HomeLatest ThreadsGreatest ThreadsForums & GroupsMy SubscriptionsMy Posts
DU Home » Latest Threads » Forums & Groups » Retired » Retired Forums » 2016 Postmortem (Forum) » ***(CVOTER/UPI (moonies))...

Sun Nov 6, 2016, 03:24 PM

***(CVOTER/UPI (moonies)) national poll goes from Hillary +1.25% lead to Hillary +3.34% lead***



https://www.documentcloud.org/documents/3214801-US5th-Nov-Report.html

Yet her odds drop?



Note that Nate's adjusted lead went from +3 to +5, the sample size was larger and her average lead is just under 3%, thus this poll HAD to increase her odds, no? The WAPO poll did earlier. It's not logical that this dropped her odds. Can a statistician here please explain this to me?

6 replies, 823 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
Arrow 5 replies Author Time Post
Reply ***(CVOTER/UPI (moonies)) national poll goes from Hillary +1.25% lead to Hillary +3.34% lead*** (Original post)
Dem2 Nov 2016 OP
smorkingapple Nov 2016 #1
Stallion Nov 2016 #2
Dem2 Nov 2016 #4
Dem2 Nov 2016 #6
Doctor Jack Nov 2016 #3

Response to Dem2 (Original post)

Sun Nov 6, 2016, 03:26 PM

1. C+ Poll probably replaced an older higher rated one. His model thinks momentum matters too much

Reply to this post

Back to top Alert abuse Link here Permalink


Response to smorkingapple (Reply #1)

Sun Nov 6, 2016, 03:31 PM

2. The Data in Post Says Otherwise

Clinton posted better than last 4 polling periods

Reply to this post

Back to top Alert abuse Link here Permalink


Response to smorkingapple (Reply #1)

Sun Nov 6, 2016, 03:32 PM

4. His model has a massive inertia problem!

You call it momentum, I call it inertia, we are in agreement here, just using different words to mean basically the same thing. The swings are too wild compared to the actual polling averages seen at, say, Huffpo that doesn't use garbage polls.

The new poll should replace the poll from the same pollster, which you see the old poll's weight go to almost zero when replaced with a newer from the same pollster, thus is really should have given her a double boost - it's not logical!!!

Huffpo (4-way):

Reply to this post

Back to top Alert abuse Link here Permalink


Response to smorkingapple (Reply #1)

Sun Nov 6, 2016, 04:02 PM

6. A "D" rated poll of Michigan just upped her overall odds of winning the election by 0.4%

http://projects.fivethirtyeight.com/2016-election-forecast/updates/

It was also a +2 swing toward Hillary from the same pollster's previous release. I know state polls carry more weight, but his model does account for "a rising tide lifts all ships" model of national polls, that +2 improvement day-to-day in cvoter should have raised her odds.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Dem2 (Original post)

Sun Nov 6, 2016, 03:32 PM

3. Im sure 538 will drop Clinton by 5 more percentage points

Because it is more evidence of "uncertainty" in the polls or some bullshit

Reply to this post

Back to top Alert abuse Link here Permalink

Reply to this thread