians are becoming increasingly familiar with sudden squalls of email filling up their inboxes, and policy makers with responses to public consultations arriving via a web intermediary. But not surprisingly many of these can be phoney, inflating the true size of what purports to be "grassroots" campaign.

The shortcomings of the web-based approach were illustrated here recently. Photographers on a shoestring budget successfully mobilised against Clause 43 - but internet campaigners concerned about file-sharing who used a site to send 20,000 emails about the Digital Economy Act failed to make an impression, resulting in a triumph for the BPI.

Earlier this month Obama's internet guru, Harvard academic Cass Sunstein, warned departments that internet opinion shouldn't be used as an opinion poll or focus group. He advised that:

Agencies exercise good judgment and caution when using rankings, ratings, or tagging. Specifically, agency use of the information generated by these tools should be limited to organizing, ranking, and sorting comments. Because, in general, the results of online rankings, ratings, and tagging (e.g., number of votes or top rank) are not statistically generalizable, they should not be used as the basis for policy or planning.

That's pretty conclusive. Four years ago, Sunstein published a love letter to Web 2.0 called Infotopia: How Many Minds Produce Knowledge that praised Wikipedia, blogs and prediction markets. But this is an altogether more sober assessment. He seems to have had second thoughts.

A vivid illustration of how a few single-issue fanatics can skew the results of an opinion poll is currently being digested in Canada. Lawyer Richard Owens has investigated the responses and found something quite interesting.

In response to a copyright form paper, over 8,000 responses were submitted, but 65 per cent of these were an identical form email from one IP address, the "Canadian Coalition for Electronic Rights". Owens notes that these:

...Included Submissions in which: no names were used; only first names were used (there were, for example, sixty-eight “Chris” and seventy-two “John” who made Submissions); and, suspect names, such as - “D Man”, “El Qwazo”, “pr0f1t”, “Cereal”, and “Eagle” - were used. Given the ability to submit anonymously or under false identification, is highly probable that there are multiple Submissions from the same persons.

The CCER form letter had been circulated around Bittorrent P2P fan sites. But most of the visitors to these sites aren't Canadian. Quantity overruled quality.

“Asking for people's identity is a pretty massive process”

Observers wondered whether something similar might have happened in the UK. When the Open Rights Group ventured into the real world, the numbers were small: it mustered just over 100 bodies for its main demo, and only single figures for its "flash mobs". The ORG's "Your Message To Mandelson" campaign launched last year rapidly gathered 300 anonymous messages - but stalled at around the 500 mark.

Such disparities led people to question how representative the online activity really was. "Is this a particularly well-focussed campaign by a relatively small group of activists?", asked the BBC's Rory Cellan Jones.

Our interest in the ORG write-in was piqued when one reader told us he'd successfully sent a Digital Economy email from a US IP address - he'd just made up a postcode.

Sponsored: Today’s most dangerous security threats

Next page: Fear the swarm?