(C) Daily Kos
This story was originally published by Daily Kos and is unaltered.
 . . . . . . . . . .



Our BROKEN Polling System -- Senate Edition [1]

['This Content Is Not Subject To Review Daily Kos Staff Prior To Publication.', 'Backgroundurl Avatar_Large', 'Nickname', 'Joined', 'Created_At', 'Story Count', 'N_Stories', 'Comment Count', 'N_Comments', 'Popular Tags']

Date: 2022-11-13  

Do our polls really mean anything anymore?

There has been a lot of chatter about the general unreliability of polls in the past election cycles. In particular, this has concerned in this cycle with those slew of right-wing funded polls which flooded the market, particularly in the 4 weeks leading up to voting day which seemed designed to skew the poll aggregator averages and game the 538 algorithm.

Now that the results are (mostly) in, I wanted to see if that was true. So I took the 9 senate races that the aggregators thought were going to be closest (Pennsylvania, New Hampshire, Nevada, Arizona, Georgia, Washington, Wisconsin, North Carolina, and Ohio) and let the math do the talking. Specifically, I took the LAST poll that each organization released for each race up to one month before voting and compared it to the election result to see how accurate it ended up being (I should say that the vast majority of polls I looked at were released in the week leading up to Election Day). I then averaged the accuracy of each polling organization to see how they did across at least 4 of the races I looked at (I felt that lower than that the sample size was too small). Lastly, I gave each organization one of the following grades: Good (if polling average was within 2% of the result); Bad (if polling average was within 4% of the result); Absolute Garbage (if polling average over 4% of the result). The reason that only those polls within 2% of a result get a passing grade is that YOUR JOB AS A POLLSTER IS TO PREDICT RESULTS and if you cannot get within 2% of the result (which is a 4% spread), then you are not worth anything. With that said here are the results (in brackets are the number of polls that were averaged), while number represents the percentage deviation from the actual election result.

Insider Advantage (6) — 6.8% — Absolute Garbage

Trafalgar (9) — 6% — Absolute Garbage

Remington (5) — 4.2% — Absolute Garbage

Marist (5) — 2% — Good

Emerson (9) — 4.2% — Absolute Garbage

Suffolk (5) — 1.6% — Good

Siena (6) — 1.8% — Good

Fox News (4) — 2% — Good

You Gov (4) — 1% — Good

CNN (4) — 1% — Good

Fabrizio/Anzalone (4) — 2.5% — Bad

Data for Progress (7) — 4.3% — Absolute Garbage

So, of the 12 organizations that polled enough senate races, 5 are absolute garbage, 1 is bad, and 6 are good. And of the 6 organization that are either bad or absolute garbage, 5 of them are right-wing polls.

There is an additional point that needs to be made: other than the Marist poll, EVERY SINGLE OTHER POLLED ERRED IN FAVOR OF THE REPUBLICANS. That means that the percentage deviation was all in one direction (other than Marist).

Verdict — the popular opinion is correct — right-wing polls did flood the market and game the aggregators. The math says so. 538 and RCP should wake up to this fact.

[END]
---
[1] Url: https://www.dailykos.com/stories/2022/11/13/2135820/-Our-BROKEN-Polling-System-Senate-Edition

Published and (C) by Daily Kos
Content appears here under this condition or license: Site content may be used for any purpose without permission unless otherwise specified.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/dailykos/