Have you noticed 538's recent Senate forecast trend?

[Anonymous]

Anonymous wrote:In the past few days, the forecast on 538 has gone from less than a 15% chance of Democrats taking the Senate to 19% today. It had been relatively stable before.

Similarly, the House forecast has trended in the Democrats' favor over the same short period (84% to 88%), after a long period of stability.

I wonder if this is a sort of polling wobble at the last minute, or if this trend indicates something significant.

Thoughts?

I still think the GOP will hold the Senate.

[Anonymous]

Anonymous wrote:In the past few days, the forecast on 538 has gone from less than a 15% chance of Democrats taking the Senate to 19% today. It had been relatively stable before.

Similarly, the House forecast has trended in the Democrats' favor over the same short period (84% to 88%), after a long period of stability.

I wonder if this is a sort of polling wobble at the last minute, or if this trend indicates something significant.

Thoughts?

I think that there's probably a systematic error in favor of Democrats in most of the polls, but that the trend in Democrats' favor is probably the result of the mail bombs, Pittsburgh and the yoga class attacker. I think there are plenty of conservative voters in America who oppose violence; they've just assumed that Democrats have been hyperventilating for nothing about Trump's rhetoric. Now, they see the rhetoric actually leads to problems.

[Anonymous]

If you look at the NYT’s Upshot project on polls, you can see it’s a huge struggle for them to get people to answer calls.
https://www.nytimes.com/interactive/2018/upshot/elections-polls.html

It seemed to take them 15-40k calls to get just 500 people to answer each poll. That response rate means only about 1 out of every 50 calls got a response. And you can rightly ask what sort of person has time and inclination to answer an anonymous phone poll, and what that says about the quality of any polls results.

[Anonymous]

Anonymous wrote:

Anonymous wrote:In the past few days, the forecast on 538 has gone from less than a 15% chance of Democrats taking the Senate to 19% today. It had been relatively stable before.

Similarly, the House forecast has trended in the Democrats' favor over the same short period (84% to 88%), after a long period of stability.

I wonder if this is a sort of polling wobble at the last minute, or if this trend indicates something significant.

Thoughts?

I still think the GOP will hold the Senate.

There are a bunch of articles out there that say Republicans really need to not just hold the Senate, but also extend their lead by 3+ seats, because that’s the only way they’re going to survive the 2020 election cycle when there are something like 21 Republican senators up for election. If they merely break even, or just gain one seat, then there’s a good chance they lose control in 2020.

[Anonymous]

I’ve said all along that the Dems will be up 50-49 after election night. The MS runoff will be key. If McDaniel is the runner up to Espy, it’s Espy versus McDaniel with Espy, a moderate, probably winning that race. If Espy loses, were at 50-50 for the 2019 senate. Then ... the potential party switchers.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:In the past few days, the forecast on 538 has gone from less than a 15% chance of Democrats taking the Senate to 19% today. It had been relatively stable before.

Similarly, the House forecast has trended in the Democrats' favor over the same short period (84% to 88%), after a long period of stability.

I wonder if this is a sort of polling wobble at the last minute, or if this trend indicates something significant.

Thoughts?

I still think the GOP will hold the Senate.

There are a bunch of articles out there that say Republicans really need to not just hold the Senate, but also extend their lead by 3+ seats, because that’s the only way they’re going to survive the 2020 election cycle when there are something like 21 Republican senators up for election. If they merely break even, or just gain one seat, then there’s a good chance they lose control in 2020.

Yes, but holding it, even at 51-49, will let them get many judges confirmed. They could get some through at 50/50, but Murkowski will say no to others.

[Anonymous]

Anonymous wrote:

[Anonymous]

Anonymous wrote:I’ve said all along that the Dems will be up 50-49 after election night. The MS runoff will be key. If McDaniel is the runner up to Espy, it’s Espy versus McDaniel with Espy, a moderate, probably winning that race. If Espy loses, were at 50-50 for the 2019 senate. Then ... the potential party switchers.

I think 50-49 is possible but not in any scenario that involves party switchers and Mississippi. I want some of what you've been drinking!

[Anonymous]

Anonymous wrote:

Anonymous wrote:

That was not the final prediction on election night. It was 71/29. Which is a third of a chance. That isn't terrible odds. And they were right about the popular vote she won that by about 2 points.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

LOL

Based on 538's past performance, I will predict R hold both houses.

Final results:

R 56
D 44

R 219
D 216

[Anonymous]

Anonymous wrote:Polls all rely on phone calls, right? So how do they sample young voters? I have 3 kids over 18 and they all told me they never answer their cell phone if they don't know who's calling them.

I wish I could find the article, but I had read the other-day that pollsters had to make 300 phone calls to fully poll 1 person.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:Polls all rely on phone calls, right? So how do they sample young voters? I have 3 kids over 18 and they all told me they never answer their cell phone if they don't know who's calling them.

My understanding is that they weight the responses they get. So (as an overstated example) if they know the population age 18-28 is 20% of the final vote count in most elections, and their polling sample has only 10% of respondents age 18-28, then they’ll double-count their responses to make up the difference. I suppose another way to handle it is to just poll more until you get enough young voters.

But every poller might take a slightly different attitude as to how to calculate the “normal” percent of each age group in each election, and how best to make up the difference. So that explains why different polls get systematically different results (eg Rasmussen skews heavily Republican because of some undetermined aspect of heir “secret sauce”).

It is not just the younger generation that don't answer their phone to a number that is not recognized. I routinely do not answer the phone whether my cell phone or my landline. Not sure how this is factored into the polls.

Unless I'm expecting a call from a student's parent or doctor's office, I never answer an unfamiliar number.

[Anonymous]

Sadly- I think the GOP will win big tonight.

But I would gladly be wrong on this.

[Anonymous]

THIS ELECTION WILL TELL ME WHICH STATES I WILL NEVER LIVE IN

[Anonymous]

Can someone explain why 538 has changed their forecast from 9% chance of Republicans keeping the House to 27% chance in the past 15 minutes?