USC Dornsife / LA Times Presidential Election "Daybreak" Poll [Trump +4.1% ...actually +10% ]

The USC Dornsife / LA Times ^ | July 25, 2016 | NA

[Future Useless Eater]

Trump ...................45.4
Clinton..................41.3

[Future Useless Eater]

The number you see today is an average of past polls PLUS the most recent. The assumption that the 4 point lead for Trump represents the response for 1 days polling is incorrect. Look at this as if it were the RCP average of polls.

You don't have to lecture me on how that poll works...
 
What you say is not wrong (other than your being vague) and you accusing me of being wrong.
It would be more correct for you to have said... 
     What you see today is an average of the latest six past polls plus the most recent.
I am NOT making any assumption that the 4 point lead for Trump represents the response for 1 day of polling.

I studied this for hours with a sophisticated spreadsheet, and my analysis is near-bullet-proof.

Here is a very-simplified spreadsheet test for you to verify that what I said was right....
Plug into your spreadsheet this example....
Hypothetically, lets say that 300 people got polled each day for a long period of time, 
and for 20 days straight Candidate "T" polled at exactly 42%.   
If that were the case, you certainly would agree that the 7-day average was 42% and the sample size was 2100.

Then presume a step-function occurred, and Candidate "T" truly did jump suddenly, 
(lets say starting on a FRIDAY) and consistently polled 48% for three days in a row.
What would the next THREE (seven-day) averages be reported as, for Friday, Saturday, and Sunday?

Here is the answer... 
  On Friday: 43.7%
On Saturday: 44.4%
  On Sunday: 45.1%
.........................Note how much this hypothetical simplification matches what really happened.
Then maybe you can understand that Trump REALLY IS at roughly 47 or 48 percent right now, and has been for about 3 days, 
and it will just take 4 more days before it shows up in this poll (assuming it holds ~steady for the full 7 days)

Likewise, Hillary really is in the 36-37% range right now, and she has been for about three days., 
If she were to stay stable at those values for four more days, those numbers (~37%) would also show up in this poll.

[Future Useless Eater]

TYPO Correction....

Hypothetically, lets say that 300 people got polled each day for a long period of time, 
and for 20 days straight Candidate "T" polled at exactly 43%.   
If that were the case, you certainly would agree that the 7-day average was 43% and the sample size was 2100.

Then presume a step-function occurred, and Candidate "T" truly did jump suddenly, 
(lets say starting on a FRIDAY) and consistently polled 48% for three days in a row.
What would the next THREE (seven-day) averages be reported as, for Friday, Saturday, and Sunday?

Here is the answer... 
  On Friday: 43.7%
On Saturday: 44.4%
  On Sunday: 45.1%
.........................Note how much this hypothetical simplification matches what really happened.
Then maybe you can understand that Trump REALLY IS at roughly 47 or 48 percent right now, and has been for about 3 days, 
and it will just take 4 more days before it shows up in this poll (assuming it holds ~steady for the full 7 days)

Likewise, Hillary really is in the 36-37% range right now, and she has been for about three days., 
If she were to stay stable at those values for four more days, those numbers (~37%) would also show up in this poll.

[Windflier]

Reports of unidentified flying ashtrays over DC.

[ctpsb]

Couldn’t agree more. I’ll take all the help/good news we can get but couldn’t agree more we can take nothing for granted. Vote fraud, the media, etc etc.

All conservatives MUST get everyone they know to vote the same as them and take nothing for granted!!

[LS]

I predicted in Jan or so Trump would win, and by about Feb I said it would be 300-320 EVs, but that the pop vote would be close.

I was wrong. Cankles is utterly crashing, and it’s possible that so many Dems will go 3D party Trump could get 52% Cankles as low as 40%, with Trump exceeding 320 EVs.

[billyboy15]

You misunderstood what I said and I haven’t the patience to explain it any more than I already have.

I will stand by what I said despite your spread sheet because I know how poll averaging works.

[Simon Foxx]

ROTFLMAO!!!!

[Ouderkirk]

and this is most likeky with a democRAT over-sample of 7 to 10 percentage points....so It’s likely to be Trump +15 were teh sample to be normalized in that manner.

[MaxistheBest]

Sorry...I didn’t see that ot is a 7 day rolling average..my mistake...I get it now.