10 Reasons the Polls Could Be Wrong

In less than six hours we will have the next best thing after real election results, as voting ends at 10 P.M. and the three main television channels broadcast their exit polls. A few hours later, the first true results will come in. Meanwhile, all we have to go on are the hundreds of polls conducted over the last 100 days – and all of them have pretty much said the same thing.

• LIVE BLOG: Israelis Go to the Polls

The right-wing-religious bloc of parties supporting Prime Minister Benjamin Netanyahu has a clear majority in the next Knesset, while Likud-Yisrael Beiteinu has been steadily losing votes, but almost primarily within the rightist bloc, and almost always to Naftali Bennett's Habayit Hayehudi.

A few polls published toward the end of last week – the last polls allowed by law to be published – showed the overall gap in seats between the rightist-religious bloc and the center-left bloc slowly shrinking, down from about 66-54 in previous weeks to around 63-57.

The parties are continuing to conduct their own polls even today, and from what I’m hearing the main trends are holding – but, of course, it's difficult to separate reliable data from disinformation. So we will have to wait.

But while we wait, here are the 10 main reasons why the polls, though generally consistent, could all be wrong, and why the gap in favor of Netanyahu may either be much wider than we think, or might not exist at all.

1. Lack of voting patterns – Pollsters rely on previous elections to build their voter samples. But in an election with new parties in the center (Yesh Atid and Hatnuah) and the mergers of four parties on the right (Likud with Yisrael Beiteinu, and Habayit Hayehudi with National Union), there are no previous voting patterns to rely on and it is difficult to accurately balance the sample.

2. Election Day logistics – Israeli politicians like to say that strong organization and good logistics on Election Day can win parties an extra seat in the Knesset. This is true and favors particularly the older, more established parties (Likud and Labor), who still have veterans, and religious parties such as Habayit Hayehudi and Shas, who can rely on religious institutions to supply them with steady streams of volunteers. Newer parties just don't have the manpower or the experience.

3. Turnout – The 2009 elections had a relatively low turnout of 65 percent, and most polls predicted even less enthusiasm this time around. The polls were weighted accordingly, but the figures released by the Central Elections Committee so far seem to indicate a much higher turnout than expected. Polls that tried to analyze the wavering and stay-at-home voters showed that they were more left-leaning than the general public. If some of them do vote, it could change the overall picture.

4. The Israeli-Arab vote – Traditionally, Israeli Arabs have voted in smaller numbers than Jewish voters (only 52 percent in the last elections). If all of them voted, there would never be a right-wing government in Israel, but even if the Arab turnout is only similar to the Jewish one, that could boost the leftist bloc by at least three seats.

5. Small parties – Parties on the brink of the electoral threshold (2 percent of the vote) could have a significant effect on the division of seats. If, for example, Kadima scrapes through, that would mean two or three additional centrist Knesset members. The same is true for the right-wing bloc if neo-Kahanist Otzma Leyisrael gets in. If either fails to cross the threshold, the votes will be lost to their respective blocs and the balance will go to the other side. It has been almost impossible for the polls to predict the fortunes of the splinter parties.

6. Non-communal voters – Another method pollsters use to build their sample is the relative size of various communities and their expected votes. But the sector's political allegiances are not always clear-cut. Large numbers of Haredi voters prefer voting for secular parties, and many secularists seem to be planning to vote for religious Habayit Hayehudi. It is unclear also how many "Russian" voters are planning to vote for Avigdor Lieberman now that his Yisrael Beiteinu has merged with Likud.

7. Compromised pollsters – Many Israeli pollsters simultaneously conducted the same surveys for news organizations and for different political parties. Without casting aspersions against any specific pollster, data collected for parties – who have a clear interest in framing the polls in their favor – is tainted and should be considered suspect.

8. Untrusting voters – Many Israelis, especially those who have been brought up in totalitarian societies or live in insular communities, are extremely suspicious of phone calls from polling companies. They find it hard to believe that their voting intentions will be kept secret and tend to play it safe by saying they will vote for the party in power or the party that their friends and family expect them to vote for.

9. Outdated polling techniques – Most political polling is still conducted by phoning landlines, but many voters, especially younger ones, only use mobile phones these days. Some polls have included cell-phone users, but the polling companies are still unsure what the correct proportion of these should be in the sample. This could result in an entire young generation of voters being under- or over-represented in the polls.

10. It's the system – The truth is that Israeli polls generally get the results wrong and at best are reliable enough to detect voting trends, not predict the actual division of seats in the Knesset. This isn't due to a lack of professional pollsters, but to the unique nature of Israel's voting system. A multi-party system of proportional representation contains a vastly larger number of variables to be worked out than systems based on first-past-the-post races, most of which are between two or three large parties.