Text size

This was no doubt the most polled election campaign, meaning the one with the most polls in the history of the state. Not only were there five or more polls a week, but each daily paper, each TV station, and the two main radio stations were all equipped with their own pollsters. In the last 10 days of the campaign, there were almost daily polls. On Election Day itself there were five polls and samplings, including one by Ynet and another by Israel Radio.

Even before the exit polls began, one could see the long line of people waiting to watch the pollsters fail. But it's not the pollsters who deserve complaints, but rather those who worship them and use them, meaning the politicians and media, including, of course, this writer.

It is important to remember that it is not the pollsters who told us the elections had been decided before the voting. It was Olmert and the commentators who told us so, and to a large extent they were right. The pollsters were correct about quite a bit. They told us there would be a large group of abstainers; they said 30 Knesset seats were up for grabs and that it was impossible to predict which way they would go; they said Kadima was declining and Yisrael Beiteinu was on the rise; and they even spotted the start of the fashionable trend to vote for the Pensioners, causing that trend to accelerate.

To prove that it was not possible to predict the Pensioners' achievement, Rafi Smith, who polled for Globes, conducted a post election poll in which he asked, "When did you finally decide who you would vote for?" Fully 21 percent said they decided on Election Day, and another 8 percent said that they decided during the last three days of the campaigning. It can and should be asked, what are pollsters worth if they cannot predict a last-minute phenomenon like the Pensioners? The answer is that no matter how limited the polls are, we still need them. The parties need them in order to know where they stand and what they should improve. And to be cynical about it, at least some of the parties did not know without the pollsters what policy to adopt and which vision to stand behind.

The press needs the polls because they are excellent entertainment, a little like Sudoku. The problem is that the exaggerated attention to the polls trivializes the political debate: It is a lot less important what is correct and just, a lot more important who is ahead and who has chances, who is rising on the charts, who is popular and who wins the big pot at the end. The polls not only simulate political life, they take its place. Not that the media has a choice. It is the only game in town, and everyone surrenders to its rules.

The result is that it does not matter how much the pollsters try to make clear that they are dealing with the science of probability; as far as the viewers are concerned, they are competitors in a TV game show. They are judged by the final result, and those who fail are attacked; and after all, everyone makes mistakes, because one has to guess the results of more than 10 parties, and it is impossible to be absolutely right all the time. Once every three years we get a chance to let off steam about the statisticians "who don't understand a thing," instead of the weathermen "who don't understand a thing." The statisticians might not be as sexy as the forecasters, but they are a lot more respectable. And for a few days, statistical science itself is made a mockery.

But the pollsters have nothing to worry about, because the elections for them are what Independence Day is for performing artists - an exhausting marathon of appearances that is sometimes humiliating and sometimes a little amateurish, but brings in a lot of money and fame. In one week, the public debate over their mistakes and failures is over. And by the next elections, nobody remembers the mistakes and everyone needs the pollsters just like this time (and maybe more), worshiping them just like this time (and maybe more) and accusing them after the elections of being wrong just like this time (and maybe with a more justifiable reason).