Were the Polls Wrong
As usual I am reading a lot about how the latest shocking election outcome shows the polls were wrong and that polling has become unreliable (no links I’ve been reading this on Twitter). This is not a new assessment of the polls. It was widely argued (again Twitter now I cite Yglesias by name) that the polls were probably wrong.
The alleged errors are exact opposites. Before the election, it was widely (to universally) asserted that polls implied an over optimistic assessment of Democrats’ chances (in particular to hold the Senate). I became very aware of this compulsively comparing different fivethirtyeight forecasts
Note here Steve Benen “In the runup to Election Day 2022, Republicans were optimistic about retaking the upper chamber, and forecasts suggested it was the most likely outcome. “
The forecasts (sic) to which he refers is the deluxe fivethirtyeight forecast which gave the Republicans a 59% chance to take the Senate. Now, first of all, an outcome with rated probability 49% should not be a shocker (and it is clear that the Senate outcome shocked a lot of people). But also the deluxe forecasts are made combining “expert” forecasts with the classic forecasts which are based on “polls, fundraising, past voting patterns, and more”. The classic forecast gave the Republicans a 51% chance of winning the majority. The polls only forecast gave them a 50% chance. The polls weren’t wrong (at least not as processed by fivethirtyeight which weights pollsters based on past performance and removes something which is a weighted average of an estimated house effect and zero.
Similarly, the polls said 2016 was too close to call (or at least fivethirtyeight gave Trump a 28% chance) and showed Brexit to be an almost exact tie as it was. In each case, the conventional wisdom was not based on polls, was held with great confidence, was wrong, and lead to little (not zero this time) self criticism.
The people dismiss polls and claim to know better before the election, turn out to know less well than the pollster, and blame the polls for their error.
I am getting tired of this.
Robert:
What is the Mid Term history for first term presidents? What happened during Obama’s first term?
Even so, Biden was able to accomplish more than Obama in his first two years of the presidency with a far weaker Senate which will still exist in his second two years. The polls as I saw them were optimistic in Repubs retaking the Senate and the House.
It also appears as one writer put it, there were more polls favoring Republicans than Democrats. What does 538 show as far as Biden’s polling popularity? The common meme here in Arizona is inflation caused by Democrat policy. Much of the supply chain issue is a rerun of 2008 (I was chasing semiconductors then).
the 538.com polls only forecast estimated the probability of a GOP takeover of the Senate to be 50% (not roughly 50% that is the number on the website). Considering history fundraising and a bunch of stuff (but not “expert” forecasts) the probability was estimated to be 51%.
a simple average of polls is highly highly vulnerable to (hypothetical) deliberately biased polls designed to influence the narrative. Fivethirtyeith.com polls only estimates are less vulnerable (because they estimate pollster quality and partially correct for estimated pollster house effects).
Fivethirtyeight.com is (as noted by Nate Silver) partially vulernable to an effort to influence the narrative with (hypothetical) deliberately biased polls. Years ago, Silver noted that , based on the outcome of the election, the null hypothesis that Rasmussen was not biased was rejected and standard significance levels (standard hypothesis testing not a mere hypothesis)
I find myself pretty glad that the polls were wrong, to the extent they were.
It looks unlikely that the Dems will control the House, and real control of the Senate requires 60 votes, which the Dems don’t have. But at least they will control committees and do one investigation for every one initiated by the GOP in the House.
It is often said that responders lie to poll-takers, dontchaknow.
The polls weren’t wrong. I am glad that the pundits were wrong to the extent they always are. I am also glad that, since 2004, I have refrained from making my personal forecasts public.
In any case, don’t all ordinary readers prefer optimistic polls over pessimistic ones.
One could understand why professional pols might make good use of ‘harsh’ polls, but for true believers they make unpleasant reading.
Fred:
If you told everyone, we have the unofficial results (which I saw yesterday for Maricopa County, AZ) would anyone continue reading the Polls? At this point all they are doing is confirming the vote. It makes for great anticipation and news.
We’re in the denouement phase just now.
It is universally assumed that optimistic polls are useful to those they favor (band wagon effect). However, in an evenly split country (which split was clearly indicated both by the polls and the outcome of the elections) a pollster who aims to please has no clue as to how to please the majority.
Robert
they don’t aim to please the “majority.” They aim to please their sponsors.
The real issue with polls is who shows up. Polls vastly underestimated the deplorables showing up in 2016 and to a lesser extent in 2020. The Senate was saved by somewhat less turnout among the deplorables— and some non deplorable Republicans and Independents voting for the Dems.
Part of the motivational aspect of polling is to get voters to go vote, of course.
There was however that whole ‘Red Wave’ nonsense that was pushed by GOPsters, that their victories in the House and Senate were imminent & unavoidable. Turned out to be only an appendix to the Big Lie.
Five Takeaways From a Red Wave That Didn’t Reach the Shore
NY Times – Nov 9
… The Democratic base showed up.
The biggest question hanging over Democrats all year was just who, exactly, would show up to vote for them. In a typical midterm election, like 2010 and 2014, turnout drops by about 20 percentage points from a presidential year. …
That might help explain why polling failed to capture the widespread feeling among Democrats, which grew after the Supreme Court’s reversal of Roe v. Wade and the Jan. 6 hearings over the summer, that their core democratic rights were increasingly at risk. …
At least part of the problem is poor use of available data. For instance:
– 538 (I believe) offered evidence that presidential popularity is a poor predictor of mid-term election results. How often did were Biden’ approval numbers cited by election pundits, compared to the number of times pundits notedthat presidential approval may not be a useful metric?
– How often were we told that the president’s party loses seats at the mid-term, without any mention of coattails during the presidential election? Clearly, coattails are a big factor in mid-term losses for the president’s party, but few members of the chattering classes point that out.
– Generic ballot results swung back and forth between Republicans and Democrats, a sign that swing voters were not sure who they’d vote for. The story we were given was that there was a final swing toward Republicans, without much mention of the tentativeness of that swing.
– The gerrymander mattered, but mostly went unmentioned.
These are all due to poor understanding of the number, not bad numbers. “Experts” failed us far more than pollsters.
macroduck:
Biden is like a fish out of water. Lacks charisma like Obama. Although, he did many things Obama could not accomplish to help people beyong the ACA. I do not fault Obama extremely for such. He lacked the experience to do such.
Biden came through with many programs to help the little guy. Chief amongst them is healthcare at an extremely low cost or nothing at less than 200% FPL. He eliminated the gap between 100 and 138% FPL. 200 -250% FPL has low premiums. If you are above 600% FP, your premium is capped at 8.5% – no cliff anymore. Forty states now have Medicaid expansion. Maternal healthcare in these states has been lengthened to 1 year.
He was the right president for this downturn. Many people did miss this.
which goes to show that “statistics” are unreliable predictors of events determined by unknown or non-repeatable factors.
in this case some of the “known” or knowable factors were ignored, as pointed out here by others…the big one being that polls are not exactly honest, due to the fact that politicians know that polls affect elections, so they are motivated to look for fudge factors.
all polls tell you is how well your lies are working at the time the poll is taken. it looks as if at least some voters have learned to lie to the pollsters. me too, sorta after i got the impression that the pollster was “pushing” or at least asking nonsense questions. i didn’t exactly lie, but i refused to answer the nonsense questions.
The point of the ‘Red Wave’ campaign from the GOP seemed to be to motivate their base to vote, and to suggest to the Dems that ‘resistance is futile’ (so don’t bother to vote, you losers.)
That was probably not a well thought out strategy. But it may have played on polls indicating that Biden was not exactly popular, which the GOP thought would be advantageous to them.
After denialism faltered in the midterms, Republicans seek distance from a movement they unleashed
Boston Globe – Nov 17
Relax
Kari Lake is still out there.
Crying in the Wilderness.
Trying hard to sit at Trump’s side in 2025.