 |
|
On polling accuracy...
Posted on: November 4, 2016 at 10:41:20 CT
JeffB MU
Posts:
72172
Member For:
21.41 yrs
Level:
User
M.O.B. Votes:
0
A lot of the polls seem to try and save money by polling fewer people but trying to get a randomized effect by selecting rough percentages of whatever groups they think should be represented. There are allegations that some deliberately skew their samples to affect the polling results. That assertion is buttressed by the fact that emails Wikileaks has released show both a strong media bias in some mainstream media and instructions sent detailing how to skew those results.
In any event, most publish percentages of those in their polls claiming to be Democrats, Republicans or independents. Looking at those numbers vs the numbers in the population at large can show whether or not their sample is representative of the nation at large.
Per Gallup 29% of the voters describe themselves as Democrats & 26% as Republicans. http://www.gallup.com/poll/188096/democratic-republican-identification-near-historical-lows.aspx
Here's an article on Zerohedge noting the discrepancy in some of the polling numbers:
Latest Polls: ABC Manufactures A 3-Point "Lead" For Hillary As IBD/TIPP Still Sees A Dead Heat
http://www.zerohedge.com/news/2016-11-04/latest-polls-abc-manufactures-3-point-lead-hillary-ibd-still-sees-dead-heat
The latest goal-seeking effort by the ABC / Washington Post found Hillary to have a 3-point national advantage against Trump despite her tanking in "trustworthiness" and her "voter enthusiasm" waning. Their latest poll included a 7-point democrat sampling advantage though we suspect there are other pro-democrat "ethnic oversamples" incorporated which ABC does not disclose.
This ABC News/Washington Post poll was conducted by landline and cellular telephone Oct. 30-Nov. 2, 2016, in English and Spanish, among a random national sample of 1,151 likely voters. Results have a margin of sampling error of 3 points, including the design effect. Partisan divisions are 37-30-29 percent, Democrats-Republicans-independents. ...
In any event, most publish percentages of those in their polls claiming to be Democrats, Republicans or independents. Looking at those numbers vs the numbers in the population at large can show whether or not their sample is representative of the nation at large.
Per Gallup 29% of the voters describe themselves as Democrats & 26% as Republicans. http://www.gallup.com/poll/188096/democratic-republican-identification-near-historical-lows.aspx
Here's an article on Zerohedge noting the discrepancy in some of the polling numbers:
Latest Polls: ABC Manufactures A 3-Point "Lead" For Hillary As IBD/TIPP Still Sees A Dead Heat
http://www.zerohedge.com/news/2016-11-04/latest-polls-abc-manufactures-3-point-lead-hillary-ibd-still-sees-dead-heat
The latest goal-seeking effort by the ABC / Washington Post found Hillary to have a 3-point national advantage against Trump despite her tanking in "trustworthiness" and her "voter enthusiasm" waning. Their latest poll included a 7-point democrat sampling advantage though we suspect there are other pro-democrat "ethnic oversamples" incorporated which ABC does not disclose.
This ABC News/Washington Post poll was conducted by landline and cellular telephone Oct. 30-Nov. 2, 2016, in English and Spanish, among a random national sample of 1,151 likely voters. Results have a margin of sampling error of 3 points, including the design effect. Partisan divisions are 37-30-29 percent, Democrats-Republicans-independents. ...
Please explain why this message is being reported.
MESSAGE THREAD
On polling accuracy... - JeffB MU - 11/4 10:41:20