Sunday November 18, 2012
US pollsters were actually spot-on
I REFER to the letter ďThey all failed to predict outcome of US presidential electionĒ (Sunday Star, Nov 11).
The writer stated that the opinion polls showed that the outcome was to be a tight race but the actual result was an easy win for incumbent President Barack Obama and therefore all the analyses in the world cannot predict the outcome of an election.
And certainly thatís the way it looks!
But let me (as a pollster in Malaysia) defend the US pollsters.
For the month before the election, the opinion polls, covering the whole country, indeed showed a very tight race with Obama usually at 51% and Republican challenger Mitt Romney at 49%.
After counting more than 100,000,000 votes, the actual result was 51.4% v 48.6% ó a result remarkably close to the opinion polls.
So how was it that the pollsters did not forecast Obama ending up the easy victor with 332 electoral college votes and Romneyís 206 electoral college votes?
Well, the pollsters did forecast this result but the media chose to play up the overall national vote to keep interest alive.
It is the peculiarity of the US electoral system that translates a narrow win in the popular vote to an overwhelming victory.
One must analyse the polls state by state where the candidate who wins the state by any margin ó big or small ó takes all the votes. Yes, the winner takes all. That is the peculiarity of the US system.
I followed the polls very closely. Hundreds of state polls were summarised by some media such as the Huffington Post and the BBC. The Huffington Post, the BBC forecast and others were almost 100% accurate even as far out as a month before the election.
This shows that polling conducted properly with the right analysis can correctly forecast results.
Chairman, TNS Malaysia