Analysis: When polls dictate reality

Eligible voters in Labor run-off race will decide whether votes will match up with poll whose smell may be just as harmful as its consumption.

Labor 311 (photo credit: Marc Israel Sellem)
Labor 311
(photo credit: Marc Israel Sellem)
President Shimon Peres is known for comparing polls to perfume. As he says, polls are nice to smell but harmful to drink.
Peres should know. He has finished first in the polls on multiple occasions only to finish last in ballot boxes where the votes actually count.
Labor candidates concerned about turnout in Monday vote
Politics: Labor’s long road to resurrection
He won the exit poll in his 1996 race against Binyamin Netanyahu but Netanyahu became prime minister. He lost the Labor chairmanship to Amir Peretz in 2005 despite a massive lead in the polls.
When done professionally, polls reflect reality and provide a valuable picture of public opinion at the point when they are taken.
When handled unprofessionally, polls can end up dictating an opinion to the public and forcing reality to change against its will. That’s what apparently happened in the Labor leadership race, whose first round ended Monday with a strikingly close race between MKs Shelly Yacimovich, Amir Peretz and Isaac Herzog.
Channel 2 aired a poll taken by the Sarid Institute July 20 that declared Yacimovich the front-runner in the race for the first time.
From then on, the press treated her as the inevitable winner, despite serious problems with the poll.
The poll was taken using a partial list of Labor members released by the party. Two media outlets, including The Jerusalem Post, purchased the list from the party for a considerable sum, but didn’t end up using it, because the data was very problematic.
A quarter of Labor members’ phone numbers were missing from the forms, especially in key sectors. Another quarter of the people were only on the list conditionally, due to problems with payment methods for their membership fees.
Channel 2’s regular pollster Mina Tzemach declined to take the poll and wrote an article about why polls during primaries are historically inaccurate. Rafi Smith, the Post’s pollster, begged out of the poll, leaving no alternative survey that would question the Sarid poll’s veracity.
Also, Channel 2 never purchased the list from the party, leaving open the question of whether they received it from a candidate who could have doctored the data.
Less than a week before the vote, Channel 2 struck again, broadcasting another Sarid Institute poll on Wednesday predicting a landslide victory for former Channel 2 anchorwoman Yacimovich in the first round with 42 percent, compared to 25% for Peretz, 16% for Herzog, and 15% for Mitzna.
In the actual result at the ballot box, the gap between Yacimovich and Peretz was less than 2%, not 17, and Yacimovich’s gap with Herzog was less than 8%, well below the 26% predicted.
Knowing the results of the Sarid poll that would be broadcast an hour later, Yacimovich announced in her speech to Wednesday night’s Labor convention that the race was between her and Peretz and any vote for Herzog was a wasted vote.
No one can know how many more votes Herzog would have received had such polls been banned by law, as has been proposed in the past. It’s possible that he would have made the second round instead of Yacimovich or Peretz.
Last Wednesday’s Sarid Institute poll also predicted that if a second round were held between Yacimovich and Peretz, she would win an overwhelming 66% of the vote.
The eligible voters in the run-off race will now have to decide whether their votes will match up with a poll whose smell may be just as harmful as its consumption.