Polling – or perhaps more importantly, the abuse of polling data – has a checkered history in Venezuela. I will never forget the day of the presidential recall referendum in 2004. I was sitting in the BBC studio in Washington, between a TV and radio interview, and the fax machine spat out a release from what was then one of the most influential Democratic polling and consulting firms in the United States: Penn, Schoen, Berland and Associates. “Exit Poll Results Show Major Defeat for Chavez” was the headline. The agency claimed to have interviewed an enormous sample (more than 20,000 voters at 267 voting centers), and found that Chávez had been voted out of office by a margin of 59% to 41%. I looked up at the producer who gave me the fax with more than a bit of perplexity. With good reason: the actual result of the referendum was the opposite: 58% to 41% against the recall. The Organization of American States and the Carter Center observed the election and made it clear that there was no doubt that the results were clean. Given the actual vote, the probability of the variation in PSB’s result was less than 1 chance in 10 to the 490th power, if you can imagine something that unlikely. The producer put the press release aside. “I’m not doing anything with this unless there’s another source,” she said.
Which brings us to the current presidential election on Sunday. The most recent polls show a wide range of possibilities, from a 4% lead for President Hugo Chávez’s challenger, Henrique Capriles, to a 27.7% lead for Chávez himself. The average is a lead of 11.7 points for Chávez over Capriles. My CEPR colleague David Rosnick did a statistical analysis of the most recent polling data to adjust for the biases of the various polling firms, using data from 2004-2012. The adjusted lead increases to 13.7 percentage points, with Capriles having an estimated 5.7% chance of winning the election.
Nevertheless, much of the US media is making it look like a close race, in spite of the polling data. It’s not as bad as 2004, when most of the major foreign media were pretending – ridiculously, as it turned out – that the recall referendum was “too close to call”. There is progress in history.
Full Article: Venezuela election: why the voting is more trustworthy than the polling | Mark Weisbrot | Comment is free | guardian.co.uk.