Some 150 years after Friedrich Nietzsche issued his famous warning on the dangers of public opinion polls, Professor Yitzhak Katz, the founder and CEO the Maagar Mochot polling institute cites this very prophetic worry to demonstrate just how complicated polls are, despite making a living out of them for the past 30 years.
Katz,65, is one of Israel's veteran pollsters and has recently written a book on his experience called The people of the poll: behind the scenes of the polls, politics, and business
Follow Israel Hayom on Facebook and Twitter
Katz's point of view is two-pronged: as a self-critical pollster who is well aware of the limits and shortcomings of the polls, the pollsters and those who commission the polls; and as someone who knows his flock, the respondents, as he prepares for the Israel's third election in a year.
Katz doesn't have a lot of good things to say on the people he has worked with. In his book, he talks about three prototypical mayoral candidates: The stars, who win and also do a good job; the posers, who have great public relations despite not having much to show for as mayors; and the professionals, whose actions have truly made their towns better but don't get good PR.
"The most common race is between the posers and the professionals, and the former usually has a much better chance of winning," Katz says.
As a historian of polls, and not just someone who conducts them, Katz says that the data shows most of the Israeli public doesn't trust polls. In one poll, conducted by Tel Aviv University, 60% of respondents said they had a lack of confidence in the polls and only 19% said that it was important to get polls and assessments regarding the election. In another poll, a more recent one conducted ahead of the April election,, 51.5% said they did not believe polls and 42% said they do.
Katz pulls no punches. "Many polls have been a fraud," he says. "This is particularly common in municipal elections, but also in national election polls and in the runup to various elections in unions." in his book he provides examples of fraudulent polling in various local elections, but also in the April 2019 elections, when someone created a fake poll and falsely claimed his institute had conducted it. The "poll" claimed that Zehut, a libertarian right-wing party, was expected to win eight seats.
Q: How common are those fake polls? Has this been evident in this current campaign?
"I am sure that it's happening now. It is very easy to create a fake poll. It is primarily done through social media, and it happens quite often. Today people don't just fake polls, they also have no shame in admitting doing it. The most updated method is through an organized effort to plant respondents into online panels. Unlike past methods of fraud, you don't wait for polls to publish and then manipulate the results, you just plant your preferred respondents into the polls to begin with."
Q: How do they do that?
"There are about four or five serious samples in Israel that have tens of thousands of registered respondents. Each one of those samples is created through online recruitment of people who are willing to take part in polls. They willingly provide important information on themselves. This is how pollsters such as myself know the degree to which they are representative of society. They are rewarded financially for their time. This is how a real pollster knows his or her advantages and disadvantages. All of us know what we miss in our respondent pools and what needs to be corrected. Usually, it means we need more haredi or Arab respondents, as well as older respondents. But in recent years, more and more parties and other groups have tried to put their own constituents as respondents through an organized effort, or through a trickle, signing up their would-be respondents through the web."
Q: Do you have an example?
"I have to be careful. I know it happens, but i have no proof."
Q: Can you spot these fake respondents as having links with the ultra-nationalist Otzma Yehudit or the libertarian Zehut parties ?
"I won't comment on specific cases, but I can say that we have been aware of this phenomenon and there are ways to identify it and take care of it."
One of the most fascinating parts in Katz's books talks about a genre of polls that has become obsolete: asking voters to compare politicians to animals to see how they are perceived by the general public. A lion is considered strong and daring; a dog is loyal, but also considered derogatory; a donkey is considered hard-working but is also used for belittling someone.
In 1981, then opposition leader Shimon Peres was compared to a fox or a snake. In 1984, when he was already prime minister, he was perceived to be a tiger or rhinoceros, and in 1990, when he was accused of plotting to unseat the incumbent prime minister by reneging on a unity deal, he was compared to a hyena. Katz assumes that had the poll been conducted while Peres was president between 2007 and 2014, he would have seen more flattering comparisons.
Katz says there is a list of 13 variables that could affect the poll's quality and sometimes pollsters don't pay enough attention to them: the sample pool, which often includes outdated and partial databases; overrepresentation of online respondents, which results in the underrepresentation of haredim, older voters, Arabs or others who have limited access to the internet; the undecided voters, who only make up their mind at the very last moment; the low turnout; relying too much on the patterns from previous elections; and tactical voting in which people vote for a party they do not support just because it will help their overall ideological camp.
"On top of all that, there are respondents who just flat-out lie," he says. "In one campaign, there was a Facebook group called 'Let's mess with the pollsters.' There are also voters who just want to go with the flow rather than express their own opinion and there are also those who decide whom to vote for, or whether to vote at all, after seeing a poll. Some voters want to jump on the bandwagon if someone has momentum, while others see the poll and choose to vote for the underdog. The funds available to the pollster also play a role and impact the methodology and professional considerations. Sampling errors also have an impact of course, and sometimes there is a deliberate skewing of results, and of course, the fact that there are high expectations from pollsters does play a role."
Q: Why have pollsters stopped polling primary races?
"During party primaries, you have a variety of problems. The registry of party members is corrupted in many cases, and some members are only nominally registered and don't take part in the voting. There is also a lack of basic information on the candidates and the party and there is an impact of backroom deals and various voting arrangements that are dictated by certain wheelers and dealers to their various groups and camps. There is also the effect of the ground game on voting day. All those cannot be fully accounted for in polls for primary races, so it is best to avoid them altogether."
Katz managed to successfully predict the outcome of the Labor leadership race in 2005. Underdog Amir Peretz defied most polls and ended up defeating former prime minister Shimon Peres. He says that even back then voters gave impossible answers.
"We would ask them, 'Who are you voting vote for?' and they would respond: 'Whomever Itzik tells me to vote for.' Such answers referred to specific party bosses who had a following within a party."
Katz laments that "outside elements, who want to influence the results of the polls and interfere," referring to the relationship between media outlets and the pollsters. "This can involve a cordial relationship, but it could also involve threats."
Katz says that one candidate whispered in his ear while being interviewed, that if he published an unfavorable poll he will make sure to tell everyone that polls conducted by Maagar Mochot should not be trusted. Katz also reveals that one candidate tried to buy off the pollster.
"I had senior politicians who would call me and dispense professional advice that is aimed at influencing the results of the polls. For example one of them told me, 'It's important that you include respondents from the towns where we had a good showing during the municipal elections."
According to Katz, PM Netanyahu is among those who regularly commission polls and his conduct is very much influenced by them. But Katz refuses to give examples, perhaps because he worked with Likud over the past two elections.
In the grand scheme of things, despite its many faults, polling works, he claims. "It is a success story," he insists. "The fact of the matter is that people try to create fake polls just like they try to create knock-off brands."
Polls play a role in the political decision-making processes, in the vision leaders create and the strategic planning they undertake. But, he concedes, you cannot ignore the many constraints that come with the territory.
The Maagar Mochot Institute also conducts polls for Israel Hayom.
