If I take a poll for a Democratic candidate and poll 527 Democrats and 443 Republicans?

[kentuck]

What percentage would you expect the Democrat to receive? Almost 20% more Democrats polled than Republicans.

Would you expect the Democrat to get at least 50% in the polls?

Those were the numbers polled in the recent Gallup poll for Donald Trump, except it was 527 Repubs polled and 443 Democrats. Even with almost a 20% advantage in Republican voters, he still could not reach 50%.

What would have been the numbers if it was a more balanced poll?

[Arkansas Granny]

1. I like to check Real Clear Politics from time to time. They average several national polls to

arrive at their figure. They show 44.9% approve and 52% disapprove as of of 2/3/2020. He has never reached 50% approval in their average polling figures.

https://www.realclearpolitics.com/epolls/other/president_trump_job_approval-6179.html

[PRETZEL]

6. I like RCP also,

but at the same time, polls are really just snapshots in time, just like any other picture.

Three months from now, based upon the events of the day, any particular poll could easily show something totaly different.

Six months from now, it'll be another snapshot.

And in as much as it's real easy to just use national polls as indicators, those indicators don't really reflect battleground sentiment. That poll, if possible to project on a national level, would have more significance to me than just a national poll or even an average.

[OliverQ]

7. 45% approval is still disgustingly and frighteningly high.

He should be in the low 30s minimum.

[TwilightZone]

2. Pollsters adjust for demographics

They use numbers adjusted for a variety of factors, not raw numbers.

From Gallup:

"Gallup weights samples to correct for unequal selection probability and nonresponse. Gallup also weights its final samples to match the U.S. population according to gender, age, race, Hispanic ethnicity, education and region. Other adjustments may be made to fit the needs of a specific study. Demographic weighting targets are based on the most recent Current Population Survey figures for the aged 18 and older U.S. population."

https://www.gallup.com/174158/gallup-panel-methodology.aspx

[kentuck]

3. How about cult membership?

Do they correct for that?

[TwilightZone]

11. That's what they're measuring, isn't it?

Whether Trump's approval rating is 40, 45, or 50% is pretty much irrelevant. It's inexplicably high, regardless of what the "real" number is.

We underestimate Trump's appeal at our peril and denial isn't going to help us in November. What matters is what we do about it, starting with turnout.

[uponit7771]

5. Then they should've adjusted down seeing there are less republicans now than in 16. That makes

... red dons numbers among republicans go up because there's more purity among the sample

[Maraya1969]

9. If they do so much adjusting for accuracy how come they are always

skewed to the right compared to other polling companies?

You Gov, Morning consult, Monmouth University and most others always have lower ratings for Trump that Gallup.

[ECSkeptic]

4. Data is messy...

...even when weighting results by the party identity of respondents. Random noise and error creep into any measurement, and measurements of human behavior are especially messy. I'm sure Gallup is reporting honestly, but they cannot change the inherently random nature of measurement. The polling averages computed by RCP and FiveThirtyEight do a better job of teasing the signal from the noise than single polls. Those averages have been stubbornly stable for a long time, with an average approval of around 43% for Trump, give or take. I think that approval is more or less baked in at this point, with most voters living fat, dumb, and happy in their respective ideological bubbles. People know Trump and their opinions of him are basically locked down. Any other President with this economy would be blowing the doors off all opponents, not languishing in the low 40's. The fact that he's not suggests that nearly all the electorate have made up their minds and won't change them.

If there is something that could significantly move the needle, I'm not smart enough to think of it. Mobilization, motivation, and turnout will win (or lose) the day on Nov. 3.

[rockfordfile]

8. The fact is this was setup by a right-wing pollster as propaganda for lies.

I believe Gallup polling was investigated years ago?

[GeorgeGist]

10. Why not just start your post with the truth?