US Politics Thread | Page 398 | Vital Football

US Politics Thread

Last time, millions of Hillary voters stayed at home, because they thought she had it in the bag. Another significant group voted for an Independent, feeling they were free to vote for who they liked, not tactically.

2018 saw massive Democrat turnouts, and they won by over 10,000,000 votes.

If Trump wins this time, there will be something dodgy going on.
 
I have to say Trumpy is the most undiplomatic person I have ever see in a public position. I have had a laugh at his twattish and childish behaviour.

Its wearing very thin now.
 
I have ended up with a couple of American 'friends' on Facebook through my cousin, who lives in Ohio. It's instructive to see the posts they put up. They are dyed in the wool Trumpers and believe that the whole rioting situation has been inflamed by Marxists and ultra left liberals. Apparently, the Democratic Party is riddled with them. They are solidly behind Trump.
An example of their mindset is revealed by one post showing showing looters with the caption,
"This is why I need an AR15 and 30 round magazines."
I think Trump has a great chance at getting re-elected.
 
ironside said:
I have ended up with a couple of American 'friends' on Facebook through my cousin, who lives in Ohio. It's instructive to see the posts they put up. They are dyed in the wool Trumpers and believe that the whole rioting situation has been inflamed by Marxists and ultra left liberals. Apparently, the Democratic Party is riddled with them. They are solidly behind Trump.
An example of their mindset is revealed by one post showing showing looters with the caption,
"This is why I need an AR15 and 30 round magazines."
I think Trump has a great chance at getting re-elected.
Click to expand...
Polls show a spread in favour of Biden, you're looking at an average spread of 6 points across the following set of polls https://www.realclearpolitics.com/epolls/2020/president/us/general_election_trump_vs_biden-6247.html.

I'd be cautious with the margin of error, response rate issues create bias, some statisticians would recommend doubling the margin of error in order to be cautious, if we do this then you can see things could theoretically be a lot closer.

Biden has come across very well of late and has cashed in on a great opportunity for his campaign. I still maintain questions over his mental acuity given his age and he has a proclivity for gaffs, what will help him is that given these riots his party will fall entirely behind him. A few months ago some individuals in the democrat party were pursing sexual harassment allegations and he has been called a racist during the primaries, this will all go.

With COVID reducing opportunities for a good ground game I do believe the debate between Trump and Biden will be all the more crucial.
 
Pangloss said:
Polls show a spread in favour of Biden, you're looking at an average spread of 6 points across the following set of polls https://www.realclearpolitics.com/epolls/2020/president/us/general_election_trump_vs_biden-6247.html.

I'd be cautious with the margin of error, response rate issues create bias, some statisticians would recommend doubling the margin of error in order to be cautious, if we do this then you can see things could theoretically be a lot closer.

Biden has come across very well of late and has cashed in on a great opportunity for his campaign. I still maintain questions over his mental acuity given his age and he has a proclivity for gaffs, what will help him is that given these riots his party will fall entirely behind him. A few months ago some individuals in the democrat party were pursing sexual harassment allegations and he has been called a racist during the primaries, this will all go.

With COVID reducing opportunities for a good ground game I do believe the debate between Trump and Biden will be all the more crucial.
Click to expand...
I did however read that the all important swing states are still behind Trump.
It will certainly be an interesting US election next time around.
 
Pangloss said:
Polls show a spread in favour of Biden, you're looking at an average spread of 6 points across the following set of polls https://www.realclearpolitics.com/epolls/2020/president/us/general_election_trump_vs_biden-6247.html.

I'd be cautious with the margin of error, response rate issues create bias, some statisticians would recommend doubling the margin of error in order to be cautious, if we do this then you can see things could theoretically be a lot closer.

Biden has come across very well of late and has cashed in on a great opportunity for his campaign. I still maintain questions over his mental acuity given his age and he has a proclivity for gaffs, what will help him is that given these riots his party will fall entirely behind him. A few months ago some individuals in the democrat party were pursing sexual harassment allegations and he has been called a racist during the primaries, this will all go.

With COVID reducing opportunities for a good ground game I do believe the debate between Trump and Biden will be all the more crucial.
Click to expand...

As a Statistician, I have a Masters in Statistical Analysis, I would be interested on where you got the margin of error information from.
 
ironside said:
I have ended up with a couple of American 'friends' on Facebook through my cousin, who lives in Ohio. It's instructive to see the posts they put up. They are dyed in the wool Trumpers and believe that the whole rioting situation has been inflamed by Marxists and ultra left liberals. Apparently, the Democratic Party is riddled with them. They are solidly behind Trump.
An example of their mindset is revealed by one post showing showing looters with the caption,
"This is why I need an AR15 and 30 round magazines."
I think Trump has a great chance at getting re-elected.
Click to expand...

Only about 30% are like this, though. It's his base, and it has neither increased nor decreased since 2016.

He needs the Independents, and some disillusioned democrat votes to stand a chance. Last time some Obama voters did go with Trump. I can't see any of them doing that again.
 
HeathfieldRoad1874 said:
As a Statistician, I have a Masters in Statistical Analysis, I would be interested on where you got the margin of error information from.
Click to expand...
Is the MoE field in that table not margin of error in that link? Or have I misunderstood what that stood for?

Interestingly since I posted the swing has increased and now 8 points.

I have an undergraduate degree in Computer Science and Statistics and a Masters in Mathematical Economics and Econometrics if we're getting our academic dicks out :D
 
Last edited:
HeathfieldRoad1874 said:
Sorry, I didn't make myself clear. You suggested doubling it, but with the normal 95% confidence interval, why? And who is suggesting it?
Click to expand...

I guess you could construct a CI fairly easily...

Anyway the doubling thing, this was something I read from David Spiegelhalter (https://en.wikipedia.org/wiki/David_Spiegelhalter ) in one of his books. His justification is that there's obviously random variability in these things and we can cover ourselves there. But polls are highly fallible due to sampling methodology, low response rates etc especially on information gained via telephony which have poor response rates in general.. He rather heuristically suggested that he likes to double the MoE's as a result, good enough for him...

I guess he's not a pollster so take from that what you will... Incidentally if there's significant issues with systematic bias introduced by the survey methodology then your confidence interval: estimate +\- 2*standard Error of estimate (if assuming normal) may not be very good because of the error in variable problem no?
 
The polls seem to have been quite accurate in 2018, and all the special elections since 2016.

Indeed, they weren't far out with Hilary, they just didn't account for many of her voters staying at home. Not every pollster asks intention to vote.

The larger the sample, the more accurate the results, so that's why I like someone like FiveThirtyEight' results. They also track bias, and adjust accordingly.
 
HeathfieldRoad1874 said:
The polls seem to have been quite accurate in 2018, and all the special elections since 2016.

Indeed, they weren't far out with Hilary, they just didn't account for many of her voters staying at home. Not every pollster asks intention to vote.

The larger the sample, the more accurate the results, so that's why I like someone like FiveThirtyEight' results. They also track bias, and adjust accordingly.
Click to expand...

I'm not aware of FiveThirtyEight conducting polls, I thought that 'just' created pipelines to capture that data from other sources and then aggregate them. If, and I stand to be corrected (so please do), they are just pulling from elsewhere and aggregating then it's difficult to see how they can make sophisticated adjustments for bias IF they only have access to that aggregated data, these are adjustments that you'd look to make at the row level no?
 
They can analyse past results and detect trends in polling by different pollsters. They do conduct a few of their own polls, but the daily approval tracking for Trump is an amalgamation of all the others out there. They also rate them based on methodology. Then they adjust for anything that varies consistently from the consensus.

An example would be Rasmussen, who only poll people with a landline. As you said above, telephone polls are inherently inaccurate, and this is compounded by landline being more likely to be elderly voters. That's why they adjust the numbers downwards by 5% for them.

It is classic Type A methodology as described by Spiegelhalter .
 
You must log in or register to reply here.