Excuse the multi post but this article also points out how much the divisions are less between states than within them.
I see they mention generational replacement as a force driving change, which is what I was referring to as a source of flipping in the 20th Century. Immigration created ethnic neighborhoods in big cities. Those Little Italys and Poletowns helped newcomers adapt to the New World. Their kids and grandkids, however, tended to assimilate and move into less defined areas, making it harder to predict their votes in the short term. Because they moved to all-white suburbs, the long-term effect coalesced along conservative lines, although that has been changing again in the 20th Century.
Partisan sorting is the modern inverted version of this dispersal, and an echo of the ethnic neighborhood. People are now deliberately moving into areas that have known political preferences, making them increasingly partisan.
That’s a known effect. Why it winds up with a balance, though…
I would put it down to three things:
-
Old politics. You could draw a smiley face on a cushion and it would get 30% of the vote if it had an R next to its name (or, to a lesser extent, a D). The “baked in” politics are that Democrats are basically communist and the economy always tanks under Democrat rule. It has always been the case that a large proportion of Americans believe this and can never be convinced to vote D.
-
New politics. A lot of people, perhaps the majority, are receiving news and opinions according to their past viewing preferences, and thereby becoming entrenched in their worldviews. At least 30% of people live in an alternate reality where Jan 6th was entirely peaceful, climate change is a hoax, it’s been proven that Joe Biden is corrupt and covid was a conspiracy between Fauci and the Chinese.
-
The true “silent majority” are people with virtually zero knowledge or interest in politics. And the “at a glance” factors somewhat shift Trump’s way. Yet another indictment does look like they are hounding Trump. And inflation does look like a fuck up if you are unaware of the levels of inflation around the world. And Trump just continues to garner more screen time.
Biden’s done a great job but he’s not as good at countering 1, 2 or 3 as someone like Newsom would be.
It would probably win a Republican primary in a landslide if it was a MyPillow.
I also want to clarify what I mean by “to a lesser extent, D”.
It’s the nature of two-party, first-past-the-post, politics that many people find themselves voting for the same party over and over, even if they are not entirely happy with the candidate or promised policies…if you have an ideological lean, you really only have one choice.
But I’d still like to think that for many people like me, if the Democrat candidate had broken the law repeatedly and with his words wanted to foment further violence and autocracy, then I’d grit my teeth and vote Republican.
Tax cuts for the rich are not as bad as the end of democracy.
Or, just not vote in that race at all (or, just not vote in that election).
Nah, for me, if one guy is sowing division and conspiracy theories, I’d vote for the alternative (assuming that they are not doing the same). It’s really important that they lose.
Heck, I’d probably go for Christie over RFK and likely lose some teeth from the degree of gritting.
Now that is a Hobson’s choice.
I can recall back around 2015 many D-leaning posters (myself sometimes included) occasionally pronounced something about like “I’d vote for the worst imaginable D over the best imaginable R”.
To which I’d comment “Don’t tempt Fate. The D’s can someday somehow anoint a truly awful candidate. They haven’t yet, but that doesn’t mean its impossible. And in fact what you’re advocating is exactly what causes R-leaning voters to vote Trump.”
If it’s Christie vs RJK Jr, count me in as a vote for Christie. Then take a long hot shower. And maybe ipecac.
Yes. Me too.
Conventional R beats wacko R wearing D clothing every time.
I ageee with some of your points, but disagree with your hypothesis for why the Democrats face the danger they do from not having a propaganda machine. The problem is that we no longer have a bell shaped curve of distribution of voters. Back in the days of LBJ, Ronald Reagan, and even as recently as Bill Clinton, moving to the center to win over moderates worked because there was a large number of moderates to be won over, and relatively few extremists on the fringe to lose.
For whatever reason, the moderate Republican no longer exists in large numbers, and so moving to the center wins a Republican fewer votes than it used to in the past. The Republicans now have to rely on getting out the MAGA vote. Trying to win over a few moderates that now mostly don’t exist is no longer a viable strategy for them (see Liz Cheney as a prominent example).
The flip side is that trying to radicalize the left won’t work for Democrats because a leftist version of MAGA is going to be smaller than the right wing version of MAGA and be a guaranteed loser every time. Trying to build a left wing MAGA would just lose Democrats the moderate base. That’s why a Democratic propaganda machine hasn’t been tried.
ETA. To tie back into the main topic, I suspect a large part of Biden’s low numbers are disappointed liberals. Presumably most of them will “come home” as we get closer to the election. Yes, a propaganda machine might have them more excited at this time, but it would come at the cost of possibly losing part of the moderate Democratic base.
Polls only reflect people reached by polls. I don’t answer any polls. The people watching Foxnews and the Weather Channel all day are the ones answering phones.
The sample size in the polls is also too small. There is probably a bias to some groups, but since the same people are sampled many times (they answered the phone) then any major shift in a group that is missed will not be reflected. Yet if the group sampled does change, then that trend does get captured. That data has a bit more meaning.
The effect of advertising or sensational news(“Hunter Biden’s Laptop found!”) seems to affect the last months most. Fall of 2024.
Who in the World Is Still Answering Pollsters’ Phone Calls?
What about cellphones? Finally, an easy one: We call cellphones and landlines! About three-quarters of our calls go to cellphones nowadays — including nearly every call to young people.
I read the NYT article. The author says they do take steps to ensure that the poll is “demographically and politically representative”, though not without also saying upfront that the death of telephone polling may be near.
Per the article, however, less than one in 200 dials results in a completed interview. If it’s gotten as bad as all that, I don’t see how polling can be anywhere near as reliable as it might have been fifty years ago, when most people probably would have been eager to particpate.
No question that the declining population willing to answer is a major problem.
Interestingly, younger people are more likely to answer unknown numbers.
One answer is to throw more resources at it. I’ve done phone-bank surveying. We were quartered in a very large room with many long tables of callers. We each had a computer and phone. Each number popped up individually and randomly from a database. I think I just pushed a button to make the call so that my fingers couldn’t hit a wrong digit. I had a very low percentage of completed calls even back then, but the total number of calls made up for it.
Election polling is an interesting branch of cold-calling. The eventual answer is known exactly and publicly. The audience for the polls can compare them to the real-world result and decide for themselves whether they were accurate enough to spend money on. So can the general public for that fraction of polls that are released to the news.
We can be sure that given the money at stake and the consequences of getting the right answers, thousands of professional pollsters are working every minute to determine new techniques that will produce more responses and better results. That may not happen. The times may have changed or the techniques might be too expensive or time-consuming to be practical. Or maybe someone will have a game-changing breakthrough in the future when something new comes along. I never bet against outcomes if getting there will make someone a fortune.
Speaking as a market research professional: sample size isn’t generally an issue for political polls.
You may feel like 1000 or 3000 people is a tiny sample, but if the sample is representative, that sort of sample size is large enough that the results will be within a few percentage points of what you’d see with a “census” (i.e., if you talked with everyone). Increasing the sample size to, say, 5000 or 10,000 respondents would actually not add very much additional accuracy or statistical power.
However, that “if the sample is representative” is the increasingly tricky part, and as noted, by yourself and others, many people are no longer willing to answer a call from a number they don’t recognize, or to participate in a poll. This is a big issue for the market research industry as a whole, but is exacerbated in political polling, as there are organizations that do “push polling” (conducting what appear to be political polls, but are actually biased campaigning calls).
Exactly so, and a big part of the issue is that online polling (which is likely where it’s going to go) isn’t necessarily demographically representative, either.
Young people do not answer the phone. Never.**
So the polling is not going to be accurate. If you only find one participant out of 200 dials , then you’re not sampling the general population.
In 1948, using the telephone was what caused the massive
blunder in the famous “Dewey beats Truman” headline.
Back then, many people didn’t use phones
And now, we’re back to a similar situation.
-
- (Cite: my family)
- (Cite: my family)
I hope your family was the source for the rest of your cites because they are equally untrue.
The 1948 mistake was caused by stopping the polling two weeks before the election on the presumption that Dewey had an unassailable lead. Roper in fact stopped two months early. In fact, many people made up their minds at the last minute, as has been found to be true in all later elections.
And many people used phones, since more than 60% of the population had one by 1950. That didn’t matter, since phones were not the source of interviews.
But the most profound change was in how the pollsters selected the people they questioned. In 1948, they used quota sampling, where interviewers were told to find so many young white men, so many elderly black women, and so on. The interviewers then picked the individuals, sometimes on street corners.
After 1948 pollsters switched to the more scientific probability polling, giving everybody a theoretical equal chance to be polled.
Also, the result of the 2020 election was 51% to 47% Trump. People tend not to admit they were wrong. This may happen more with Trump voters, but they will then not show up the third time to vote at all. They would never vote for Biden.
Are you talking popular vote? You’ve got the votes and the candidates reversed.