Biden Approval Rating By Entrecampos

[Entrecampos]

RATING	POLLSTER	Biden Approval %	Biden Disapproval %	Biden Approval - Disapproval
S+	Trafalgar	40	56	-16
B+	YouGov	42	48	-6
B+	Ipsos	48	47	1
B	Rasmussen	43	56	-13
B	Morning Consult	46	47	-1
C+	Quinnipiac University	38	53	-15
D-	Redfield & Wilton Strategies	45	37	8
	Entrecampos	42.3	52.5	-10.2

[Entrecampos]

I use an updated version of my 2020 Presidential election pollsters model that did a very good job predicting portuguese mayoral elections and USA 2020 elections (the best prediction in voting shares that I have knowledge).

Edited October 8, 2021 by Entrecampos

[Hestia]

9 minutes ago, Entrecampos said:

RATING	POLLSTER	Biden Approval %	Biden Disapproval %	Biden Approval - Disapproval
S+	Trafalgar	40	56	-16
B+	YouGov	42	48	-6
B+	Ipsos	48	47	1
B	Rasmussen	43	56	-13
B	Morning Consult	46	47	-1
C+	Quinnipiac University	38	53	-15
D-	Redfield & Wilton Strategies	45	37	8
	Entrecampos	42.3	52.5	-10.2

S+ to Trafalgar? You've got to be kidding. That's a complete joke.

[Hestia]

They did better in 2020, but they had some of these beauties too.

[Entrecampos]

26 minutes ago, Hestia said:

S+ to Trafalgar? You've got to be kidding. That's a complete joke.

Yes, governors and senate elections are harder to predict, the big majority of the other pollsters, have the same problems in these type of elections, they have very good records in presidential elections of 2016 and 2020, and this is my main way to evaluate pollsters (2.75% mean error versus 4.80% of Quinnipiac for example). They were rated as C- in 2020 by 538, and this was a enormous giant joke that I detected by that time. Comparing with Presidential predictions of Quinnipiac (I removed other type of elections) for example that has the same rating of Trafalgar in 538 now (A-), you see they are much worse. I know that 538 tries to do the overall between Presidentials and other elections, but for me Presidential polls have much more importance than the others.

 

Edited October 8, 2021 by Entrecampos

[Hestia]

Just now, Entrecampos said:

Yes, governors and senate elections are harder to predict, the big majority of the other pollsters, have the same problems in these type of elections, they have very good records in presidential elections of 2016 and 2020, and this is my main way to evaluate pollsters (2.75% mean error versus 4.80% of Quinnipiac for example). They were rated as C- in 2020 by 538, and this was a enormous giant joke that I detected by that time. Comparing with Presidential predictions of Quinnipiac (I removed other type of elections) for example that has the same rating of Trafalgar in 538, you see they are much worse. I know that 538 tries to do the overall between Presidentials and other elections, but for me Presidential polls have much more importance than the others.

 

So completely ignoring an entire cycle in between because it's poor for that pollster? That doesnt seem like a good idea. Several pollsters were much closer, in fact 2018 was better polled than 2016 generally, even though some bumps still in the Midwest. They may be better, but S tier is REALLY overblowing it. 

[Entrecampos]

7 minutes ago, Hestia said:

So completely ignoring an entire cycle in between because it's poor for that pollster? That doesnt seem like a good idea. Several pollsters were much closer, in fact 2018 was better polled than 2016 generally, even though some bumps still in the Midwest. They may be better, but S tier is REALLY overblowing it. 

I need to admit one thing, I lost most of my time analysing pollsters in presidential elections, my elo system has 90% of data put in Presidential elections and 10% in other type of elections (because lack of time), and 0% of primaries, this ranking is made to be good in presidentials, not in other type of elections, I need to adjust Approval ratings from this Presidential elections rating to general elections rating (I will need some weeks maybe). Sorry for that.

Edited October 8, 2021 by Entrecampos

[Hestia]

Just now, Entrecampos said:

I need to admit one thing, I lost most of my time analysing pollsters in presidential elections, my elo system has 90% of data put in Presidential elections and 10% in other type of elections (because lack of time), this ranking is made to be good in presidentials, not in other type of elections, I need to adjust Approval ratings from this Presidential elections rating to general elections rating (I will need some weeks maybe). Sorry for that.

That's ok. Your method, I just don't think Trafalgar is a gold standard with some of these stats.

[Entrecampos]

4 minutes ago, Hestia said:

That's ok. Your method, I just don't think Trafalgar is a gold standard with some of these stats.

My 2020 Elo System was much better than Trafalgar in Presidential Elections, but even my model had 0.5 to 1.0% GOP bias, and was enough to be wrong in the winner of Wisconsin, Arizona and Georgia (if my national voting share was 100% right, i woud be right in all states). Unfortunately its hard to search and find the perfection.

[Timur]

6 hours ago, Hestia said:

S+ to Trafalgar? You've got to be kidding. That's a complete joke.

Well, Trafalgar is the only pollster to get a couple states in the Midwest (I'm not sure if they predicted Pennsylvania correctly) correct while the others got them wrong in the 2016 Presidential election...

[Hestia]

Just now, Timur said:

Well, Trafalgar is the only pollster to get a couple states in the Midwest (I'm not sure if they predicted Pennsylvania correctly) correct while the others got them wrong in the 2016 Presidential election...

Sure, they did fine. S+ tier? Hardly.

[Entrecampos]

Ranking	Pollsters	Trump	Biden
1	Trafalgar Group	47	48
2	AtlasIntel	49	51
3	Emerson College	45	53
4	Susquehanna	46	49
5	Siena College	41	52
6	Marquette Law School	43	47.5
7	Gravis Marketing	43	54
8	CNN/SSRS	44	51.5
9	YouGov	43.5	52
10	Morning Consult	41	54
11	Research Co.	42	50
12	Swayable	45	55
13	Fox News	44	48.5
14	Change Research	45	53
15	Ipsos	44	53
16	RMG Research	44	50
17	ABC News	39.5	56.5
18	Civiqs	47	51
19	SurveyMonkey	43.5	54.5
20	Redfield & Wilton Strategies	45	49
Wisconsin	Entrecampos	47.46	47.68

[Entrecampos]

Results for Wisconsin 2020 with my Presidential model, I understand that some pollsters are better for different types of elections, Trafalgar is very good for Presidentials, not this good in the others. In Portugal this also happens, Catholic pollster is very good in mayorals and not this good in general elections, for example.

Edited October 17, 2021 by Entrecampos

[Entrecampos]

Ranking	Empresa	Trump	Biden
1	Trafalgar	49	47
2	Wick	50	47
3	Emerson College	46	52
4	Susquehanna	47	46
5	Rasmussen Reports	47	48
6	St.Pete Polls	48	49
7	Siena College	44	47
8	Insider Advantage	48	47
9	Quinnipiac	42	47
10	Morning Consult	45	52
11	AYTM	43	45
12	Swayable	46	53
13	Data for Progress	48	51
14	Frederick Polls	49	51
15	Change Research	48	51
16	Ipsos	46.5	50.5
17	RMG Research	47	51
18	ABC News	49.5	47.5
19	SurveyMonkey	48.5	49.5
20	Redfield & Wilton Strategies	44.33	50.67
Florida	Intracampos	47.38	43.80

[Entrecampos]

For Wisconsin and Florida we can say, assuming that 98% votes GOP or DEM

49.1% Biden
48.9% Trump

50.9% Trump
47.1% Biden
 

Edited October 17, 2021 by Entrecampos

[Entrecampos]

I made some changes in my rankings. Pennsylvania 2020.

[Entrecampos]

Biden with -8.1% net approval with the actual model.

[Entrecampos]

Ranking	Empresa	DEMs	GOPs
1	WPA Intelligence*	44	42
2	Quinnipiac University	43.5	46.5
3	YouGov	42	37
4	Marist College	46	38
5	CNN/SSRS	45	42
6	Public Opinion Strategies	42	42
7	Lake Research Partners/Emerson College	45	41
8	McLaughlin & Associates	47.00	48.00
9	Echelon Insights	48.5	40
10	Change Research	47.5	45
11	The Winston Group*	46	43
12	Morning Consult	45	40
12	RMG Research	40	41
13
14
15
16
17
18
19
20
21
22
23
24
25
USA 2021 Generic Election	Entrecampos	51.47	48.53

[Entrecampos]