• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Another Presidential Poll Late June 2016

Wise people use poll averages or even better, weighted poll averages, to avoid the problem of cherry picking one poll to show what you want while being oblivious to the actual data. I could just as easily link the ABC News / Washington Post poll from Thursday with Clinton at +12%. That is just as improperly misleading as your cherry picked poll with Clinton at +2%.

Polls have errors. Thus, candidates will float up and down in any individual poll. Cherry picking the extremes is never a good idea, since that is likely to be statistical noise rather than real results. You can cherry pick the middling polls if you really dislike poll averaging. I could give you a half dozen polls from the last week with Clinton near +5% which is probably much more realistic.
 
dullard said:
Wise people use poll averages or even better, weighted poll averages, to avoid the problem of cherry picking one poll to show what you want while being oblivious to the actual data. I could just as easily link the ABC News / Washington Post poll from Thursday with Clinton at +12%. That is just as improperly misleading as your cherry picked poll with Clinton at +2%.

Polls have errors. Thus, candidates will float up and down in any individual poll. Cherry picking the extremes is never a good idea, since that is likely to be statistical noise rather than real results. You can cherry pick the middling polls if you really dislike poll averaging. I could give you a half dozen polls from the last week with Clinton near +5% which is probably much more realistic.
Click to expand...
I'm not cherry picking. There are other polls that have been brought up in P&N. This is just the latest one. Just because you don't like it, doesn't mean it shouldn't be discussed.
 
TheGardener said:
I'm not cherry picking. There are other polls that have been brought up in P&N. This is just the latest one. Just because you don't like it, doesn't mean it shouldn't be discussed.
Click to expand...

What conclusions do you draw from this poll? Why have you not brought up other ones that show Clinton leading by double digits?

You know as well as we do that you are drawn to this poll because it tells you what you want to hear. Dullard is 100% right that focusing on any one poll is stupid, no matter which side it favors.
 
Polls aren't working anymore because the pollsters no longer know how to weight their polling data.

Part of that is because, in the last 5 years, the number of self-identified independents has skyrocketed. Some of the Gallup surveys indicate it's as high as 45%, which would make it 50-100% larger than either of the main parties.

Another reason is that they look at 'likely voters'. That is to say, people who voted in the past. A lot of people who didn't vote in the past will be voting in Nov.

Also on this nonsense of the so-called "accuracy" of polling averages a la RealClearPolitics.

Just look at this and tell me how accurate it was :

BobgRka.jpg
 
shady28 said:
Polls aren't working anymore because the pollsters no longer know how to weight their polling data.

Part of that is because, in the last 5 years, the number of self-identified independents has skyrocketed. Some of the Gallup surveys indicate it's as high as 45%, which would make it 50-100% larger than either of the main parties.

Another reason is that they look at 'likely voters'. That is to say, people who voted in the past. A lot of people who didn't vote in the past will be voting in Nov.

Also on this nonsense of the so-called "accuracy" of polling averages a la RealClearPolitics.

Just look at this and tell me how accurate it was :

BobgRka.jpg
Click to expand...

Primary polling and general election polling are almost nothing alike. Large errors are common in primary polling in any year. Large errors are not common in general election polling. If anything here is nonsense it is your attempt to equate the two. No competent pollster would do that.

Furthermore, according to Gallup the percentage of self identified independents is not significantly different today than it was in 2012 and poll aggregation models did fine. Not perfectly, but well enough for them all to correctly predict the outcome.
 
TheGardener said:
I'm not cherry picking. There are other polls that have been brought up in P&N. This is just the latest one. Just because you don't like it, doesn't mean it shouldn't be discussed.
Click to expand...

Do you know what cherry picking means?

Did you actually read Dullard's post?
 
FiveThirtyEight was wrong about Trump because they didn't listen to the polls as much as they did in the past. The polls show Hillary with a healthy lead in aggregate.
 
TheGardener said:
Ig'm not cherry picking. There are other polls that have been brought up in P&N. This is just the latest one. Just because you don't like it, doesn't mean it shouldn't be discussed.
Click to expand...

This is the definition of cherry picking. You're picking one poll that you like, while ignoring others.
 
I could spend hours a days posting the same comments to each one of you, every time you create a thread. And you know it.

But here is another article, referencing the same poll, BUT it also has a graph of the other polls. If that is not sufficient, tough. I'm not going to spend my time whining like you. Let's face it, each of you post and linked topics which don't reflect all sides. It's not that I cherry picked. It's that you don't like what I said. If I post another poll that showed Hillary ahead, you'd be slapping one another on the back.

Hypocrisy will rule the day, when you believe that you can exploit a situation for your benefit. TheGardener

http://www.dailymail.co.uk/news/art...earth-campaign-two-candidates-don-t-like.html
 
Only morons could think the OP believes that this poll is the only one out there. Had a guy post about some battle ground state polls that seemed completely out of whack. Not a peep from any of the liberals about cherry picking. Cherry picking is what the complainers are doing by only complaining about cherry picking when polls they don't like are posted.
 
buckshot24 said:
Only morons could think the OP believes that this poll is the only one out there. Had a guy post about some battle ground state polls that seemed completely out of whack. Not a peep from any of the liberals about cherry picking. Cherry picking is what the complainers are doing by only complaining about cherry picking when polls they don't like are posted.
Click to expand...

buckshatting
 
can't believe any of the poles..they are heavily skewed towards any liberal candidate in an attempt to influence popularity.
 
shady28 said:
Just look at this and tell me how accurate it was :

BobgRka.jpg
Click to expand...



Looks like they nailed each candidate's actual result (individual state polls are never very accurate).
  • Carson was only off by 1.8%
  • Kasich was only off by 2.5%
  • Rubio was only off by 0.3%
  • Trump was only off by 1.5%
  • Cruz was off by 6.6%, but the polls being less than 800 voters (often far less), a 6.6% error expected to happen fairly regularly.
Remember, polls are not very accurate when subtracting the two candidates values. You should only be looking at each individual candidate's amount in polls. Thus, if there is a 5% margin of error per candidate in an individual poll (about typical of the polls in that image), then margin of error for subtracting two candidates is close to +-10% (it is not perfectly accurate to just double the poll margin of error for any one candidate but the end result close enough for this post). Thus Cruz's actual win of 17.1% is within the +-10% range expected of RCP prediction of Cruz winning by 9% (9% +-10% is anywhere from Cruz losing by 1% to Cruz winning by 19%).
 
Last edited:
What is ironic (not really) is that all of you whiners, who have suddenly joined this cherry picking bandwagon, have not commented on the substance of this latest poll. Can you spell D-I-V-E-R-S-I-O-N?
 
TheGardener said:
What is ironic (not really) is that all of you whiners, who have suddenly joined this cherry picking bandwagon, have not commented on the substance of this latest poll. Can you spell D-I-V-E-R-S-I-O-N?
Click to expand...

It's either that or just claim this is another of the dozen reposts of the same "X poll this week says this!" thread littering P&N these days.

So no, this isn't diversion. It's treating this oft-repeated topic with the proper amount of substance that it deserves, yet again.
 
dullard said:
Looks like they nailed each candidate's actual result (individual state polls are never very accurate).
  • Carson was only off by 1.8%
  • Kasich was only off by 2.5%
  • Rubio was only off by 0.3%
  • Trump was only off by 1.5%
  • Cruz was off by 6.6%, but the polls being less than 800 voters in the polls, a 6.6% error expected to happen fairly regularly.
Remember, polls are not very accurate when subtracting the two candidates values. You should only be looking at each individual candidate's amount in polls. Thus, if there is a 5% margin of error per candidate in an individual poll (like many of the polls in that image), then margin of error for subtracting two candidates is close to +-10% (it is not perfectly accurate to just double the poll margin of error for any one candidate but the end result close enough for this post). Thus Cruz's actual win of 17.1% is within the +-10% range expected of RCP prediction of Cruz winning by 9% (9% +-10% is anywhere from Cruz losing by 1% to Cruz winning by 19%).
Click to expand...

That is a very good and important point. This is true when looking at any poll that if you're trying to examine the spread between two different candidates you need to conduct a difference in proportions test, which is not what those poll numbers do.
 
The internals of this poll look funny. He gets 33% support from Hispanics in this poll. More than both McCain and Romney did. That really doesn't seem believable.

White voters back Trump 47 - 34 percent. Black voters back Clinton 91 - 1 percent and Hispanic voters back her 50 - 33 percent. Voters 18 to 34 years old go Democratic 48 - 23 percent, while voters over 65 years old go Republican 51 - 35 percent.

http://www.qu.edu/news-and-events/quinnipiac-university-poll/national/release-detail?ReleaseID=2363
Click to expand...
 
Last edited:
You must log in or register to reply here.

TRENDING THREADS

Back
Top