Rasmussen Polling is suddenly deeply suspicious to me
Perky
(1000+ posts)
Send PM |
Profile |
Ignore
Wed Dec-19-07 09:07 PM
Original message
Rasmussen Polling is suddenly deeply suspicious to me
Edited on Wed Dec-19-07 09:16 PM by Perky
Help me out here. Look at the other pollsters and the consistency of the sample size and the thred on the recent order in Iowa Now Look at Rasmussen Much larger polling samples, but inconsistent quantities and CLinton leads in every one. Poll Date Sample Obama Clinton Edwards Richardson Biden Spread RCP Average 12/07 - 12/17 - 30 26.7 23 7.6 4.2 Obama +3.3 American Res. Group 12/19 - 12/23 600 LV 10 31 20 1 2 Clinton +11.0 American Res. Group 11/26 - 11/29 600 LV 27 25 23 4 8 Obama +2.0 American Res. Group 11/10 - 11/14 600 LV 21 27 20 12 5 Clinton +6.0 American Res. Group 10/26 - 10/29 600 LV 22 32 15 7 5 Clinton +10.0 American Res. Group 09/26 - 09/29 600 LV 24 30 19 10 3 Clinton +6.0 American Res. Group 08/26 - 08/29 600 LV 23 28 20 13 1 Clinton +5.0 American Res. Group 07/26 - 07/30 600 LV 15 30 21 13 3 Clinton +9.0 American Res. Group 06/26 - 06/30 600 LV 13 32 29 5 2 Clinton +3.0 American Res. Group 05/23 - 05/26 600 LV 11 31 25 8 3 Clinton +6.0 American Res. Group 04/27 - 04/30 600 LV 19 23 27 5 6 Edwards +4.0 American Res. Group 03/19 - 03/22 600 LV 16 34 33 1 2 Clinton +1.0 American Res. Group 01/29 - 02/01 600 LV 14 35 18 1 2 Clinton +17.0 Rasmussen 12/17 - 12/17 775 LV 27 31 22 9 5 Clinton +4.0 Rasmussen 12/10 - 12/10 1106 LV 26 29 22 7 5 Clinton +3.0 Rasmussen 11/26 - 11/27 1156 LV 25 27 24 10 4 Clinton +2.0 Rasmussen 11/12 - 11/12 1239 LV 24 29 25 10 3 Clinton +4.0 Rasmussen 10/10 - 10/14 1007 LV 21 33 22 7 4 Clinton +11.0 Research 2000 12/18 - 12/20 400 LV 22 10 22 1 1 Tie Research 2000 11/12 - 11/14 400 LV 25 27 21 10 4 Clinton +2.0 Research 2000 07/23 - 07/25 400 LV 16 22 27 11 3 Edwards +5.0 Research 2000 05/14 - 05/16 400 LV 22 28 26 7 2 Clinton +2.0 Strategic Vision (R) 12/08 - 12/10 600 LV 33 25 24 4 4 Obama +8.0 Strategic Vision (R) 11/30 - 12/02 600 LV 32 25 25 3 5 Obama +7.0 Strategic Vision (R) 11/23 - 11/25 600 LV 29 29 23 6 4 Tie Strategic Vision (R) 11/09 - 11/12 600 LV 27 29 20 7 5 Clinton +2.0 Strategic Vision (R) 10/12 - 10/14 600 LV 23 28 20 9 6 Clinton +5.0 Strategic Vision (R) 09/21 - 09/23 LV 21 24 22 13 4 Clinton +2.0 Strategic Vision (R) 08/17 - 08/19 600 LV 22 21 23 14 5 Edwards +1.0 Strategic Vision (R) 06/22 - 06/24 600 LV 21 20 26 11 4 Edwards +5.0 Strategic Vision (R) 05/18 - 05/20 600 LV 24 16 29 9 3 Edwards +5.0 Strategic Vision (R) 03/30 - 04/01 600 LV 20 19 27 4 4 Edwards +7.0 Strategic Vision (R) 02/16 - 02/18 600 LV 18 18 24 3 5 Edwards +6.0 Strategic Vision (R) 01/19 - 01/21 600 LV 17 15 25 1 4 Edwards +8.0 Zogby 11/06 - 11/06 502 LV 25 28 21 9 3 Clinton +3.0 Zogby 08/17 - 08/18 503 LV 19 30 23 10 3 Clinton +7.0 Zogby 05/14 - 05/15 511 LV 22 24 26 6 4 Edwards +2.0 Zogby 03/26 - 03/26 506 LV 23 25 27 3 3 Edwards +2.0 Zogby 02/07 - 02/09 500 LV 18 24 24 -- 4 Tie Zogby 01/15 - 01/16 596 LV 17 16 27 1 3 Edwards +10.0 Data take from the RCP site; copy past into excel ans then resorted befor pasting here.
Kurt_and_Hunter
(1000+ posts)
Send PM |
Profile |
Ignore
Wed Dec-19-07 09:09 PM
Response to Original message
1. He uses Robo-calls, which may well have different biases than live calling
Perky
(1000+ posts)
Send PM |
Profile |
Ignore
Wed Dec-19-07 10:12 PM
Response to Reply #1
17. The data seems to confirm a bias I wonder why its Clinton though.
Older and more apt to be home maybe?
OKNancy
(1000+ posts)
Send PM |
Profile |
Ignore
Wed Dec-19-07 09:10 PM
Response to Original message
2. One finds them suspicious
when the results are not to ones liking
mmonk
(1000+ posts)
Send PM |
Profile |
Ignore
Wed Dec-19-07 09:12 PM
Response to Reply #2
3. Or the data conflicts and conflicts consistently
Perky
(1000+ posts)
Send PM |
Profile |
Ignore
Wed Dec-19-07 09:23 PM
Response to Reply #3
7. Which is what is being seen here
I don't like the result..but that is secondary to the issue I raise.. why does everyone else use consitente quanities and the results in the same time fram are consistently different the Rasmussen? WHy the larger polling sample? which is fine, but then why the inconsistent numbers? WHy so much smaller this time. I am looking for explantions here not justifications....I dimply do not understandf why Ras is so different. What do they know about polling that no other firm knows?
mmonk
(1000+ posts)
Send PM |
Profile |
Ignore
Wed Dec-19-07 09:35 PM
Response to Reply #7
11. Extrapolation is guess work and some are better at it I guess.
I think they each develop their own methods at time. I look for trends mostly. Also, what is the sample such as likely voters, registered voters and such.
Perky
(1000+ posts)
Send PM |
Profile |
Ignore
Wed Dec-19-07 09:41 PM
Response to Reply #11
12. they are all LV Polls (compare with NH Polling)
Edited on Wed Dec-19-07 09:42 PM by Perky
Compare with New Hampshire Poll Date Sample Clinton Obama Edwards Richardson Biden Spread RCP Average 12/09 - 12/18 - 33.4 27.4 16 7 2 Clinton +6.0 American Res. Group 11/26 - 11/29 600 LV 34 23 17 10 3 Clinton +11.0 American Res. Group 10/26 - 10/29 600 LV 40 22 10 5 4 Clinton +18.0 American Res. Group 09/26 - 09/29 600 LV 41 22 10 8 3 Clinton +19.0 American Res. Group 08/26 - 08/29 600 LV 37 17 14 7 4 Clinton +20.0 American Res. Group 07/26 - 07/30 600 LV 31 31 14 7 2 Tie American Res. Group 06/27 - 06/30 600 LV 34 25 11 6 4 Clinton +9.0 American Res. Group 05/24 - 05/27 600 LV 34 15 18 9 3 Clinton +16.0 American Res. Group 04/26 - 04/29 600 LV 37 14 26 3 2 Clinton +11.0 American Res. Group 03/19 - 03/22 600 LV 37 23 20 2 2 Clinton +14.0 American Res. Group 01/31 - 02/01 600 LV 39 19 13 2 1 Clinton +20.0 American Res. Group 12/26 - 12/27 600 LV 24 25 21 2 1 Obama +1.0 CNN/WMUR 12/13 - 12/17 469 LV 38 26 14 8 2 Clinton +12.0 CNN/WMUR 12/06 - 12/10 378 LV 31 30 16 7 -- Clinton +1.0 CNN/WMUR 11/14 - 11/18 389 LV 36 22 13 12 2 Clinton +14.0 CNN/WMUR 09/17 - 09/24 307 A 43 20 12 6 3 Clinton +23.0 CNN/WMUR 07/09 - 07/17 333 LV 33 25 8 10 4 Clinton +8.0 CNN/WMUR 06/06 - 06/10 309 LV 36 22 12 10 5 Clinton +14.0 CNN/WMUR 03/27 - 04/02 339 LV 27 20 21 4 2 Clinton +6.0 CNN/WMUR 02/01 - 02/05 353 LV 35 21 16 1 3 Clinton +14.0Rasmussen 12/18 - 12/18 791 LV 31 28 18 8 2 Clinton +3.0 Rasmussen 12/11 - 12/11 841 LV 28 31 17 8 4 Obama +3.0 Rasmussen 11/29 - 11/29 959 LV 33 26 15 9 4 Clinton +7.0 Rasmussen 11/05 - 11/05 628 LV 34 24 15 8 -- Clinton +10.0 Rasmussen 10/23 - 10/23 841 LV 38 22 14 7 -- Clinton +16.0 Rasmussen 09/16 - 09/16 500 LV 40 17 14 11 -- Clinton +23.0 Rasmussen 08/09 - 08/09 453 LV 37 22 14 9 -- Clinton +15.0 Rasmussen 06/27 - 06/28 428 LV 38 21 10 9 -- Clinton +17.0 SurveyUSA 05/04 - 05/06 589 LV 40 24 22 -- -- Clinton +16.0 SurveyUSA 01/26 - 01/28 582 LV 40 25 23 -- -- Clinton +15.0 Zogby 12/01 - 12/03 502 LV 32 21 16 6 4 Clinton +11.0 Zogby 09/26 - 09/28 505 LV 38 23 12 8 2 Clinton +15.0 Zogby 05/15 - 05/16 503 LV 23 24 11 8 1 Obama +1.0 Zogby 04/02 - 04/03 LV 29 23 23 2 2 Clinton +6.0 Zogby 02/07 - 02/08 502 LV 27 23 13 3 2 Clinton +4.0 Zogby 01/15 - 01/17 502 LV 19 23 19 1 3 Obama +4.0
mmonk
(1000+ posts)
Send PM |
Profile |
Ignore
Wed Dec-19-07 09:46 PM
Response to Reply #12
13. Iowa is going to be harder to predict than New Hampshire IMO.
But then we don't know how the results will impact New Hampshire and I think it will to some extent. And yes, I give more credence to LV than RV.
durrrty libby
(1000+ posts)
Send PM |
Profile |
Ignore
Wed Dec-19-07 09:13 PM
Response to Reply #2
5. Yes indeed that is the most likely case
rinsd
(1000+ posts)
Send PM |
Profile |
Ignore
Wed Dec-19-07 09:13 PM
Response to Original message
4. Actually the extreme consistency of the other samples is more suspicious than Ras.
ARG SV ALWAYS have 600 in their sample Research 2000 always has 400. At this point I will stick with Ras and Zogby.
suston96
(1000+ posts)
Send PM |
Profile |
Ignore
Wed Dec-19-07 09:18 PM
Response to Original message
That large samples yield deceptive polling? I don't get it.
Perky
(1000+ posts)
Send PM |
Profile |
Ignore
Wed Dec-19-07 09:25 PM
Response to Reply #6
8. no on the contrary...the larger the sample the greater the accuracy
IN fact In trust larger pols more for that reason...but then I look at the sample size incosistency and I am gobsmacked! Particulary when this last one is so much smaller than the others.
suston96
(1000+ posts)
Send PM |
Profile |
Ignore
Wed Dec-19-07 09:30 PM
Response to Reply #8
9. Oh, OK. I misunderstood your intent....but....
...the sample size can be smaller because of the time of the year. Nevertheless, his 775 size is still larger than the other pollsters' samples.
Perky
(1000+ posts)
Send PM |
Profile |
Ignore
Wed Dec-19-07 09:34 PM
Response to Reply #9
10. But inconsitent over time.....WHy?
suston96
(1000+ posts)
Send PM |
Profile |
Ignore
Wed Dec-19-07 09:46 PM
Response to Reply #10
14. Because that is all he can get as he polls on a daily basis?
Are those other pollsters polling on a daily basis?
Perky
(1000+ posts)
Send PM |
Profile |
Ignore
Wed Dec-19-07 09:51 PM
Response to Reply #14
but someone said above that he uses robo-ocalling the other guys us real people they actually call alot more... but they use 600 to get a scientifically accura sample. I am not sure how y9u get that with robcalling.
mmonk
(1000+ posts)
Send PM |
Profile |
Ignore
Wed Dec-19-07 10:09 PM
Response to Reply #10
16. I really think the answer to your question is that
the election process becomes more fluid the longer it lasts or over time if you will. Different things and issues start to take shape which will influence the polling and give it different ebbs and flows as far as candidates go.
DU
AdBot (1000+ posts)
Wed May 08th 2024, 10:45 AM
Response to Original message
Powered by DCForum+ Version 1.1 Copyright 1997-2002 DCScripts.com
Software has been extensively modified by the DU administrators
Important Notices: By participating on this discussion
board, visitors agree to abide by the rules outlined on our Rules
page. Messages posted on the Democratic Underground Discussion Forums are the
opinions of the individuals who post them, and do not necessarily represent
the opinions of Democratic Underground, LLC.