Joe Biden approval rating discrepencies

[bmw]

https://www.realclearpolitics.com/epoll ... -7320.html

How do you have one poll with him at -2 and another at +23? That is an insane discrepancy of 25 points.

Looking back at the Trump Presidency, Politico had him at -29 in late January while Rasmussen had him at +3. That's 32 points of difference.

https://www.realclearpolitics.com/epoll ... -6179.html

That means when comparing Politico to Rasmussen, Rasmussen has Trump +5 compared to Biden while Politico has it -52, or a 57-point difference.

In reality, I think Trump was probably -7 or so in late January while Biden is probably +7 or so today, thus making Rasmussen an outlier and Politico an outright joke.

[Rate This]

https://www.realclearpolitics.com/epoll ... -7320.html

How do you have one poll with him at -2 and another at +23? That is an insane discrepancy of 25 points.

Looking back at the Trump Presidency, Politico had him at -29 in late January while Rasmussen had him at +3. That's 32 points of difference.

https://www.realclearpolitics.com/epoll ... -6179.html

That means when comparing Politico to Rasmussen, Rasmussen has Trump +5 compared to Biden while Politico has it -52, or a 57-point difference.

In reality, I think Trump was probably -7 or so in late January while Biden is probably +7 or so today, thus making Rasmussen an outlier and Politico an outright joke.

It’s not an exact science. Rasmussen is the only poll that saved Trump from being double digits down and is the only one preventing Biden from being more like plus 15-17.

[bmw]

There's no way Biden is +16 right now - no way. A recent CNN poll had him at +7 which I think is probably pretty accurate. To be at +16 he'd have to have significant support among Republicans which I just don't see being the case.

Anyways, my point was that these polls usually claim margin of error around 5 points give-or-take, and in both instances, these polls are seemingly WAY outside of those error bounds. In theory the polls should all be within 10 points of each other, but they're not, and 57 points relative to each other under difference Presidencies should raise serious questions about the methods used by both of these organizations.

[Honeyman]

I pretty much agree with beemer. While Rasmussen appears to be the outlier here, something is wrong with polling methodology these days. They simply have not shown to be very accurate. Either people are lying to them, they are not getting a proper cross-section of viewpoints, the data is being misinterpreted, etc., something is clearly flawed.

[TC Talks]

There's no way Biden is +16 right now - no way. A recent CNN poll had him at +7 which I think is probably pretty accurate. To be at +16 he'd have to have significant support among Republicans which I just don't see being the case.

Anyways, my point was that these polls usually claim margin of error around 5 points give-or-take, and in both instances, these polls are seemingly WAY outside of those error bounds. In theory the polls should all be within 10 points of each other, but they're not, and 57 points relative to each other under difference Presidencies should raise serious questions about the methods used by both of these organizations.

LOL, sorry the data hurts your feelings....

[Rate This]

There's no way Biden is +16 right now - no way. A recent CNN poll had him at +7 which I think is probably pretty accurate. To be at +16 he'd have to have significant support among Republicans which I just don't see being the case.

Anyways, my point was that these polls usually claim margin of error around 5 points give-or-take, and in both instances, these polls are seemingly WAY outside of those error bounds. In theory the polls should all be within 10 points of each other, but they're not, and 57 points relative to each other under difference Presidencies should raise serious questions about the methods used by both of these organizations.

Rasmussen has a serious Republican bias... Politico may have a serious bias the other way.

[Deleted User 14896]

I got a question Rate This may know:

How do they poll? Is it random phone calls? Standing in a food court at the mall with a clipboard? A little of both?
Thanks.

[Rate This]

I got a question Rate This may know:

How do they poll? Is it random phone calls? Standing in a food court at the mall with a clipboard? A little of both?
Thanks.

Random phone calls. The catch is that you can’t robocall cell phones (tell that to some folks!) and so they miss a lot of folks who don’t have land lines anymore. That’s among other issues including an absolutely abysmal response rate. It’s just terribly expensive to pay somebody to sit there and do it manually.

[Deleted User 14896]

I got a question Rate This may know:

How do they poll? Is it random phone calls? Standing in a food court at the mall with a clipboard? A little of both?
Thanks.

Random phone calls. The catch is that you can’t robocall cell phones (tell that to some folks!) and so they miss a lot of folks who don’t have land lines anymore. That’s among other issues including an absolutely abysmal response rate. It’s just terribly expensive to pay somebody to sit there and do it manually.

Thank you.
In my opinion, this method of opinion retrieval would not result in an accurate poll.

I myself do not answer any unknown numbers. I haven't done that since I got caller ID years and years ago. Even before I had a cellphone, I didn't answer the landline if I didn't know the number.

I think it's fair to say everyone here on Buzzboard is above average intelligence. Does anyone on here answer unknown numbers? I'm thinking probably not. But I'm confident we can think of a few family and friends that do answer every call they receive, not knowing who is calling.

Think of the ones you know that do that. Are they a little "special"?. Are they in the low end of the spectrum? That is a fair assessment of the folks I know that just can't let their phone go to voicemail. They're the ones that don't have quite as much as most. They're the ones where every conversation seems to be a bit superficial. They lack depth in their life.

These are the ones the pollsters are catching on the phone.

[MotorCityRadioFreak]

Yep, they call landlines which skews older and more Republican. So, if anything, Biden's numbers are even better.

[paul8539]

It also depends on who is doing the polling. Liberal group pollings would have high Biden and low Trump ratings, and conservative group pollings would have low Biden and high Trump ratings. It depends on the polls.

[jadednihilist]

The secret recipe in polling are the assumptions made in how to extrapolate a random sample of 400ish people to approximately 100 million actual voters. Raw polling numbers are never released. When you are polled, your answers are weighted based off the demographic information you provided; which is one major potential source of error. The other source of error is sampling methods. Polls conducted by landlines generally skew older, whereas online polls may skew younger or higher income; finding the right balance is tricky.

What is completely fair to say is that most polling firms are struggling with these assumptions in a Trump/Post-Trump era.

[Matt]

Yep, they call landlines which skews older and more Republican. So, if anything, Biden's numbers are even better.

Except that election polls vastly overrated democrats. Nice try...

[Lester The Nightfly]

I think it's fair to say everyone here on Buzzboard is above average intelligence.

[zzand]

I get robocalls on my cell all damn day long