Thread by @lxeagle17

No update to our nonpartisan generic ballot poll tracker this morning for @SplitTicket_, mainly because we have seen no new generic ballot polls (!). But I thought I'd take this thread to go through where our aggregator was at specific days, compared to 538 and RCP.

For the sake of transparency, I'll put up that we have data for our aggregate from June 1 onwards. The data is in the table, on our site. However, we only display the GCB from September 1 onwards in our graphic, which is around when the LV screens begin to hit polls. OK, now...

CA Primary (6/07):
@SplitTicket_: R+2.5
@FiveThirtyEight: R+2.2
@RealClearNews: R+3.4

Dobbs v. Jackson (6/24):

@SplitTicket_: D+0.4
@FiveThirtyEight: R+2.3
@RealClearNews: R+2.6

¯\_(ツ)_/¯

#NE01 special, which was an R+5 win in a Trump +11 seat on 6/28

@SplitTicket_: D+2.7
@FiveThirtyEight: R+2
@RealClearNews: R+2.5

#MN01 special on 8/9, R+4 in a Trump +10

@SplitTicket_: D+1.7
@FiveThirtyEight: D+0.1
@RealClearNews: R+0.1

#AKAL special on 8/16, D win in Trump +10 state

@SplitTicket_: D+2.1
@FiveThirtyEight: D+0.5
@RealClearNews: D+0.2

#NY19 special, 8/23, D+3 in a Biden +2 seat

@SplitTicket_: D+3.1
@FiveThirtyEight: D+0.3
@RealClearNews: R+0.2

LABOR DAY (9/5)

@SplitTicket_: D+3.6
@FiveThirtyEight: D+1.0
@RealClearNews: R+0

TODAY (10/31)

@SplitTicket_: D+0.4
@FiveThirtyEight: R+0.8
@RealClearNews: R+2.9

As you can see, FiveThirtyEight's aggregator is the most stable one out there (and honestly, the best IMO). I have no idea what RCP is doing, because their criteria for poll inclusion/exclusion is not transparent at all. Ours is quite noisy, as you'd expect most aggregators to be

Directionally, though, all of them show that the election has shifted R since Labor Day. How much of that is because of the introduction of LV screens, and how much of that is persuasion? Honestly, I don't know. My guess is, they both have something to do with it.

Anyways, use this for what you will. Public polling is noisy, as shown by the fact that our aggregator had Dems up on the day of the leak by 0.4 but down by 3 on the CA primary date -- that's because of a paucity of polling in that period that caused polls to "age out".

This is less of an issue right now, and less of an issue when you have a consistent volume of polls, but...well, let's hope we get some polls soon this week!