The most insane thing about this poll is that four parties that fall under the threshold:

Otzma Yehudit 1.9%
Derech Eretz 1.4%
Gesher 1.4%
Labor 1.1% (!)

What about the top-line numbers? Is not B&W's rise to 19 more interesting?

Not necessarily. https://twitter.com/TheJeremyMan/status/1249755756506873856

Let's start with being aware of what *exactly* the pollster is asking the public - more important than usual in the current chaos.

The Direct Polls survey released yesterday assumed that Labor and Blue & White would merge; this Midgam poll asks about Labor running separately.

But be careful about the conclusions you draw based on that awareness. You might be tempted to say: Midgam showed B&W alone at 19, and Direct Polls showed B&W-Labor at 16! That means B&W will actually *lose* seats if it merges with Labor!

This is probably not the case.

Let's run through the numbers.

Midgam had Labor separately at 1.1%. That's less than a seat and a half. Call it two. So Midgam shows Blue & White and Labor getting a total of 21 seats if they run separately, while Direct Polls shows them getting only 16 if they run together.

Where do those seats go? Yesh Atid. Midgam shows Yair Lapid's party at 10, but Direct Polls has him rising to 14.

Are 20% of Blue & White's voters so economically liberal that they will object to a B&W-Labor merger so strongly that they will abandon the party for Yesh Atid?

But the reality is much more likely to be this:
* B&W is at ~17
* Labor is at ~2.
* Midgam is overcounting B&W by 2 seats (only 1.7% of the vote)
* Direct Polls is undercounting B&W-Labor by 3 seats (only 2.5% of the vote).

Those are very small differences.

Consider, too, that Blue & White and Yesh Atid draw from the same pool of disaffected centrists. So their poll numbers will swing with the full force of the margin of error; their numbers won't be over-stabilized due to demographic weighting the way most Israeli parties are.

But Direct Polls's margin of error for a party at 17 seats is only ~1.2%. (Midgam didn't publish their MoE.)

Where does the rest of the swing come from?

Direct Polls has a pro-Lapid and anti-Gantz house effect. Their March 27 poll actually had Lapid at 16 and Gantz at only 15!

(Three notes:

1) I'm not equipped to analyze pollsters' house effects the way @NateSilver538 does.

2) This is Midgam's first poll since the Blue & White split so we don't know their house effect.

3) Having a house effect does not necessarily mean Direct Polls is wrong.)

So it's likely that Labor and Blue & White joining forces doesn't bring Blue & White down, despite what a superficial glance at the top-line numbers would have you believe. It's just a difference between pollsters.

"But don't you keep telling us that mergers are counterproductive?"

Yes, but by that I mean that Party (X+Y) will get less together than Party X and Party Y would get separately.

I do *not* mean that Party (X+Y) will get less than Party X alone.

That sort of overreaction requires a massive voter backlash and a very good explanation - like if Blue & White merged with Otzma Yehudit or something.

A merger with the Labor party? No. Not when the #3 Blue & White candidate is the former head of Israel's largest labor union.