Five Thirty Eight - a site that “combines semi-professional expertise with a dollop of eye-witness reporting,” according to the FT’s John Gapper - has neatly summarised Mark Blumenthal’s comprehensive dissection of the media’s reliance on exit polls as an electoral indicator.
Caveat - Blumenthal’s collection of essays dates back to 2004, but the conclusions remain valid. Some highlights:
Exit polls have a much larger intrinsic margin for error than regular polls
Exit polls have consistently overstated the Democratic share of the vote
Exit polls were particularly bad in this year’s primaries - They overstated Barack Obama’s performance by an average of about 7 points.
Democrats may be more likely to participate in exit polls.
Obama has a strong lead in the (hopelessly preliminary) major polls, and the results of the exit polls in major states will start rolling out within the next hour. Take those results with a liberal sprinkling of salt.
(H/T Dan McCrum)
Article Series - US Elections 08
- First blood to Obama
- "We are not sure if an election result has ever mattered so little to the markets."
- Why you should ignore exit polls
- Market talking heads: election means nada, redux
- Intrade odds favour Obama (shock)
- Pink Picks, White House edition
- Obama, the markets and the perils of presidential data mining
- Further reading, White House edition
- Presidential data mining, redux
- Gideon Rachman doesn't give a twit
- Dirty dirty exit poll data, round one
- McCain gets Kentucky and South Carolina, Obama claims Vermont
- Obama vs McCain - how the networks see it
- 270, 538, swing states, electoral college - what does it all mean?
- Obama wins Ohio, White House in sight
- As Obama sweeps Ohio, Pennsylvania and New Mexico...
- Electoral vote tally - Obama 207, McCain 129
- Barack Obama wins Virginia, California - and the White House
This entry was posted by Stacy-Marie Ishmael on Tuesday, November 4th, 2008 at 17:08 and is filed under Capital markets, People.
Tagged with Barack Obama.