Pretty good agreement between state and national polling - both show a Harris national lead of +4. The state polling stabilized the week of August 10 and has been pretty stable since.

I my pro bono work in the past I have represented patients in their commitment hearings. You lost most of these cases - judges defer to the psychiatrist's judgement. That a hospital lost 99% of their commitment hearings is astounding. Great NYT piece. nytimes.com/2024/09/01/bus…

With margins this close in the EC, I am skeptical that a model really helps people understand the election. The best thing people knowledgeable about election data can say is that the election is really close—inside the normal range of polling error.

On today's episode of stupid polling tricks, rounding. Emerson releases their numbers with decimals - so the lead without pushing is 2.5% and when pushed 3.7%. RCP of course comes up with a lead of Harris +2 - but if you rounded the lead it would be 3%.

The Emerson Polling history - and it does provide story consistent with the election since June. Emerson had a 3 point shift before/after the debate, but the shift was smaller when they pushed undecideds. They found nearly a 9 point shift post Biden withdrawal.

Updated with the Yougov polling. There is here a glimmer of a sign WI/MI are moving into lean harris territory. If true NV becomes big as it allows Harris to win with either GA or NC. Ground 0 is PA. Harris the favorite here - though not a big one.

Is this a replay of '20? The average shift in the Yougov polling from the results in '20 is .1%. Still inside the average error everywhere except WI.

-Employment at All-Time Highs -Wages: ATH -Consumer spending: ATH -Corporate profits: ATH -Credit card defaults: Off lows, but still modest -Corporate + Personal bankruptcies: Low What data are you looking at?

Virtually identical to the margin in 2020. 2024 is a rematch of 2020 - which should surprise no one.

This a good answer from Walz.

The key numbers in the Unemployment report: Average earnings up⬆️ .4% on the month (versus CPI up .2%) Year of Year average earnings ⬆️3.8% versus inflation of 2.9%. Inflation is up .1% since May, Wages are up .6% since May. bls.gov/news.release/e…

TL;DR - Nates model is wrong and he knows it. Put another way this election is so unique (debate/assassination attempt/convention/Biden withdrawal) that any model based on past elections is wrong.

I first created this chart in 2012 as a way to step back from the inevitable day to day reaction from individual polls and try and understand the election in a broader perspective. The election clearly broke in the 1st week in August and has been pretty stable since.

Trump won Texas by 5.5%. So another poll showing a replay of 2020 (like the Emerson poll from Florida earlier) and the Yougov state polling yesterday (7 polls, and an average shift of .1% from 2020). This isn't complicated folks. Any shift in margins are tiny.

Emerson has 2 numbers - a normal trial heat and a second where they push undecideds. In their polling this am, this is how the undecided split: FL Harris 2.3 Trump 1.6 OH Harris 1.7 Trump 1 TX Harris 1.8 Trump .9 Undecideds in this morning's Emerson polling break 2-1 for Harris.

Very interesting - in yesterday's Emerson national poll undecideds broke 1.8% for Harris and .7% for Trump. I have been tracking the undecided in Emerson for months: it is young and more likely to be POC. In yesterday's national poll 74% had a strongly unfavorable view of Trump

% of undecided by race: Emerson Polling National - 27% black, 25% Hispanic, 7% Asian, 32% white Texas - 73% white, 18% black, 10% Hispanic Florida - 30% white, 17% Hispanic, 52% black Ohio - 64% white, 18% black, 9% Hispanic

The nonsense being repeated about the African American vote is indefensible. Since 8/13 Harris has been running AHEAD of Biden's pre-election polling among African Americans in 2020 and WAY ahead of D performance in 2022. This becomes REALLY clear when you look at the states.

To my knowledge I am the only 1 looking at the state x-tabs - there is a hell of a lot more data available when you do that. The sample sizes in National polls are too small. For example about 30% of Georgia voters are African American - which means you get MUCH bigger samples.

This is basically the experience of every young prosecutor when they meet with the DA on a case they are trying. The footnote questions were in my experience a sign that the DA had READ the document. Why is an oft answer question in a DA's office where preparation is expected.