Number Cruncher Politics

Number Cruncher Politics is a political analysis and forecasting website and blog launched in 2014. The site is non-partisan and focusses on UK opinion polling, psephology, and statistical analysis. It is perhaps best known[1][2][3][4][5] for correctly predicting the polling failure at the 2015 general election.

History

On the eve the 2015 UK election, the site published a lengthy analysis[6] of opinion polling accuracy by its founder Matt Singh, based on decades of polling and election data. The report suggested that, contrary to the close race suggested by opinion polls and forecasts, the Conservative Party would win by more than six percentage points and could win an overall majority. In the event, the Conservative lead in the popular vote was 6.6 points, giving the party an overall majority of 12 seats. After the publication of the broadcasters' exit poll, the post containing the analysis attracted sufficient traffic to crash the website.[7]

At the Scottish independence referendum, the site was credited by FiveThirtyEight with the most accurate regional prediction.[8]

References

  1. ↑ "If they can’t get the hang of shy Tories, polls should be axed","Sunday Times, 10th May 2015"
  2. ↑ "How 'shy Tories' confounded the polls and gave David Cameron victory","The Guardian, 8th May 2015"
  3. ↑ "The return of the shy Tories?","The Economist, 7th May 2015"
  4. ↑ "The Blogger Who Beat the British Political Pollsters","Bloomberg Business, 31st July 2015"
  5. ↑ "Was anyone right about the pre-election polls?","Understanding Uncertainty (University of Cambridge Statistical Laboratory), 15th May 2015"
  6. ↑ "Is there a shy Tory factor in 2015?","Number Cruncher Politics, 6th May 2015"
  7. ↑ "Voter samples flawed in UK opinion polls, says ex-Barclays trader" Financial Times, 14 December 2015"
  8. ↑ "Scottish Regions’ View On Self-Governance Hasn’t Changed Much Since 1997","FiveThirtyEight, 19th September 2014"

External links


This article is issued from Wikipedia - version of the 6/14/2016. The text is available under the Creative Commons Attribution/Share Alike but additional terms may apply for the media files.