APP - Curious sampling in this poll

C

canceled.2021.3

Former Vice President
http://www.reuters.com/article/us-usa-election-poll-idUSKCN12B2PV


The Reuters/Ipsos poll is conducted online in English in all 50 states. The poll of 2,386 American adults included 1,839 people who watched the debates, 1,605 people who were considered likely voters due to their registration status, voting history and stated intention to vote in the election. Among the likely voters, the poll counted 798 Democrats and 586 Republicans

Why the sampling imbalance?

Democrats represent 57.6% of likely voters, while Republicans represent just 36.5%.

Even if you use the results of the 2012 election where Obama won 51% to 47% it makes no sense weighting the sample this way. I know that many will say that I am trying to convince myself that Trump can win. Well, let me disabuse you of that notion right now. I am not confident that Trump can win.

So does anyone want to take a stab at the weighting of this poll?
 
another factor of the IPSOS polling is that it is only conducted among those who have registered to participate in the polls.....therefore it is a sampling of the same limited pool of people every week.......
 
Lots of pollsters and each wants attention so it's not odd that they get creative in methodoligy.
While it may be true that there is such a spread they probably should allow for actual participation to factor in. But chances are that othere already do so no advantage there.
It's a game, just can't ignore that
 
Celticguy said:
Lots of pollsters and each wants attention so it's not odd that they get creative in methodoligy.
While it may be true that there is such a spread they probably should allow for actual participation to factor in. But chances are that othere already do so no advantage there.
It's a game, just can't ignore that
Click to expand...

Doing these polls are tremendously challenging to be fair. For one, how can one predict who is likely to turn out? I admit I was wrong in 2012 about whether blacks would turn out like they did in 2008. They did. I am obviously making the same prediction this cycle and could be famously wrong.

This is a challenging election to predict. Both candidates are flawed. I don't think the endless ads will move the needle at all at this point. Not sure breathless condemnations from the NY Times will move the needle either. I don't think they are trying to impact November 8th as much as they are trying to impact early voting. That is the area where the GOP is woefully behind.

Personally I think we should end all early voting as it is ripe for fraud
 
Agree on early voting. Much harder to fool with when everyone is doing it at once.
I believe black turnout was off some but enough were bought to hit the mark.
I'm wondering if they will take the cash this time, that 'one of us' factor was very strong.
 
Celticguy said:
Agree on early voting. Much harder to fool with when everyone is doing it at once.
I believe black turnout was off some but enough were bought to hit the mark.
I'm wondering if they will take the cash this time, that 'one of us' factor was very strong.
Click to expand...

I agree. In the end, other than straight up partisans many don't like either. It is going to come down to who they hate more. I am not a big Trump fan. But, I have a binary choice. So I choose Trump

One other interesting thing to consider is that this election has some variables that make it hard to predict. For example, when an incumbent is running, if that incumbent is not above 45% they are likely finished and voters will break for the challenger. If you look at the polls on RCP Crooked ILLary is not above 45% in many of those polls even ones. So the question is this. Is Crooked ILLary technically an incumbent this year? I know she isn't elected, but given her 30 years, being part of the Obama White House etc., does that make her an incumbent? I can see an argument both ways.
 
You must log in or register to reply here.
Back
Top