Systematic Bias: Who Pollsters Overestimated, and Who They Underestimated

On average, campaign polls underestimated the coalition by 5.68 seats (and gave the opposition a slight boost of +2.31) — not random noise, but a consistent bias throughout the campaign. Most underestimated: Religious Zionism (-3.99 seats on average). Most overestimated: Meretz (+4.37 seats on average).

Meretz: The Campaign's Biggest Miss

In the last ten polls before election day, every single poll estimated that Meretz would pass the electoral threshold, averaging 4.5 seats. Meretz actually received 0 seats. This was a missed threshold event: not one pollster identified that Meretz would fall below the electoral threshold.

Meretz did not pass the electoral threshold and received 0 seats — a threshold failure, not just a magnitude error.

Accuracy Improved as the Election Approached — But Not at a Steady Pace

In the final 0–14 days before the election, the average error was 1.159 seats per party, compared to 2.519 at 90+ days before the election — 2.2x higher. That said, the improvement wasn't gradual: accuracy in the last month was nearly identical to accuracy in the last two weeks, meaning most of the improvement happened earlier in the campaign.

Each Pollster's Profile

Every pollster in the database most overestimated Meretz and most underestimated Religious Zionism. The direction of the error was shared by everyone — it mainly differed in magnitude.