U.S. CitiesCity-Data Forum Index
Go Back   City-Data Forum > General Forums > Politics and Other Controversies > Elections
 [Register]
Please register to participate in our discussions with 2 million other members - it's free and quick! Some forums can only be seen by registered members. After you create your account, you'll be able to customize options and access all our 15,000 new posts/day with fewer ads.
View detailed profile (Advanced) or search
site with Google Custom Search

Search Forums  (Advanced)
Reply Start New Thread
 
Old 10-12-2016, 02:31 PM
 
1,432 posts, read 917,563 times
Reputation: 333

Advertisements

Quote:
Originally Posted by OscarTheGrouch View Post
You are assuming a systematic error - ie every poll is biased the same way, which is clearly not true. Look at the mean-reverted bias:
http://projects.fivethirtyeight.com/pollster-ratings/

Some polls are biased one way, others are biased the other way. When you average a large number of polls with such variability in their bias you end up reducing your bias. This is fairly basic statistics.
Ah, ah....not quite as easily as you try to make it out. I am scientist, and not a political scientist, but I do know that by averaging polling data with biased results do not necessarily mean making them better. Or what we call smoothing the data. Systematic bias average do not allow more precision. So you can explain the mean-reverted analysis scientifically?

If I have 5 measurements, and each is skewed in one direction by 5%, averaging only tells me a point estimate that is 5% greater than the true mean, not where the true estimate lies.
Reply With Quote Quick reply to this message

 
Old 10-12-2016, 02:34 PM
 
1,432 posts, read 917,563 times
Reputation: 333
Quote:
Originally Posted by HeyJude514 View Post
You used the terms "skewed polls" versus "true polls." What do you mean by the term "true" polls if not the opposite of the "skewed" ones? You might not have used the term, but it's obvious that's what you meant.

The skewed versus the true (i.e. unskewed) polls is the argument many Romney fans made in 2012. They, like you, didn't believe the pollsters knew how to weight their polls. Romney believed it, too, which is why he had a fireworks display ready to go off when he was declared the winner, and he never wrote a concession speech.

But on election night it turned out that the pollsters--who do this for a living--knew exactly how to weight the polls and they called the election correctly.

But go ahead and believe you know more than the pollsters. You certainly won't be the first to look foolish on election night.
I said hypothetical. Guess you missed that part. My example had nothing to do with Romney. As someone who understands about how to do research and stats, and does for living, I know when I see poor methodology. Some of the polls may be very accurate, but some are dog crap. I am not saying I know more than pollsters, but I know when they are doing a poor job at sampling populations.
Reply With Quote Quick reply to this message
 
Old 10-12-2016, 02:40 PM
 
497 posts, read 302,141 times
Reputation: 576
I too am a scientist, and am wondering where you are getting your data that shows that the polls are systematically biased in one direction?

Both the fivethirtyeight ratings and past performance ( Are the polls skewed against Trump? - POLITICO ) would indicate that there is little to no 'mean bias' in the polls (although individual polls may have a few % point bias in either direction) and in the primary the polls actually over estimated Trumps performance.

Quote:
Originally Posted by Secchamps98 View Post
Ah, ah....not quite as easily as you try to make it out. I am scientist, and not a political scientist, but I do know that by averaging polling data with biased results do not necessarily mean making them better. Or what we call smoothing the data. Systematic bias average do not allow more precision. So you can explain the mean-reverted analysis scientifically?

If I have 5 measurements, and each is skewed in one direction by 5%, averaging only tells me a point estimate that is 5% greater than the true mean, not where the true estimate lies.
Reply With Quote Quick reply to this message
 
Old 10-12-2016, 02:41 PM
 
Location: Columbia, SC
18,230 posts, read 10,155,773 times
Reputation: 7073
Quote:
Originally Posted by Secchamps98 View Post
I have seen his list, and most are good in my opinion. I keep hearing this crying of plutocracy. That's total crap. I have one vote and a billionaire has one vote. I understand that money makes an unequal playing field, but the billionaires have the most to loose. Shouldn't they have a greater say in how people want to tax and use their money? Why don't we start tracking taxes dollars paid/per unit benefit gained so we can see how much each person contributes to paying their fairs share - and then decide how much voice they get in determining in how they will be taxed.
No. That is the problem now, as underscored by the fact that Trump can take deductions not available to "the little people".

That ain't Democracy, and it ain't republicanism. It's plutocracy.
Reply With Quote Quick reply to this message
 
Old 10-12-2016, 02:46 PM
 
1,432 posts, read 917,563 times
Reputation: 333
Quote:
Originally Posted by cuebald View Post
No. That is the problem now, as underscored by the fact that Trump can take deductions not available to "the little people".

That ain't Democracy, and it ain't republicanism. It's plutocracy.
Then we disagree. You can take the same deductions, build a business that is real estate focused. Sure he is paying SS tax for his employees, state taxes on R/D goods for new construction. Your argument is like, oh..I don't own a home, those who own one shouldn't be able to deduct their real estate interest since I can't. Deductions are in place to incentive behavior. If you don't like it elect someone who will change the tax code....but don't expect RC to do it for you.
Reply With Quote Quick reply to this message
 
Old 10-12-2016, 02:54 PM
 
1,432 posts, read 917,563 times
Reputation: 333
Quote:
Originally Posted by OscarTheGrouch View Post
I too am a scientist, and am wondering where you are getting your data that shows that the polls are systematically biased in one direction?

Both the fivethirtyeight ratings and past performance ( Are the polls skewed against Trump? - POLITICO ) would indicate that there is little to no 'mean bias' in the polls (although individual polls may have a few % point bias in either direction) and in the primary the polls actually over estimated Trumps performance.
I look at polls individually.

CNN Poll today (wheres the breakdown of voter registration - can't tell anything form this)

Poll finds Hillary Clinton and Trump tied in traditionally red Utah - CNNPolitics.com

The Y2 Analytics poll was conducted October 10-11, and used responses from 500 likely Utah voters and has a margin of error of plus or minus 4.4 percentage points.

2016 Poll Tracker - USA TODAY
Not data on methods - how can we trust this?

2016 Presidential Election Polls: Clinton vs. Trump
Not data on methods?

My point, if one uses an average of polls, we don't know the direction of the means (+/-), especially if ther is bias in sampling. So we don't know if the average would contain a good representation of where the true mean falls....

My frustration is a lack of transparency on how sampling is done...without that, each poll is suspect..
Reply With Quote Quick reply to this message
 
Old 10-12-2016, 02:58 PM
 
3,126 posts, read 1,271,731 times
Reputation: 2936
Quote:
Originally Posted by PCALMike View Post
That's certainly not going to happen with Trump. He promises us it will be impossible for anyone to win office again without begging the super rich for money, by eliminating all restrictions on campaign finance donations. Whats so great about that? His supreme court picks are a disaster for salt of the earth Americans.

People claim that massive income and wealth inequality is not a problem when in reality it destroys democracies. When a super rich family can pour billions of dollars into the political process to get what they want, thats not democracy, thats plutocracy. The 99.9% are the losers in this game and the sooner people realize that, the better.
You are obviously a victim of the left wing education system, Very sad...
Reply With Quote Quick reply to this message
 
Old 10-12-2016, 03:00 PM
 
497 posts, read 302,141 times
Reputation: 576
Ahhh, because you don't have the gumption to look up a polls methodology, doesn't mean that it isn't there. I would suggest you go to the RCP polls page:
RealClearPolitics - Election 2016 - General Election: Trump vs. Clinton

Click on any poll you want and wallah - you get the raw poll data and the methodology. For example:

NBC News/Wall St. Jrnl 10/8 - 10/10 806 LV 3.5 50 40 Clinton +10
Methodology: https://www.scribd.com/document/3272...-10-11-RELEASE

Ignorance is not really an excuse for making unsubstantiated claims (particularly for a self proclaimed scientist).


Quote:
Originally Posted by Secchamps98 View Post
I look at polls individually.

CNN Poll today (wheres the breakdown of voter registration - can't tell anything form this)

Poll finds Hillary Clinton and Trump tied in traditionally red Utah - CNNPolitics.com

The Y2 Analytics poll was conducted October 10-11, and used responses from 500 likely Utah voters and has a margin of error of plus or minus 4.4 percentage points.

2016 Poll Tracker - USA TODAY
Not data on methods - how can we trust this?

2016 Presidential Election Polls: Clinton vs. Trump
Not data on methods?

My point, if one uses an average of polls, we don't know the direction of the means (+/-), especially if ther is bias in sampling. So we don't know if the average would contain a good representation of where the true mean falls....

My frustration is a lack of transparency on how sampling is done...without that, each poll is suspect..
Reply With Quote Quick reply to this message
 
Old 10-12-2016, 03:00 PM
 
3,126 posts, read 1,271,731 times
Reputation: 2936
Quote:
Originally Posted by Unsettomati View Post
It's always good to get input from a poster whose history reveals lots of 'Romney's gonna win!' posts from 2012.

But I'm fascinated by the fact that some people want to record their cluelessness for posterity...
Yes, it's very brave of you to do that.....
Reply With Quote Quick reply to this message
 
Old 10-12-2016, 03:02 PM
 
3,126 posts, read 1,271,731 times
Reputation: 2936
Quote:
Originally Posted by jambo101 View Post
The whole smear campaign against Hillary is an election strategy Trump is using to deflect from the fact that he is a political novice with no clue how to run the country,he says nothing of substance and all his bluster is directed at trashing Hillary.As you claim to be highly educated i'd have thought you would see through the obvious bs.
Wow, you don't get much air or light in your basement do you???
Reply With Quote Quick reply to this message
Please register to post and access all features of our very popular forum. It is free and quick. Over $68,000 in prizes has already been given out to active posters on our forum. Additional giveaways are planned.

Detailed information about all U.S. cities, counties, and zip codes on our site: City-data.com.


Reply
Please update this thread with any new information or opinions. This open thread is still read by thousands of people, so we encourage all additional points of view.

Quick Reply
Message:

Over $104,000 in prizes was already given out to active posters on our forum and additional giveaways are planned!

Go Back   City-Data Forum > General Forums > Politics and Other Controversies > Elections
Follow City-Data.com founder on our Forum or

All times are GMT -6.

© 2005-2019, Advameg, Inc. · Please obey Forum Rules · Terms of Use and Privacy Policy · Bug Bounty

City-Data.com - Archive 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35 - Top