HomeLatest ThreadsGreatest ThreadsForums & GroupsMy SubscriptionsMy Posts
DU Home » Latest Threads » Forums & Groups » Retired » Retired Forums » 2016 Postmortem (Forum) » Poll Averages, people. I...

Mon Jul 25, 2016, 11:41 AM

Poll Averages, people. It's the opposite of DEMOCRATIC PANIC!!!

I predicted there would be a "bounce" prior due to the RNC convention. I even expected Trump to take a lead of a few points. So far, he hasn't, which is unabashedly good news.


http://elections.huffingtonpost.com/pollster/2016-general-election-trump-vs-clinton

This may change, and I still expect a Trump lead before our bump takes effect. But as of right now things are looking OK. The best he can do, apparently, is to get close to Clinton. Her upside is still (always) higher.


I will be releasing my own election model in a week or so, once I get the data ingest built.

11 replies, 1214 views

Reply to this thread

Back to top Alert abuse

Always highlight: 10 newest replies | Replies posted after I mark a forum
Replies to this discussion thread

Response to Loki Liesmith (Original post)

Mon Jul 25, 2016, 11:44 AM

1. Of course he was going to get SOME bounce. But he's losing, and Nate is making up his lead.

 

Reply to this post

Back to top Alert abuse Link here Permalink


Response to CrowCityDem (Reply #1)

Mon Jul 25, 2016, 11:59 AM

8. Let's be fair to Nate. His model is more sensitive to some kinds of swings than others.

Frankly I think he overfits his models a bit. Also, he uses mean values in his polls, which are more sensitive to outliers. Sam Wang at the Princeton Election Consortium uses a median-based approach instead which tends to disregard outliers.

But Nate isn't cheating. He is just reporting that outcomes his models predict.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Loki Liesmith (Reply #8)

Mon Jul 25, 2016, 12:07 PM

10. I wasn't specifically referring to today's snapshot...

 

but the one from the other day, where he had FL, OH, and PA all moving in Trump's direction, despite no new polling showing that happening. FL especially, had Hillary with the polling lead, but he turned it red based on a nebulous 'trend'.

Things like that call him into question, since he's not strictly analyzing data anymore. In that example, he's making it up.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Loki Liesmith (Original post)

Mon Jul 25, 2016, 11:46 AM

2. The one the media is pushing this morning is the CNN poll which shows a 4 point trump lead. What is

funny is that whenever the polls show a positive move for Hillary, it is hardly mentioned, but when a poll demonstrates a negative move for Hillary, it makes headlines.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to still_one (Reply #2)

Mon Jul 25, 2016, 11:50 AM

3. it's pointless to complain about the refs

Just gotta outplay the bias.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Loki Liesmith (Reply #3)

Mon Jul 25, 2016, 11:51 AM

4. I agree

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Loki Liesmith (Original post)

Mon Jul 25, 2016, 11:54 AM

5. There's ALWAYS going to be a post-convention bounce.

It happens to every candidate, on either side. So I'm not paying much attention to this one. But I'm mighty glad WE get to go second. We can beat that Trump bounce back down easily. The DNC convention will generate enough news that there'll be TONS more to talk about and report about than simply all the adverse stuff, distractions, enemy dirty-tricks and assorted attempts at rat-fucking, all of it. We've got some of the best speakers in the business, starting with a sitting President who's POPULAR, whose poll numbers are above 50% and rising (as opposed to what we saw last week - when the last two Presidents AND nominees on the other side didn't even want to show up), a convention with A-list guests and speakers and decorative, notice-me celebrities (as opposed to the B- and lower "luminaries" we saw last week). We have loads of popular name-politicians who WANT to be here (as opposed to what we saw last week).

So as the convention gets going, I expect - and HOPE - that there's plenty more news to be generated that's all positive, plus an upbeat tone to the whole thing, that will more than counteract any convention bounce the bad guys got.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Loki Liesmith (Original post)

Mon Jul 25, 2016, 11:55 AM

6. This is Trump's high water mark. He'd be more than doomed if trailing or even tied

 

at this point.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to geek tragedy (Reply #6)

Mon Jul 25, 2016, 12:06 PM

9. I still expect him to lead for a bit

Although I cannot prove it, I do believe that some media orgs poll and repoll until they get a sample that fits their narrative?

Why do I believe that? Because the changes in direction of leads in polls are too often correlated more than they should be assuming even random processes with some level of drift. One of these days I'll have enough data to demonstrate this.

But not today, sadly. So my gut says expect a Trump lead because the narrative demands it and some polling outfits are willing to keep polling until the result is what they expect.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Loki Liesmith (Original post)

Mon Jul 25, 2016, 11:56 AM

7. Surprised it isn't more. When you consider places like TYT

That a lot Dems look to for news, bashes Hillary all day everyday.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Thrill (Reply #7)

Mon Jul 25, 2016, 12:11 PM

11. Trump's numbers and favorably when up while Clinton's remained about the same.

I'm a Clinton supporter and watched TYT's RNC coverage all last week. I really didn't see it as Hillary bashing. But regardless of my opinion, the polling shows Trump's support and favorability increasing.

Reply to this post

Back to top Alert abuse Link here Permalink

Reply to this thread