Opinion Polling for the April 2019 Israeli Legislative Election

The April 9, 2019 contest was the 21st Knesset election, held for a 120-seat parliament in which any governing coalition required at least 61 seats.[4] Forty party lists ultimately participated after 47 registered, while a pre-election commentary had described 43 parties as vying for seats.[5][6] A 3.25 percent popular-vote threshold applied to Knesset representation, and nationwide turnout reached 68.46 percent.[7][5]

At the start of the campaign, Likud was widely treated as the clear favorite. The entry of former general Benny Gantz into politics and the later merger of his party with Yair Lapid's Yesh Atid changed that assessment, with polling indicating that the resulting Blue and White list could win the most Knesset seats.[8][9]

In late 2018, before a major opposition challenger had emerged, Likud was the prohibitive favorite in polling, with opposition parties registering far behind.[8] A December 2018 Channel 2 News poll found that a hypothetical joint list combining Gantz with Lapid's Yesh Atid would come within a few seats of Likud, although the report did not give exact totals.[8]

A Midgam Institute poll conducted January 14–16, 2019, surveyed 1,007 Israeli adults and carried a maximum sampling error of 3.1 percent.[10] It projected Likud at 32 seats—up four from Midgam's previous poll—while Gantz's Hosen LeYisrael was projected at 13 seats and Yesh Atid at 14. One in five respondents said they were still undecided about which party to support.[10]

A separate January 22 Washington Institute report described Netanyahu as holding a clear lead and commanding the support of roughly a quarter of the electorate, while noting that a 61-seat majority in the 120-member Knesset would be required to govern.[4] The report did not identify the specific polls on which that characterization was based.

Polling methodology varied across outlets. The January and February Midgam surveys sampled 1,007 and 515 respondents respectively, with reported maximum sampling errors of 3.1 and 4.4 percent.[10][11] The March Midgam–iPanel survey combined telephone and online interviews and reported a 4.4 percent maximum sampling error; its article body listed 512 respondents, while its methodology note listed 510.[12]

The Maagar Mohot poll used telephone and online interviewing among more than 1,000 respondents and reported a 3.1 percent margin of error.[13] The final Maariv poll combined an online panel and telephone interviews among 1,097 adults, with a 3 percent margin of error.[14] The Times of Israel survey used online panels among 708 likely voters and reported a ±3.7 percent margin of error at a 95 percent confidence level.[15]

Before the merger was announced, a February 11 analysis reported that polls on prime ministerial suitability showed Gantz running neck and neck with Netanyahu, while Israel Resilience led Yesh Atid 23 seats to 9 as separate parties.[16] Projections suggested that a joint list would win 36 seats, compared with 32 combined if they ran separately.[16]

On February 21, 2019, Hosen LeYisrael and Yesh Atid formally merged into Blue and White (Kahol Lavan). A Midgam poll with responses collected that same day—surveying 515 people with a 4.4 percent sampling error—placed Blue and White at 36 seats and Likud at 30, with the right-wing and center-left blocs tied at 48 seats each.[11] Netanyahu led Gantz and Lapid in a rotation question on prime ministerial suitability by the narrowest of margins, 37 percent to 36.[11] A Yediot Aharonot poll reported around the same time showed an identical 36-to-30 split, described as the first occasion Likud had trailed in the race.[1]

The Times of Israel conducted its own survey from February 24 to 27 among 708 likely voters, using online panels and including 18 percent Arabic-speaking respondents surveyed in Arabic; the margin of error was ±3.7 percent at a 95 percent confidence level.[15] Rather than Knesset seat projections, the survey recorded voting intention as percentages: 23 percent for Blue and White and 19 percent for Likud, with 18 percent still deciding.[15] On the leadership question, 53 percent said the time had come for a change in prime minister, while 41 percent chose Netanyahu over Gantz as their preferred premier, against 39 percent for Gantz.[15]

A Congressional Research Service report, drawing on polling aggregated from Israeli sources for March 3–5 and attributed to BICOM, described Gantz and Netanyahu as running "almost even" in most polls asking voters their preferred prime minister.[9][7] The CRS added that some polls suggested Likud could retain an advantage in forming a government even if Blue and White won the most seats, because Likud's right-leaning and ultra-Orthodox partners kept it close to a coalition majority.[7] Kulanu, then part of Netanyahu's coalition, was identified as a potential swing party that might join either a Netanyahu-led or a Gantz-led government.[7]

A Midgam poll conducted March 13–14 in cooperation with iPanel and published by Channel 12 projected Blue and White at 31 seats and Likud at 28.[12] The right-wing bloc was put at 51 seats and the center-left at 46—a reversal from the preceding poll, which had favored the center-left.[12] Six lists were projected near the four-seat electoral threshold, and Gesher was projected at 2.3 percent, below it.[12] A Jerusalem Post review of polling during this period noted that Likud led Blue and White in mid-March polls from Kan and Walla, and again in a March 27 Walla poll, after Blue and White had fallen below 30 projected seats in a March 19 Walla poll.[17]

Several polls during the campaign tracked specific parties and demographic groups. An Israel Democracy Institute poll found Netanyahu supported by approximately two-thirds of voters aged 18–24 and by more than half of those aged 25–34.[5] The same source noted that 56 percent of Israelis described themselves as right-wing, compared with 40 percent fifteen years earlier.[5]

Polls indicated that Labor could win as few as five seats, signaling a historic low for the party, which had held 24 seats in 2015 as part of the Zionist Union alliance with Hatnuah.[18] Polls also forecast that Tzipi Livni's Hatnuah would fall short of the 3.25 percent threshold; Livni subsequently announced that she was leaving politics and Hatnuah would not run.[18] Yisrael Beiteinu, whose leader Avigdor Lieberman had resigned from the government in November 2018 to protest Netanyahu's ceasefire agreement with Hamas, was projected to barely clear the threshold.[19] A pre-election commentary noted that polls had consistently shown Moshe Feiglin's Zehut party at five to seven seats in a campaign with 43 parties described as vying for seats.[6]

Israeli law barred the publication of new polls during the days immediately before the election, making April 5 the last date on which results could be released.[20] Globes compiled five final polls: Yediot Ahronot put Blue and White at 30 seats and Likud at 26; Maariv–i24News at 28 and 27; Israel Hayom at 32 and 27; Channel 12 at 30 and 26; and Kan at 30 and 31.[20] The average across those five was 30 seats for Blue and White and 27 for Likud.[20] Despite Blue and White's average lead, Globes noted that Netanyahu could potentially assemble a 63-seat right-wing coalition.[20]

The final Channel 12 poll projected 30 seats for Blue and White and 26 for Likud, with Labor at 10; United Torah Judaism and Hadash-Ta'al at 7 each; New Right at 6; and Shas, Union of Right-Wing Parties, Meretz, Zehut, Kulanu, and Yisrael Beiteinu at 5 each. Ra'am–Balad was projected at 4, while Gesher remained below the threshold.[21] Its four bloc subtotals—52 for the right, 12 for the Haredi parties, 45 for the center-left, and 11 for the Arab parties—amounted to 64 seats for the right-plus-Haredi group and 56 for the center-left-plus-Arab group.[21] On the leadership question, 36 percent preferred Netanyahu as prime minister and 35 percent preferred Gantz, while 58 percent believed Netanyahu would ultimately form the government.[21]

The Maagar Mohot Institute poll for Israel Hayom and i24NEWS, conducted among more than 1,000 respondents by phone and online with a reported margin of error of 3.1 percent, projected Blue and White at 32 seats and Likud at 27.[13] Among respondents certain to vote, Blue and White was projected at 34 and Likud at 26, yet the right-wing bloc still led 65 to 55 on that subset.[13]

The poll found 52 percent believing Netanyahu would be the next prime minister, 27 percent expecting Gantz, and 21 percent uncertain. It also found 46 percent opposed to a unity government between the two large parties, against 31 percent in favor.[13] A Maariv poll from the same day, combining an online panel with telephone interviews among 1,097 adults at a 3 percent margin of error, put Blue and White at 29 and Likud at 27, with the right-Haredi bloc at 66 and the center-left at 54.[14] Previous surveys indicated that nearly one in six or seven voters made their final choice only in the last two days before the election.[14]

The Washington Institute reported that polls conducted around April 5 projected 18–19 seats combined for Labor and Meretz and a center-left total of 58 seats, characterizing that as a five-seat gain for the center-left.[2] One contributor cautioned, however, that undecided voters might be choosing among parties within the same ideological camp rather than crossing between blocs.[2]

The high proportion of undecided voters was a persistent feature of pre-election polling. Four days before the vote, polls indicated that as many as 35 percent of Israelis still did not know how they would vote.[2] The January Midgam poll had similarly found 20 percent undecided.[10]

A panel-survey study of undecided voters in the April 2019 election, using samples of 1,427 and 912 respondents across two waves, distinguished more politically sophisticated undecided voters from less sophisticated ones based on levels of political interest and knowledge.[22] Sophisticated undecided voters were more likely to follow the campaign through both traditional and social media, discuss the election online, ground their final decision in policy issues, choose between parties within the same ideological camp, and ultimately turn out to vote.[22] In the final Maagar Mohot survey, 26 percent of respondents who remained undecided leaned toward right-wing parties, 27 percent toward the center, and 19 percent toward left-wing parties.[13]

Three broadcast exit polls were released simultaneously when polling stations closed on the evening of April 9.[23] Channel 12 projected Blue and White at 37 seats and Likud at 34, with Netanyahu's potential supporting parties reaching 60 seats.[23] Channel 13 projected a 36–36 tie between the two parties and a right-wing, ultra-Orthodox coalition of 66 seats.[23] Channel 11 projected Blue and White at 37 and Likud at 36, alongside a 64-seat Likud-supporting coalition.[23]

Gantz announced what he framed as a victory based on Channel 12's projection when fewer than 5 percent of votes had been counted; as the count approached 20 percent, Channel 13's figures appeared closest to the emerging result.[23]

Preliminary reporting initially put Likud at 36 seats and Blue and White at 35. The certified final result instead gave each party 35 seats, with one seat moving from Likud to United Torah Judaism; their final vote shares were 26.46 percent and 26.13 percent respectively.[5]

Post-election reviews found that pre-election polls had correctly identified the right-wing bloc's advantage but had misjudged the distribution of seats within each bloc.[24] An Israel Hayom analysis gave the average pre-election forecast as 67.5 seats for the right-wing bloc and 52.5 for the center-left, against an actual right-wing result of 65 seats.[24] The more significant error lay in the large-party projections: nearly all polls had anticipated a combined Likud–Blue and White total of around 60 seats, while the two parties together won 70.[24]

Polling had placed Blue and White at roughly 30 seats, Labor at 9–11, and Meretz at 5–6; the final results were 35, 6, and 4 respectively.[24][5] The Israel Hayom analysis attributed the late surge of both large parties to voters shifting away from smaller allied parties in the final days, estimating that Blue and White gained four seats from Labor and one from Meretz during the last three days of the campaign.[24]

On the smaller parties, the last six major-media polls gave Kulanu five seats while it won four, and the Haredi parties were underestimated before finishing with 16 seats in the 21st Knesset.[24] Zehut—often projected at five to seven seats—failed to cross the electoral threshold, as did New Right and Gesher.[6][25]

A Globes retrospective noted that threshold-level projections were particularly consequential because a party polling around four or five seats could end up with no representation.[3] It listed possible sources of polling error, including voters changing their minds after being surveyed, withholding or misrepresenting their intentions, difficulty sampling some populations, and undecided voters making late choices during the pre-election publication blackout.[3] Pundits who had forecast Blue and White winning around 25–30 seats saw the party finish at 35, one fewer than Likud's preliminary 36-seat tally and level with Likud in the certified result.[26][5]

One study examined an inaccurate exit poll broadcast by a leading Israeli television channel during the April 2019 election, using focus groups, interviews with journalists, pollsters, and experts, and qualitative analysis of news coverage.[27] It identified three categories of public explanation for the error: statistical error in the polling process, alleged bias or misconduct by media and poll creators, and alleged deception by right-wing voters.[27] These were competing public narratives identified by the study, not established causes of the error.

A separate study examined media-sponsored pre-election polls across Israel's three elections of 2019–20, defining outlet political leaning by reference to attitudes toward Netanyahu's corruption charges.[28] Its reported finding was that polls commissioned by anti-Netanyahu media underestimated the pro-Netanyahu bloc's eventual Knesset seats on average. The catalogue abstract covered all three elections of that period rather than isolating the April 2019 election.[28]