2018 Midterm Elections Coverage

[sardia]

I’m watching the O Rourke/Cruz debate right now (because, you know, I’m in TX and am a voter) and Cruz is frankly pissing me off. He’s just basically ignored all the questions and all of O Rourke’s statements and returned to standard talking points. For instance, in the very first question about Dreamers, Cruz entirely ignored the part about Dreamers and takes about illegal immigrants in general — not focusing on the fact that they often were brought over by their parents instead of doing it themselves. They’re both walking all over the moderators, too — the moderators seem nervous, too.

Latest turnout projections are bad for Democrats. 538 assumes high turnout is better for Beto. The conversion from all voters to likely voters shows much lower turnout. Do you need a citation?

https://fivethirtyeight.com/features/el ... wer-votes/
Turns out that Republicans waste votes too.

Simply put, if you live in a district with fewer voters, your vote makes up a bigger share of the total — which makes your vote worth more — and it takes fewer votes to carry the district. So while Democrats may win solidly blue districts with relatively few total votes, GOP candidates are more likely to run up the score in reliably Republican districts, winning by huge margins. In both cases, all votes beyond the 50 percent mark are “wasted” (after all, no matter how much you win by, you can only win one seat per district), but the size disparity means that more votes are wasted in bigger, redder districts. But those votes still count toward the national popular vote, making it theoretically possible for Republicans to win the national popular vote but still lose majority control because Democrats got more bang for their buck — or, in this case, votes.

Tldr Democratic districts have more people, but only low% of them can or will vote. GOP districts have lots of voters, and lots of them vote, which wastes all the votes over 50%.

The effect size isn't as big as gerrymandering/self sorting, but it does fuel GOP arguments to redistrict by voters instead of total population. This is already baked into the model, but it'll provide fuel for future census/redistricting fights.

[iamspen]

It's hard to imagine a world in which current events won't significantly affect voter turnout, and if that is indeed the world we live in, my disappointment will be palpable. The Kavanaugh fiasco seems to be further withering the Republican party down to its Trumpist base, a trend which is surely disastrous for the long-term future of the party, regardless of how much weight it carries in these midterms.

Unfortunately, Democrats are having a hard time trading advantage of current events because, IMO, they're awful at telling blunt truths; nobody, for instance, directly called out Kavanaugh's blatant lies and asked why he was telling them. Dems' overall lack of courage and inability to stop being diplomatic in a political climate that demands it is frustrating and not doing much to further energize the base, especially the young vote, which is what they need the most.

[Tyndmyr]

It's hard to imagine a world in which current events won't significantly affect voter turnout, and if that is indeed the world we live in, my disappointment will be palpable. The Kavanaugh fiasco seems to be further withering the Republican party down to its Trumpist base, a trend which is surely disastrous for the long-term future of the party, regardless of how much weight it carries in these midterms.

Scandals have historically had fairly little effect on voter turnout. Reversion to the mean happens pretty fast.

The long term effects are probably limited to Kav himself. If he decides significantly differently on cases than another Republican justice might, that's a significant long term effect, but I highly doubt it will matter for voting.

[sardia]

Speaking of scandal, remember senator menendez corruption? Turns out he has a 3% chance of losing instead of a 0% chance. Remember, this is New Jersey in a blue wave year. And yet he still might lose a la Roy Moore. Democrats everywhere t should remember the price to pay for keeping a corrupt piece of shit in office. You might lose an easy election. https://fivethirtyeight.com/features/de ... nate-seat/
The fundamentals point to an easy win, but the polling points to a tight race.

[Tyndmyr]

Speaking of scandal, remember senator menendez corruption? Turns out he has a 3% chance of losing instead of a 0% chance. Remember, this is New Jersey in a blue wave year. And yet he still might lose a la Roy Moore. Democrats everywhere t should remember the price to pay for keeping a corrupt piece of shit in office. You might lose an easy election. https://fivethirtyeight.com/features/de ... nate-seat/
The fundamentals point to an easy win, but the polling points to a tight race.

Fair. It's a bit subjective. It's one thing if you've got a case like Moore where the corruption is really blatant and the candidate's integrity is directly challenged. A really bad candidate matters. But a gaffe, misspeak, or particular vote, generally not.

It would be pretty surprising for Jersey to go Red, but hey...the numbers are the numbers.

[sardia]

Moore is the pedophile, menendez is corruption. FYI.

Alright, the me too movement and/or sexism/Kavanaugh. Before the polling settles down, is it Good or bad for Democrats? For Republicans? My initial thought is metoo is going to cause a reckoning in the GOP. And yet... I wanna see more signs it's true. Timeframe is midterm elections, not next decade demographic stuff.

[Zohar]

According to an NPR poll from this morning, Democrats were already pretty motivated, so it's mostly Republicans who are being advantaged by this.

[Tyndmyr]

Moore is the pedophile, menendez is corruption. FYI.

Sorry, shouldn't have used corrupt in that context, twas unspecific. Meant it as morally corrupt.

Alright, the me too movement and/or sexism/Kavanaugh. Before the polling settles down, is it Good or bad for Democrats? For Republicans? My initial thought is metoo is going to cause a reckoning in the GOP. And yet... I wanna see more signs it's true. Timeframe is midterm elections, not next decade demographic stuff.

I'm gonna pull up an old 538 article with regards to this. https://fivethirtyeight.com/features/a-gaffe-can-matter-when-it-motivates-the-base/ They're arguing for gaffes mattering in some cases, but other cited examples(Romney and Obama) had no to a small effect on polling. Maaaaybe 1% at most. Even the one they're arguing for mattering, eh...Iowa is fairly red. It is highly probable that they'd have elected a Republican senator regardless.

So, what separates the case where they matter, from cases where they don't? The above two seem like significant moral failings of a candidate. So much so that even folks on their side of the partisan line must admit they are falling far short of any reasonable moral standard.

It might have an effect if Kavanaugh were running for office, since at least some appear to believe that he's morally a bit short of a full deck, but this belief seems less pervasive than the two examples listed above. And he's not directly running for office, it's merely people who happen to share a party with him. Gaffes don't seem to bleed over to this degree. I can't think of a comparative example to argue that there will be an effect. So, it shouldn't budge polling for either party.

My republican friends think it's important for republican turnout, but...honestly, republicans tend to enjoy good turnout regardless in midterms. Democrats ought to get more mileage from partisan turnout bumps...but pretty much all Democrats I know of are already reasonably incensed at Trump. It's hard to imagine this motivating someone that isn't motivated by opposition to Trump.

[sardia]

https://fivethirtyeight.com/features/wh ... nate-odds/
The Senate and the house have gotten 5% worse for the Democrats. Heidi and Beto will probably lose and are viable based off fundamentals. (Read fundraising). In his Twitter account, Nate wonders if Democrats are over donating, which makes fundraising a worse metric. He's not sure why the Republicans are doing better.

Overall, this is one of those times where the “what” is easier than the “why.” The what is that Democrats’ position has worsened in the Senate as a result of declining numbers in deeply red states — where, because of their terrible Senate map, Democrats have a ton of exposure. But the why is not totally clear: It might be attributable to Kavanaugh, or it might have been baked in all along.

[Tyndmyr]

North Dakota and Missouri are playing a huge part here. The former one I noted that I thought was more accurately represented by the worse odds a while ago(Mississippi also swung worse for democrats, so I think both of those quibbles worked out more accurately for me/lite model than the other two options). A 31 point swing is huge. Note that the swing was larger for polls + fundamentals, so while some error was baked into the polls, most likely, the difference is attributable to the fundamentals.

This is true in general. The lite model is looking significantly more accurate than polls + fundamentals.

My read on this is the same as before, polls + fundamentals appears to be double-counting effects to some degree(not actually double, but polls and fundraising are strongly dependent variables). I think they're merely overestimating the level of independence here, possibly due to partisanship being higher than it historically has been. That factor is throwing off some historical relationships, I believe.

[sardia]

North Dakota and Missouri are playing a huge part here. The former one I noted that I thought was more accurately represented by the worse odds a while ago(Mississippi also swung worse for democrats, so I think both of those quibbles worked out more accurately for me/lite model than the other two options). A 31 point swing is huge. Note that the swing was larger for polls + fundamentals, so while some error was baked into the polls, most likely, the difference is attributable to the fundamentals.

This is true in general. The lite model is looking significantly more accurate than polls + fundamentals.

My read on this is the same as before, polls + fundamentals appears to be double-counting effects to some degree(not actually double, but polls and fundraising are strongly dependent variables). I think they're merely overestimating the level of independence here, possibly due to partisanship being higher than it historically has been. That factor is throwing off some historical relationships, I believe.

Let's test this theory. West Virginia 3rd district is on average+37 GOP fundamentals. But the polling shows it as+2 GOP with some Democrats leading sometimes. Are you with the classic model, or with polls lite (only) model?

[Tyndmyr]

Seems reasonable enough. WV has supported Manchin for ages. Manchin is almost certainly going to hold his senate seat. He's a fixture 'round there.

They've been tradition "red" for presidential purposes, but the state as a whole is kinda purple depending on how you look at it. I don't have a ton of personal experience with WV, mostly just driving through it on the way to somewhere else, but my gut feeling is that they're overestimating the red partisan lean in the state. Yeah, it went really red for Trump, but that may not apply equally to other races.

So closer race feels more accurate there.

As a side note, the expert opinion also skews towards a closer race over the classic model. In most cases, the expert opinion increases the gap, not narrows it, relative to the classic model. Interesting.

[sardia]

https://fivethirtyeight.com/features/el ... our-model/
3 weeks to go until mid terms. Nate and Republicans should be worried about the Democrats ridiculous fundraising numbers.
GOP should worry because it and special election dates signifies 60+seat wave (unlike polling which indicate a small wave). Nate is worried because this number is so out of scope, which could lose predictive power.

PS governor's model is finally here. Illinois GOP governor is a serious under dog.

[Tyndmyr]

Definitely thinking small wave, based on polling, rather than the large wave that fundraising would predict. And largely a congressional wave. Hard to make headway in the senate. But it's interesting all the same.

As for governors, I agree that Illinois is probably gonna go blue. Chicago runs that state to a large degree, and the incumbent almost lost a primary challenge. Strong or successful primary challenges are, IMO, a very strong indicator of a weak candidate. Polling might tighten up some as undecided voters reluctantly break Republican, but it'll probably translate to a Republican loss regardless. I don't know enough about the candidates to know why the incumbent appears weak, but going by the metrics, it doesn't look good for him.

Florida's a more interesting race. It's a pretty notable swing state, and the odds are closer. There may be some long term effects if the Democrats can maintain/pick up additional governorships.