Beginning to wonder about this USC/Dornsrife poll

LA Times ^ | 09/01/2016 | Vanity

[ctpsb]

Trump 44.0% Clinton 43.3%

[txrefugee]

This is the usual Democrat media’s “horse race,” to keep the voters interested. They will have the race tied the day before the election. It means nothing. Get your family and friends registered and take them to the polls to vote for Trump.

[Cboldt]

You should have wondered two weeks ago, when the African American support jumped from 4 to 14%, stayed there for one week (the length of time each sample stays in the pool), then dropped to 4% again.

[cloudmountain]

Polls are a farce.
We learned that DAY ONE in the two statistics class I was FORCED to take.

[xzins]

I agree.

Except that I’m in a dry county....one of the few remaining in America. We have no watering hole.

[exnavy]

You don’the still believe the media, do you?

[monocle]

Go to the website and read who and how the poll is conducted.

[kabar]

Neither one is ahead. The differences are within the margin of error. It is really a dead heat according to this poll.

[CodeToad]

Worrying about polls, bogus polls, really?

[blueunicorn6]

Americans seem to have forgotten that Hillary Clinton was fired from her job as Secretary of State.

She was canned and now she wants to be our President?

No way.

[TexasCruzin]

RCP is full of polls that are skewed with D+4 to D+12 and most sample less Independents than they should. There are more Indies than D’s or R’s according to Gallup polling.

[af_vet_1981]

Mark Levin would be partly responsible for a Clinton victory/Obama third term, especially after supporting McCain and Romney. I hope he picks up the fallen flag and fights Clinton in the only possible fashion, telling his audience they should, even reluctantly, vote for Trump.

[TexasCruzin]

Exactly. LA Times cancelled Trump’s black vote surge because they just didn’t believe it or Nate Silver told them to.

[Raycpa]

Statistically, everything within the margin of error is the same numbers. The changes being reflected are too small to be statistically significant.

[pb929]

I think this polls uses a rolling 7 day average. So the new results published each day don't only reflect the new surveys, but also the ones that rolled off, so if the new day's polling was good, but the day that rolled off was better, the new overall results could still be worse.

Looking at polls say by day is a waste of time.

I put together an average of the 4 national tracking polls to smooth out the blips, and also another chart showing the overall trend of all the polls. If you're going to look at polls, look at the trends, not the actual numbers.

Take a look at it here

[WashingtonSource]

It’s the weekend effect? Only the weekend starts earlier. Working people are on vacation and out of reach while the dependent class is always around for a call.

[Jim Noble]

I just don't understand. When Trump is supposed to have had the best week of his campaign his numbers go down from just two days ago while Hillary's go up after her worst two weeks

Stop trying to understand meaningless drivel.

People make up their minds the last two weeks in October. None of the polling data is meaningful.

And the idea that someone on September 2 is "winning", and someone else is "losing" is just stupid.

[LS]

Today’s Ipsos is Trump +1

[LS]

Sorry, IBD Tipp not Ipsos=Trump+1

[HotHunt]

It's just a poll. Not necessarily accurate. And not necessarily reflective of what the election outcome will be.

Following the blow-by-blow descriptions and the blood-and-guts details every time a poll comes out will drive one crazy. Just take a couple of deep breaths and ignore the polls.

All I can do is vote. And I will. And it will be for Trump.

Make America Great Again. Trump/Pence 2016!

[ctpsb]

To all who responded on here (and thank you even if a few were a little edgy) my point was it seemed this one had always seemed a bit fair, but as a couple of you pointed out when the AA vote went from 4% to 14% back to 4% that should’ve been a bell ringing right there and I have now taken note.