Total Posts:7|Showing Posts:1-7
Jump to topic:

Polling Bias

vortex86
Posts: 559
Add as Friend
Challenge to a Debate
Send a Message
6/3/2016 3:20:40 PM
Posted: 6 months ago
I have communicated with RealClearPolitics a few times regarding their improper display of polling data. On top of that, there is an obvious bias in polling questioning if you take a close enough look.

Asking who would win Trump vs Clinton if there were no 3rd party candidates is not effective in showing anything as there are 3rd party candidates and this is not reality. The most recent poll by Mason_Dixon showing Clinton with a meager lead in Florida follows this same biased questioning. Polling people with either no 3rd party candidates or only the libertarian Gary Johnson ignoring Jill Stein's bid for the presidency as she takes numbers away from Clinton in polling data.

This further plays to the media bias argument Trump has proclaimed.

Thoughts?
vortex86
Posts: 559
Add as Friend
Challenge to a Debate
Send a Message
6/9/2016 4:26:34 PM
Posted: 5 months ago
Now to top it all off, besides ignoring 3rd party candidates. Now we have 800 Democrats polled vs 600 Republicans and we are surprised to see that Hillary Clinton is leading in the poll..

It's pretty sad when you have to skew things this much to try and validate your candidate.
Greyparrot
Posts: 14,230
Add as Friend
Challenge to a Debate
Send a Message
6/9/2016 4:31:45 PM
Posted: 5 months ago
At 6/9/2016 4:26:34 PM, vortex86 wrote:
Now to top it all off, besides ignoring 3rd party candidates. Now we have 800 Democrats polled vs 600 Republicans and we are surprised to see that Hillary Clinton is leading in the poll..

It's pretty sad when you have to skew things this much to try and validate your candidate.

There will always be biased polls. Not a surprise.
Ramshutu
Posts: 4,063
Add as Friend
Challenge to a Debate
Send a Message
6/9/2016 8:48:23 PM
Posted: 5 months ago
At 6/3/2016 3:20:40 PM, vortex86 wrote:
I have communicated with RealClearPolitics a few times regarding their improper display of polling data. On top of that, there is an obvious bias in polling questioning if you take a close enough look.

Asking who would win Trump vs Clinton if there were no 3rd party candidates is not effective in showing anything as there are 3rd party candidates and this is not reality. The most recent poll by Mason_Dixon showing Clinton with a meager lead in Florida follows this same biased questioning. Polling people with either no 3rd party candidates or only the libertarian Gary Johnson ignoring Jill Stein's bid for the presidency as she takes numbers away from Clinton in polling data.

This further plays to the media bias argument Trump has proclaimed.

Thoughts?

Almost every single one of the polls when comparing trump vs hillary include an "other" or "neither" option to some degree which safely deals with the third party candidates, almost universally.

Indeed, in the majority of polls I have seen both at a state and national level, have varied anywhere between 10 and 25% undecided/other option throughout.

Interestingly your second comment saying "Now we have 800 Democrats polled vs 600 Republicans and we are surprised to see that Hillary Clinton is leading in the poll.." indicates a fundamental misunderstanding of the process of polling and how it works.

For example, if you polled 800 democrats, 600 republicans, sure Hilllary will come out in the lead if you take raw responses. But no one takes raw responses; normally things such as party lines and demographics are weighted: 1000 democrats and 500 republicans in a poll? Each democrat vote for hillary will be weighted as a half, so if all democrats said Hillary, and all republicans say trump; you"d get a 50/50 split.

This is pretty much an obvious no-brainer. Professional pollsters make money through accuracy in general; and weighting is one of the big ways they can deal with that; not just by dem/rep responses, but also weights on demographics such as gender, race and age.

This is where the +/-% comes in with polls, it doesn"t mean that"s where the candidate could be, but if the votes chosen are representative of the country or place being polled as a whole (and their specified weighting), the randomized selection chosen could give anywhere between that error rate.

Now, in general, the polls from the last election have typically oversampled people voting for republican (comparing the pollsters predictions on the final day with how people actually voted); in general these have tended to be overwhelmingly giving an additional 1-2% to republicans, with anything up to 5 or 6% in some cases.

Most of the problems as it stands are several fold, and include many things:

1.) The distinction and classification of "Likely voters" vs "Registered Voters".
2.) Robo vs Live Calling (I think if I recall, Robocalling under samples liberal voters who have a higher hang-up rate).
3.) Mobile vs Land line calling: if you don"t call mobiles, which some pollsters/states don't you under sample young people who tend to vote democrat.
4.) Internet vs call polls: These tend to oversample men.

In general, the polling errors and skew aren"t actually down to any malicious intent to validate a candidate (if they are it"s typically a republican one given the polls from the last election), but are instead to do with the ability (or inability) to actually weight responses appropriately to get the right polling numbers using them.
AdamEsk
Posts: 202
Add as Friend
Challenge to a Debate
Send a Message
6/10/2016 1:48:01 AM
Posted: 5 months ago
This further plays to the media bias argument Trump has proclaimed.

Thoughts?

I'm from Texas, so it is safe to say I'm surrounded by Republicans. I've asked tons of them if they've ever gotten a phone call/email for these polls. Most of them don't, and if they do, they are normally working or don't care enough to take the poll. I think its safe to say there is some bias/variables in these polls that lead to faulty results.
BlueParagon
Posts: 22
Add as Friend
Challenge to a Debate
Send a Message
6/11/2016 5:53:20 AM
Posted: 5 months ago
At 6/3/2016 3:20:40 PM, vortex86 wrote:
I have communicated with RealClearPolitics a few times regarding their improper display of polling data. On top of that, there is an obvious bias in polling questioning if you take a close enough look.

Asking who would win Drumpf vs Clinton if there were no 3rd party candidates is not effective in showing anything as there are 3rd party candidates and this is not reality. The most recent poll by Mason_Dixon showing Clinton with a meager lead in Florida follows this same biased questioning. Polling people with either no 3rd party candidates or only the libertarian Gary Johnson ignoring Jill Stein's bid for the presidency as she takes numbers away from Clinton in polling data.

This further plays to the media bias argument Drumpf has proclaimed.

Thoughts?

If there is bias, that is unfortunate.
Axonly
Posts: 1,802
Add as Friend
Challenge to a Debate
Send a Message
6/12/2016 6:09:15 AM
Posted: 5 months ago
At 6/3/2016 3:20:40 PM, vortex86 wrote:
I have communicated with RealClearPolitics a few times regarding their improper display of polling data. On top of that, there is an obvious bias in polling questioning if you take a close enough look.

Asking who would win Trump vs Clinton if there were no 3rd party candidates is not effective in showing anything as there are 3rd party candidates and this is not reality. The most recent poll by Mason_Dixon showing Clinton with a meager lead in Florida follows this same biased questioning. Polling people with either no 3rd party candidates or only the libertarian Gary Johnson ignoring Jill Stein's bid for the presidency as she takes numbers away from Clinton in polling data.

This further plays to the media bias argument Trump has proclaimed.

Thoughts?

To be fair, no polling stats will ever be bias free.
Meh!