Protected: Two ways to estimate primary outcomes without polls (transcript)
This content is password protected. To view it please enter your password below: Password:...
Virginia is for statistics lovers
Today I head for Longwood University, site of next week’s vice-presidential debate, to talk about election math. If you’re anywhere near Farmville, come out and see what a great temperament I have. Tonight, Jarman Hall, 7:00pm.
Topics:
1. Is it going to be recorded/podcasted?
2. It appears that Huffpollster (and therefore PEC) excludes landline-only polls, but includes internet-only polls. I understand their argument for excluding landline polls (and yours for just including everything). But what is the argument for including internet-only polls while excluding landline polls? Is that based on a statistical analysis of how reliable results have been?
AAF, I second your motion for recording all Dr Wang’s lectures. Perhaps a team can form to document his nonbroadcast appearances.
Today I head…..today,. Welcome to the Department of Redundancy Department. (Just a little joke for a college professor too busy to proofread.)
This debate should provide a good check of the response time of the system (ie, impulse response, or Green’s function).
I remember thinking after the first Obama-Romney debate that the dip was too sudden, how could a system of 100’s of million people respond so quickly? But if we see a similar time structure here, then it is clear.
One confounding point is that the MM had already started to creep up before the debate. The effect of the debate would be easier to discern if it had stayed flat for a couple of more days.
Convention season and pneumonia weekend suggest to me that national polls react to events as soon as they fall into the survey period. But there’s often a delay of a couple of days before they come out.
The media seem to be fascinated by tales of undecided voters who think Trump won the debate, probably for man-bites-dog reasons.
The number of undecideds is definitely higher in 2016, but the way they break should be similar to other years. Is there any data showing undecideds are different this year?
Will there be a distaff Bradley effect? Are women in GOP households afraid to state their support for Clinton in earshot of the boys?
There’s been a study on this and it appears to be true. Female senate and gubernatorial candidates outperform their final polling averages by ~2.7%. Study here: http://link.springer.com/article/10.1007/s11109-010-9137-6
Chart from The Bad Place: http://fivethirtyeight.com/features/gender-might-matter-in-this-election-but-other-factors-probably-matter-more/
I personally know two women who have said, “Don’t tell my husband, but I’m voting for Hillary.” Is there a significant number of these female voters? We have no way of knowing. I would like to think so, but it’s not something we can count on.
That’s been my theory for some time. I guess maybe we’ll find out.
Or are “boys” in Democratic households afraid to state their support for Trump in earshot of their wives – lest they risk their marital relations?
It’s dishonest that many sites are labeling their online, opt-in, self-selecting “Who won the debate?” voting schemes as “polls.”
You are a most welcome antidote to the click-bait shills of the world masquerading as fact-based analysts. And by that, I mean Nate Silver and his thumb-on-the-scale “trend” lines. I am embarrassed that I ever followed him.
Welcome to our beautiful Commonwealth! Norfolk is a bit too far for a road trip.
Nate Silver on Twitter ” Good news for Democrats: Really hard to see how that debate helped Trump.
Bad news: If it somehow did help Trump, maybe no stopping him”
Top notch data journalism.
See Deadspin: “I Think Nate Silver Is Broken, Maybe“
They just had larry sabato from U of V on CNBC….interesting discussion to say the least….the anchors were saying that 66% of CNBC viewers said Trump won!…when Sabato said that didn’t mean anything and what one would expect with that crowd…the anchors were dumbfounded, except for John Harwood…..pretty comical….
Just read an article that said more than 80 million tuned into the debate, plus even more via streaming internet.
And that over the course of 90 minutes, people stated tuned in and there wasn’t a big drop off of viewership.
If that is true, I believe it will be a huge boost to Clinton, as she started to really expose him towards the tail end of the debate.
Hopefully the polls will bear this out in the coming days.
Sam,
Will your presentation slides be available for viewing?
Prediction Markets Score the Debate a Rout: Clinton Over Trump
“All told, Mrs. Clinton’s odds of winning the election rose from around 63 percent in the minutes before the debate started to 69 percent by the end.”
http://www.nytimes.com/2016/09/28/upshot/prediction-markets-score-it-a-rout-clinton-over-trump.html?partner=rss&emc=rss&_r=0
Not statistics, so I know I shouldn’t but … unbelievable from The Arizona Republic and so well-written. Plain speaking at its best.
http://www.azcentral.com/story/opinion/editorial/2016/09/27/hillary-clinton-endorsement/91198668/
BREAKING!
Election guru extraordinaire, David Plouffe says Trump has no viable path to 270. Media is imagining contrived horse race.
http://www.bloomberg.com/politics/videos/2016-09-27/david-plouffe-on-why-donald-trump-can-t-get-to-270