Grade Pollster: Tralfagar

[Liberalrocks]

Well?...

[Gass3268]

They are called Trashfalger for a reason.

[Alben Barkley]

There’s no need for all these threads. One thread: “Grade polls this year” with a big “F” should be enough.

[Buffalo Mayor Young Kim]

Look they just took everyone else’s results and added five to seven points for Trump. When the polls overestimated Trump that makes them look better, but it doesn’t validate their ‘methodology’.

[Chancellor Tanterterg]

F

[Horus]

D-

[Hammy]

B/C. Question their methodology if you will, but it's clearly been more accurate two presidential cycles in a row, so it wasn't just a one-off. They're certainly doing better than the "quality pollsters"

[Rand]

There’s no need for all these threads. One thread: “Grade polls this year” with a big “F” should be enough.

Every poll deserves its own thread, including Patinkin, Insider Advantage, Swayable, Qriously, Redfield & Wilton, Landmark, AtlasIntel, Leger, Research2000, Targoz, Gravis, Spry, Y2, Opinium, CardinalGPS, RMG, The Winston Group, Tufts, Harper Polling, Climate Nexus, Cygnal, Echelon Insights, SoonerPoll, Google Surveys, and GBAO.

[Ljube]

C

Not too bad.

[lfromnj]

I don't get this idea they just cooked up numbers in states like Michigan, in 2018 they showed Stabenow with a healthy lead and they actually slightly overestimated her.
Most of their 2018 senate polls were within reason although they did have that Junk GA poll.
They had a few junk polls especially outside the midwest but im giving them a C+.

Interested to see the social score idea although I don't think it works.

[Ⓐnarchy in the ☭☭☭P!]

B/C. Question their methodology if you will, but it's clearly been more accurate two presidential cycles in a row, so it wasn't just a one-off. They're certainly doing better than the "quality pollsters"

I don't get how people can see them do better than almost anyone else in several key states and just brush it off as a fluke. Clearly they're R-favouring but they have a pretty good record in key states like Florida and Wisconsin even compared to 538 A+ pollsters like NYT/Siena.

[GeneralMacArthur]

As I've said in other threads, it's fake polling.  It's literally one guy just adding points for Trump because he believes all the polls are underestimating Trump.  Just because he turned out to be right about that does not mean his polls are good or his methodology is good.

If Romney had won in 2012, would UnskewedPolls.com have been hailed as something admirable or scientific?  Again they added 5-7 points to every poll for Romney based on nothing more than partisan fantasies.  In 2012, they were wrong.

Trafalgar is only "right" because the pollsters haven't figured out how to capture the Trump phenomenon of turning out low-propensity voters.  When that issue is fixed, it will be with something scientific and demonstrable, not by simply taking real poll results and adding 7 points to the Republican's score because reasons.

[Coolface Sock #42069]

I don’t know what their methodology is or how sound it is, but they were clearly on to something, which is that traditional methods of polling do not work anymore.

[KaiserDave]

D.

Just because they happen to be closer to the truth than others does not a good pollster make. They make up pro-Trump numbers, they were lucky they were close.

[Non Swing Voter]

I think they might have potentially been on to  something with the "shy Trump voter" (i.e., liars) theory.  I think it's true that Trump people are more dishonest with pollsters or refuse to talk to them more often than non-Trump voters.  However, it seems clear to me that they were just pushing spin polls to help Trump.  They did have Trump winning Minnesota, which was as wrong as other pollsters that showed Biden tied in Ohio. 

[lfromnj]

I think they might have potentially been on to  something with the "shy Trump voter" (i.e., liars) theory.  I think it's true that Trump people are more dishonest with pollsters or refuse to talk to them more often than non-Trump voters.  However, it seems clear to me that they were just pushing spin polls to help Trump.  They did have Trump winning Minnesota, which was as wrong as other pollsters that showed Biden tied in Ohio. 

Their final MN poll was Biden +3, they did have a tied one in August though.

[Amenhotep Bakari-Sellers]

B, they did figure in the same day voting that other pollsters discounted and strictly went by VBM, and the shy Trump voter made the election much closer. They said all the swing states were within the margin of error, they polled PA, WI and NC right and polled MI and AZ wrong, but Trump was within the margin of error

[SN2903]

A because they were much closer to the result than all the mainstream pollsters. They are the new gold standard.

[emailking]

B/C. Question their methodology if you will, but it's clearly been more accurate two presidential cycles in a row, so it wasn't just a one-off. They're certainly doing better than the "quality pollsters"

I don't get how people can see them do better than almost anyone else in several key states and just brush it off as a fluke. Clearly they're R-favouring but they have a pretty good record in key states like Florida and Wisconsin even compared to 538 A+ pollsters like NYT/Siena.

Because they're reverse engineering the result. Maybe we should listen to them about the shy Trump thing but what they produce I don't see as a poll.

[Vaccinated Russian Bear]

Is there ANY evidence that they have been doing "reverse engineering the result" or "literally one guy just adding points for Trump" or is this just Americans beloved conspiracy theorizing?

[SInNYC]

As far as I know, they have not released any hints about how they do polling. Every decent teacher I've ever had graded me on my reasoning and not my final answer. So, I can only give them an INC.

FWIW, I do think there is a shy Trump voter effect, though I suspect this election was more about energetic Trump voters.