Log In
Create Account
SlickerTalk
Search Archives
The Leaderboard
The FAQ
Login
Create Account
Search
Dr. S. Talk
TT/ST Wiki
How Well Do You Know ...
RSS Feed
Hosting by DigitalOcean
Support ST on Ko-Fi
Links Only
50 Results
100 Results
250 Results
500 Results
1000 Results
5000 Results
2 Weeks
2 Months
6 Months
1 Year
2 Years
5 Years
All Time
Live
Down to Post
Backboards:
Live
________________
1: Dec 2, 17:22
2: Dec 2, 11:48
3: Dec 2, 08:21
4: Dec 1, 17:33
5: Dec 1, 11:23
6: Nov 30, 15:54
7: Nov 30, 09:41
8: Nov 29, 16:44
9: Nov 29, 08:01
10: Nov 28, 16:19
11: Nov 28, 09:42
12: Nov 27, 18:07
13: Nov 27, 12:04
14: Nov 27, 08:26
15: Nov 26, 18:06
16: Nov 26, 12:05
17: Nov 26, 08:29
18: Nov 25, 18:33
19: Nov 25, 11:12
20: Nov 25, 07:08
Posts: 155
Apropos of polls—they’re doing even worse than in 2016 so far this year as per 538 -- (link)
Posted by
JackDawson (aka dawson)
May 19 '20, 06:55
(No message)
https://fivethirtyeight.com/features/weve-updated-our-pollster-ratings-ahead-of-the-2020-general-election/
(fivethirtyeight.com)
Responses:
I cut them slack with how weird the primaries were this year. And primary elections are different animals than "pick 1 out of 2" generals. -- nm
-
Reagen
May 19, 06:57
19
Oh well if you cut them some slack then no problem! -- nm
-
JackDawson
May 19, 06:58
18
Don't be weird. -- nm
-
Reagen
May 19, 06:59
17
It’s no answer to say you cut them slack because primaries were weird this year — that’s my point isn’t it? Polling has had worse and worse track
-
JackDawson
May 19, 07:02
16
I'm saying that polls that have 3 to 6 viable options *by nature* have higher variability than ones that have 2
-
Reagen
May 19, 07:08
Is it your belief that we should never use polling as a data point? -- nm
-
David
May 19, 07:04
14
“Use ir as a data” point for what though? Settle political debates? Decide policy matters? Make ourselves feel better or worse about a potential elect
-
JackDawson
May 19, 07:09
12
Polls I'd like to see: Some of the more reliable "red" states like Nebraska and Arkansas
-
Reagen
May 19, 07:48
5
I’d love a high tide lifts all boats effect but I continue to believe at this moment in time that demographics will rule the day
-
JackDawson
May 19, 08:02
4
The biggest "remains to be seen" is just how much of 2016 was people being done with the Clintons. -- nm
-
Reagen
May 19, 08:09
3
Look at Michigan’s primaries. Polling in 2016 had Hillary up almost 20%. She lost. -- (link)
-
Tim
May 19, 08:19
2
And then in March, 300,000 more people voted in what was a 2 person race by then. Bernie got fewer votes. -- (link)
-
Tim
May 19, 08:21
1
This short circuited the brains of a lot of BernieBros, claiming the DNC rigged it. No, people just really fucking hated Hillary.
-
Tim
May 19, 08:22
I would say that the value of these polls is to point to areas/states of interest/attack (IOW, Dems should absolutely be putting $ into TX and GA)
-
Reagen
May 19, 07:14
5
To use stats in a wholesome way I’d argue for a demographic approach based on population trends and a combination of polls opinion surveys and long te
-
JackDawson
May 19, 07:19
2
I think a big error people made in interpretation last time was looking only at margins and not actual support numbers
-
Reagen
May 19, 07:30
we largely agree! -- nm
-
Reagen
May 19, 07:25
Again, without fully understanding how their accuracy is increasingly waning — 2016 you saw the same exact arguments about Texas that led to incorrect
-
JackDawson
May 19, 07:17
1
But if you look at the charts, the errors in the primary polls are wildly out of scope for the errors in the general
-
Reagen
May 19, 07:24
Certainly not -- nm
-
JackDawson
May 19, 07:06
Post a message
top
Replies are disabled on threads older than 7 days.