Fun fact Luntz lived with McCarthy for most of 2021 which is um odd. The problem with Luntz is that his focus groups are always filled with people that support his narratives. It’s not organic at all. To get things so wrong one day before the election is embarrassing. I followed Michael Moores analysis- he got it mostly right.
538 wasn't much better. 51 or 52 Republicans in the Senate was tied for the mostly likely outcome. They predicted Walker, Oz, and Laxalt would win, but just slightly. I guess in fairness, they basically said it was a toss up with the Republican lean. Gave the republicans a 59% chance of controlling the Senate. In House, they said 227 republicans was the most likely outcome. Just looking around, they called Washington 3rd a "solid R" with a 98% chance of the Republican winning, for example... First time in a while that republicans under-performed their polling. They are normally winning races they not expected to win and and making races close that were supposed to be a blow out...
538 has some tweeking to do to their algorithm and sampling standard. Maybe some they sampled failed to vote, that's what Rick Scott is claiming.
The traditional polls were in the margin of error, which is enough to swing close elections. RCP and others added a fudge factor, bumping up Republican poll numbers, based upon the prior few elections. Turns out no fudge factor needed.
What these pundits, even the somewhat nonpartisan ones, miss is the attitude of the electorate toward what's going on generally in the country. It's easy to see why that's difficult to measure. Sure, there's inflation or abortion or immigration, but what about the atmosphere of animosity between the parties, which the average voter doesn't participate in? I feel strongly that the cause of it has been predominantly from right extremists, and I believe the average voter saw that, too. Had the GOP put up more moderate candidates, the outcome would have been different. I hope both parties, but especially the right, get the message: the voters want moderates who respect their oath of office and who will engage in respectful dialogue and compromise. Alas, too many see the world as black or white and believe their color perception is the only correct one.
The majority being 50.2%....I joke, but seriously we got a big job ahead making moderates even more attractive to a greater # of 'voters' to widen the winning %. Will Maga R's support/vote for a 'normy' R? Could the 'normy' Rs take independents voting for moderate Ds?
But you would think that would show up in the polling. As pointed out above, 538 does apply a bias factor based on how each poll performed in the past, but I think it is generally small. If the bias is large, they just exclude the poll from they averages. The pollsters are not getting a good cross section of voters. It's probably related to hardly anyone answering the phone to "unknown caller" anymore. Not sure how pollsters can fix this...
When respected conservatives are called RINOS by the MAGA cult - and that’s really what it is - there’s little hope for the GOP to become more moderate. I wish that weren’t so. We need both parties working together for the good of the people, and compromising on what each means by “good”.
Given the two significant outliers; Trump and Abortion; it’s no wonder their models failed. Modeling hates rare events. I got into the impact of “Balck Swans”; b4 the book; when I was taking regression at UF and reading “Foundation”. One enabled me to understand the point of the story. Even in a universe with 10k years of detailed data on human behavior, economic events, wars etc; a once in a billion personality with extra human capabilities that could not be modeled; screwed up a model that hadn’t been wrong in a millennium. I always felt the personality(The Mule) was a nod to Hitler.
Those voters with student debt were bought and paid for. And apparently women love aborting babies more than we knew.
I'd actually argue that their model is crediting too high of a likelihood on a race like WA-03 (in general, not speaking of the specifics here). If you have 50ish heavy lean districts like that, you should see one flip. I'd bet there are probably about 100 such districts and only one flipped, which is the model over-predicting low probability events.
The base for each party will remain loyal, the battle for new voters is among the independents/moderates.
