Elections 2022: How to make sense of the opinion polls

As Israel’s fifth election campaign kicks off, public opinion polls are back at the center of the national discourse. They shape media coverage, influence which parties end up running and which drop out or merge, and ultimately, impact voter behavior.

There is nothing inherently wrong with this.

To borrow and adapt the famous Winston Churchill quote, polling is the worst way of measuring public sentiment, except for all the others that have been tried. And as campaign strategists and pollsters ourselves, who have conducted polling for senior politicians and parties in Israel and over twenty countries around the world, every day we give advice to clients based on polling data.

The key is to understand exactly what the data is and is not telling us. To understand it in its proper context. To recognize its limitations.

This column aims to do just that. Every week for the duration of the campaign, we will try to unpick the polls — analyzing the prevailing trends, highlighting which issues to look for, and pointing out what the various parties and strategists will be deducing from them.

We will also include a “poll of polls,” updated weekly, which will bring together all the publicly available polling data to provide a snapshot of where the election stands as November 1’s election day draws closer.

Our first “poll of polls” is at the top of this article. It shows the number of seats parties would be expected to win if the election was held today. We are not rounding up or down — so that you can see more precisely how the parties are doing. Meretz is currently averaging below the 3.25% threshold for Knesset seats, while Yamina is right on the brink.

How polls in Israel work

Before analyzing the current state of play, it is worth looking for a moment at how polling in Israel works.

It is important to differentiate between the polling that the public is seeing, largely conducted on behalf of media organizations, and the polling the parties themselves are doing.

Public polls are designed to be cheap and quick, asking just a few questions in order to get a snapshot of the race, which can in turn inform coverage.

In contrast, the parties are doing in-depth, strategic polling, with many more questions. This quantitative research is then combined with qualitative focus groups, in order to provide a much more comprehensive understanding of public sentiment. This is used to build and refine strategy, and to test messaging.

While in the media, polls are used to give us a sense of current public opinion, the parties are using polling to help them manipulate and shape public opinion to their benefit.

Methodology is fairly similar across all the major polling companies. Polling in Israel today is largely conducted online, with a representative sample of somewhere between 500 and 600 Israeli adults taken from an online “panel” made up of tens of thousands of people. Pollsters will generally top up this sample with phone interviews to ensure an adequate representation of harder-to-reach groups, such as Haredim, the elderly, immigrants from the FSU and the Arab community.

Each pollster will have their own model of calculating seats from this data, with slightly different approaches to weighting, but the discrepancies between different companies’ polls are generally fairly small.

Prime Minister Naftali Bennett (L) and Foreign Minister Yair Lapid at a joint press conference at the Knesset, announcing the collapse of their coalition, on June 20, 2022. (Yonatan Sindel/Flash90)

For example, in the 24 hours following Naftali Bennett’s announcement that he would be stepping down as prime minister and dissolving the Knesset, five polls were conducted by different polling companies. In all five, Likud polled at 35 or 36 seats, followed by Yesh Atid with 20 to 22. In terms of “blocs,” the Likud core bloc (Likud, the Haredi parties and the Religious Zionism party) was winning either 59 or 60 seats in each of the polls.

Things to look out for

The polling over the past few weeks has been largely stable. This is likely the result of an election campaign that has yet to ignite, due in part to the public’s understandable election fatigue.

It is still unclear which parties will run and in what constellation, and how the blocs will ultimately be put together. At this early stage, we can say that the Netanyahu bloc is hovering around the 60-seat mark — very close to a majority in the 120-member Knesset.

With this election likely to be close, and decided by narrow margins, let us outline four key issues to look out for in the coming weeks and months, that could ultimately decide the election.

Bloc switching

Coverage of Israeli polling is increasingly focussed on blocs, rather than individual parties. This is because, after four rounds of elections with largely similar results, voters have become entrenched in their positions, and unlikely to switch between blocs. So while, for example, the fortunes of Labor, Meretz and Yesh Atid may fluctuate as voters move between them, the overall bloc size will likely stay fairly stable.

Yamina MK Ayelet Shaked (L) speaks with New Hope head Gideon Sa’ar at the Knesset in Jerusalem, June 2, 2021. (Olivier Fitoussi/Flash90)

In this context, a major factor in this election could well be parties, rather than voters, moving between blocs. Of particular relevance here is Yamina, now led by Ayelet Shaked, which could well switch into the Netanyahu bloc. Were this to happen, it would instantly boost the size of his bloc from 58.7 seats (in our latest average) to 62.9 — enough to form a government.

Whether Yamina voters would follow her into the Netanyahu bloc, however, or rather move to a different anti-Netanyahu center-right party, such as New Hope or even Blue and White, is another matter entirely.

The threshold

After some half-hearted attempts to lower the electoral threshold ended in failure, parties will again require at least 3.25% of the national vote in order to enter the Knesset.

For parties such as Yamina, New Hope, Meretz and Ra’am, all of whom are polling at around this level, the threshold could have major implications. In a tight race, one party dropping under could determine the outcome (as happened in the first election of this cycle in April 2019, when Bennett’s New Right party fell under the threshold by 0.03%, denying Netanyahu a coalition).

The past few election cycles have shown that the specter of the threshold looms large in voter calculations, though it can work either way. Either voters see a party on the verge of the threshold and opt not to waste their vote on them, sending them under, or voters come back to the party to save it, pushing it well above the 3.25% mark. Recent experience indicates that the latter is more common, but that is always liable to change.

When it comes to polling, the threshold provides a major challenge. In a sample of 500, one single respondent (!) can make the difference between a party polling at four seats or zero, which will in turn shape coverage of the party in question, and of the race as a whole (as, if a party slips below the threshold, those four seats will then be distributed to other parties, possibly helping the opposing bloc to a majority).

As a consumer of polls, therefore, it is important to be wary of reading too much into a party narrowly falling under the threshold in a poll, or a bloc increasing its numbers as a knock-on effect.

Undecideds

One of the biggest issues with the way polls in Israel are presented and covered is the lack of acknowledgment of the existence of undecided voters. While polls divide the public neatly into 120 “mandates,” in reality, a sizeable group are still undecided as to who they will vote for.

In polling we conducted ahead of the March 2021 election, at this stage (about four months before Election Day) a consistent 15 percent of the public — or the equivalent of 18 to 20 seats — was undecided.

This is almost never reported. Instead, pollsters either simply remove the undecideds from their calculations, or apply a model to allocate them to the party they will most likely vote for. The trouble is that this conveys certainty in an environment that is, in fact, highly uncertain.

For political strategists, identifying the undecideds and crafting messages that speak to them is often at the core of the campaign. Which way the undecideds “break” can determine the outcome of the election — so pretending they don’t exist is unhelpful.

For readers, without access to internal polling, this means there is a lot we don’t know. What we should be aware of though, is that however neatly the polls are presented, the reality is a sizeable — and potentially decisive — number of people haven’t made their minds up.

Differential turnout

One thing pollsters find extremely difficult to measure is turnout. Parties generally build various models based on different turnout scenarios, but the assumptions behind these are rarely rock solid.

Therefore, the question of differential turnout — one demographic or ideological group, such as Arabs or Haredim, voting at a significantly higher or lower rate than another — could well be a deciding factor.

Opposition leader Benjamin Netanyahu hails the collapse of the Bennett-Lapid coalition, at the Knesset in Jerusalem on June 20, 2022. (Yonatan Sindel/Flash90)

In recent weeks, there have been reports from pollsters who work in the Arab community that Arab turnout on November 1 could be lower than in the past, which would be a boost for Likud’s chances, though this of course could change as the campaign develops.

Benjamin Netanyahu, the challenger this time, having fought the past four elections as the incumbent, will be hoping that the combination of low Arab turnout and a highly mobilized right-wing base, after a year in opposition, will be enough to get his bloc to 61 seats at the fifth time of asking.

Simon Davies and Joshua Hantman are partners at Number 10 Strategies, an international strategic, research and communications consultancy, who have polled and run campaigns for presidents, prime ministers, political parties and major corporations across dozens of countries in four continents.

Leave a Reply

Your email address will not be published.