Polling before election day was intense, every poll from every known company was scrutinised and debated endlessly on Twitter and in the empty voids of the rolling news channels. But it seems the only people to have a worse election night than the Labour Party were these very pollsters. No one predicted the result accurately and some who were expected to be the closest, were out by reputation-ruining margins. The margin of error ranged from four up to a wild twelve points out. When IPOS-Mori called almost the exact result two weeks ago, the poll was dismissed as “rogue”.

So sure were the media in “Cleggmania” that any suggestion that the Lib Dem surge wasn’t going to materialise was instantly dismissed. It was no surprise then that the lefties over a Tweetminster got their predictions so widly out – they based their poll figures on mentions and “buzz” and attempted to pitch this as a serious methodology in a bid to get in on the polling frenzy. Everyone was talking about the Lib Dems, but a lot fewer actually bothered voting for them. Tweetminster’s insistence that results could be foreseen by hype looks embarrassingly naive on refection. Pollsters claim a margin of error of plus or minus 3%, most were out by some 4% on the LibDem vote. The biggest margin of error was Tweetminster’s prediction that Esther Rantzen would win in Luton South based on positive tweets for the fading celeb – the prediction garnered them a few headlines from gullible hacks.  Despite masses of media coverage Esther polled a mere 1,872 votes and lost her deposit. A margin of error of over 30%.

The real pollsters have a bit of explaining to to do as well. The cut price discounts they did for political polling, with the free hype-based advertising that comes with it, might not seem such a good idea when the corporate world who pay the big bucks see just how inaccurate their forecasts have been.

See also : Tweet Predicting Election Test Confirms GIGO Principle