Welcome to DU! The truly grassroots left-of-center political community where regular people, not algorithms, drive the discussions and set the standards. Join the community: Create a free account Support DU (and get rid of ads!): Become a Star Member All Forums Issue Forums Culture Forums Alliance Forums Region Forums Support Forums Help & Search

jimmy the one

(2,708 posts)
Tue Oct 29, 2019, 01:26 PM Oct 2019

Real Clear Politics Trump Approval Poll Avg, is badly skewed pro trump

The real clear politics (RCP) trump approval average is padded pro trump, in a couple ways.

Firstly, the RCP average includes Rasmussen, which is not quality polling, it cheats. Rasmussen persistently pads trump approval daily by 4 to 10 points (proof below), over the average of reputable polls.
Rasmussen is also included in every single RCP poll average. Rasmussen never drops out of the average, its padded results are always included. Most all other reputable polls are included only about half the time.
This is because Real Clear Politics Poll Average has a 'dropback' period of approximately 15 days, whereby polls which were conducted prior to the ~15 day dropback, are not included in the RCP average. IE: on October 15, all polls conducted before October 1 would be dropped off the RCP average, and only polls with a date between Oct 1 and 15th would be included. Since most all reputable polls are usually conducted only once a month or longer, their inputs can be excluded about half the time, from the RCP average.

Rasmussen however, is a 3 day tracking poll, where it averages its most recent 2 days results with the current days result, and then presents a 'brand new' poll every day, which RCP duplicitously counts as a daily new poll date which enables rasmussen be included in every RCP polling average, since rasmussen never drops out of the dropback. (rasmussen does not post Sat or Sunday, but this is only a 2 day delay which keeps it well within the 15 day dropback).

The net effect is the trump-padded rasmussen poll result is worth about twice what more reputable polls are worth, in the RCP average.
Since the RCP trump approval poll includes about 10 polls, rasmussens skewing does not greatly alter the average, but even so the RCP average should be adjusted by subtracting 0.5 to 1.5 points, due the rasmussen padding affect.
Harris, somehow linked to rasmussen with too often similar results, i have noticed cleverly posts a trump approval poll about every 18 days or so, which enables them to be included in the rcp average inordinately higher than more reputable polls such as gallup, quinnipiac, pew, Reuters, abc, cbs, nbc, msnbc, ppp, fox, cnn. Harris not as bad as Rasmussen tho.
I suspect there are more disreputable right wing polls other than just rasmussen, included in RCP avgs..

Rasmussen strategically pads its polls, a transparent attempt to avoid being cast as too partisan. It will pad trump while counter padding other republicans (generally in races either forlorn hopes or way out ahead).
Rasmussen was the worst of polls in 2000 bush v gore, having gwbush winning by 9 pts whereas gore won popular vote by 0.5 pts. It also had significantly higher results for most republican congressional races. Rasmussen realized then that it lost credibility &, spurred by gwbushs 'win', tightened up & produced some fairly accurate results in following presidential races. But do not be fooled, rasmussen is a right wing pro trump poll padder extraordinaire.

Nate silver's 538 pollster rating gives rasmussen a C+, which is also misleading as a C+ rating is in the lower 50% of all pollsters, and since rass shares the C+ with dozens of other polls it is not a good score. Rass is biased Republican by 1.5:

Five Thirty Eight - method - total - avg error- % corrrect - grade - bias
Rasmussen IVR/online ....... 711........ 5.3 ..... 78%.......C+ .. R+1.5
ABC News/WashPost Live...... 60 ....... 2.9 ..... 78% .........A+ .. D+0.6
Quinnipiac ....... Live......... 195 ........4.6 .......85% ......... A- .. D+0.1
NBC News/WSJ ... Live .........27 ........4.2 .......72% .........A- .. D+0.8


Note that 538 above relies heavily on whether a poll is conducted by a 'Live' person rather than online or C.R.A.P. polling as rasmussen is, which stands for Computer Responsive Automated Polling (ring ring, hello, 18? press 1 for trump 2 for hillary thank you bye), aka IVR or interactive voice response, and or online polling which is less scientific than live polling. https://projects.fivethirtyeight.com/pollster-ratings/
Categories can be stacked high to low or vv by clicking on the category at top;
Regret to report that online nbc survey monkey is not quality polling either, D- with 4.9 bias dem. It is not included in RCP polling avg, for obvious reasons, while rasmussen is thus RCP is complicit in duplicity.

Below you can see how high rasmussen padded trump during this 3 week interval, compared with other reputable polls:

RCP trump approval ..........samplesize. .......appr - disappr --- spread
Quinnipiac 9/19 - 9/23 1337 RV .................... 40 ....55... -15
Politico/Morning Consult 9/20 - 9/22 1994 RV ...41 ... 55....-14
Reuters/Ipsos 9/16 - 9/17 939..... RV .............. 43 ....55 ....-12
FOX News 9/15 - 9/17 1008......... RV .............45 ....54 .....-9
Rasmussen Reports 9/15 - 9/17 1500 LV ......51 ...48 ....+3
NBC News/Wall St. Jrnl 9/13 - 9/16 900 RV ........45 .. 53 .....-8
Economist/YouGov 9/14 - 9/17 1184 RV ............44 ...53......-9
Politico/Morning Consult 9/13 - 9/16 1994 RV .....43 ...53 ...-10
The Hill/HarrisX 9/11 - 9/12 1000 RV................47 ...53 .....-6
Reuters/Ipsos 9/9 - 9/10 ....992 RV ..................44 .... 53 ....-9
Gallup 9/3 - 9/15 1525 A .................................43 .....54 ...-11

https://www.realclearpolitics.com/epolls/other/president_trump_job_approval-6179.html#polls

In the RCP link above to trump approval ratings, note that rasmussen is included about every week or two over 1917, 18, & 19 (excepting oct, which hasn't ended yet when I bet they 'reconcile it in more'), whilst other polls included about every month.
Click 'all trump approval polls' or 'more polls' at bottom of list if all polls do not appear. Use your keyboard 'tools' at upper right to click on 'find', enter rasmussen to highlight all its polls. I listed all of trumps 3 years, & rasmussen was represented 129 times, quinn cbs & nbc about 50 or less.

5 replies = new reply since forum marked as read
Highlight: NoneDon't highlight anything 5 newestHighlight 5 most recent replies
Real Clear Politics Trump Approval Poll Avg, is badly skewed pro trump (Original Post) jimmy the one Oct 2019 OP
The other thing some of the others do is lean Republican Jarqui Oct 2019 #1
some bias is normal, & your solid opinion jimmy the one Oct 2019 #2
My feeling is that there is more than bias going on Jarqui Oct 2019 #3
If You Check Out Rass October Reports You Will See Trump is Underwater in Them Too Now Indykatie Oct 2019 #4
rasmussen, king of padding polls jimmy the one Nov 2019 #5

Jarqui

(10,110 posts)
1. The other thing some of the others do is lean Republican
Tue Oct 29, 2019, 01:34 PM
Oct 2019

throughout much of the election cycle.

Then, just before the vote, they tighten up closer to the result to mask how skewed they were through the cycle. This helps to cover their tracks.

I guess my statement is just opinion now but that opinion is based on some studies I did during the Obama years and I feel that I have continued to see with spot checking since.

jimmy the one

(2,708 posts)
2. some bias is normal, & your solid opinion
Tue Oct 29, 2019, 01:41 PM
Oct 2019

jarqui: the other thing some of the others do is lean Republican throughout much of the election cycle.

It's normal for a pollster to have a slight bias, it would be difficult to have a 0.0 bias, due to vagaries of polling.
It is when it's intentional bias that it becomes a problem.

Then, just before the vote, they tighten up closer to the result to mask how skewed they were through the cycle. This helps to cover their tracks.

Yes, rasmussen does this, a trait of unethical polling.

I guess my statement is just opinion now but that opinion is based on some studies I did during the Obama years and I feel that I have continued to see with spot checking since.

I think your opinion is solid.

Jarqui

(10,110 posts)
3. My feeling is that there is more than bias going on
Tue Oct 29, 2019, 02:00 PM
Oct 2019

Well beyond 0.8 bias.

Several pollsters are involved out of necessity - to have more of an impact on the poll of polls.

This muddies trying to reconcile with exit polls which used to seem to have some closer relevance than they do now.

And when you do that - mess with the polling, the results are exposed more to messing around with the voting machines, etc because it has been screwed with intentionally - to make it tougher to pin down.

When I looked at the polling for the recent Canadian election, I did not get the same feeling.

I wish I had more time for it. But I do not have a good feeling about it. I'm suspicious something sinister is going on.

Indykatie

(3,691 posts)
4. If You Check Out Rass October Reports You Will See Trump is Underwater in Them Too Now
Tue Oct 29, 2019, 03:25 PM
Oct 2019

You're right about Rass having a big Repub bias though. The other problem with RCP averages is that they selectively exclude some Polls in all their averages. The excluded Polls always include results more favorable to Dems.

jimmy the one

(2,708 posts)
5. rasmussen, king of padding polls
Thu Nov 7, 2019, 11:33 AM
Nov 2019

indy katie: You're right about Rass having a big Repub bias though. The other problem with RCP averages is that they selectively exclude some Polls in all their averages. The excluded Polls always include results more favorable to Dems.

I think a few other polls are right wing biased as well, pad their polls with impunity, tho rasmussen is king of the padders.
So RCP poll avg for trump is skewed right about a couple points, imo.
I'm gonna send my post in to snopes for their perusal...... watch this space.

Latest Discussions»General Discussion»Real Clear Politics Trump...