* If you’re following national polling average sites like RCP, you know that the Los Angeles Times/USC poll has shown Donald Trump leading Hillary Clinton when pretty much all other polls have shown Clinton ahead. Well it turns out that a single polling “panelist,” who is a 19 year-old Illinois African-American has been skewing the results (of a highly questionable methodology) because he is supporting Trump…
In some polls, he’s weighted as much as 30 times more than the average respondent, and as much as 300 times more than the least-weighted respondent.
Alone, he has been enough to put Mr. Trump in double digits of support among black voters. He can improve Mr. Trump’s margin by 1 point in the survey, even though he is one of around 3,000 panelists.
He is also the reason Mrs. Clinton took the lead in the U.S.C./LAT poll for the first time in a month on Wednesday. The poll includes only the last seven days of respondents, and he hasn’t taken the poll since Oct. 4. Mrs. Clinton surged once he was out of the sample for the first time in several weeks.
How has he made such a difference? And why has the poll been such an outlier? It’s because the U.S.C./LAT poll made a number of unusual decisions in designing and weighting its survey.
There’s a lot more to this, so go read the whole thing.
- AlfondoGonz - Wednesday, Oct 12, 16 @ 2:54 pm:
Not sure how race is relevant here, other than because we seem to have found the one black Trump supporter.
- Ron - Wednesday, Oct 12, 16 @ 2:59 pm:
It matters because polls weight for demographics, the problem with this “poll” is that it took a random sample of 3000 people and tracks them throughout the election. It’s not really a poll at all but a tracking of a very specific group of voters. It’s flawed for anything other than trends.
- Rich Miller - Wednesday, Oct 12, 16 @ 3:00 pm:
===Not sure how race is relevant here===
You really are completely weird. Read the freaking story. It’s completely relevant. It’s the whole story.
- Amalia - Wednesday, Oct 12, 16 @ 3:15 pm:
oh, 19 year old men……
- Responsa - Wednesday, Oct 12, 16 @ 3:15 pm:
I already posted my “polling” rant on another thread here today and it still applies–in spades– so I won’t repeat it.
- AlfondoGonz - Wednesday, Oct 12, 16 @ 3:16 pm:
Point taken. *Sheepishly acknowledges I did not read whole story*
I do, however, take “weird” as a compliment.
- Ron - Wednesday, Oct 12, 16 @ 3:22 pm:
Clinton is up by 5-10% in virtually every other recent poll. That’s like 7 vs. this 1 garbage outlier.
- train111 - Wednesday, Oct 12, 16 @ 3:23 pm:
A few weeks back there was a story about a polling company that took its raw survey data and handed it off to 4 other polling companies. Lo and behold you had 5 different results all the way from Hillary +5 to Trump +1. All these companies apply their “secret sauce” to the raw numbers to approximate the make up of the entire electorate. This LA Times one is certainly different than all the others - especially with their transperancy about how they “massage” the raw numbers.
- Ron - Wednesday, Oct 12, 16 @ 3:24 pm:
Trump will hand the Democrats the Senate and the House. The Republican party is a joke.
- DuPage Saint - Wednesday, Oct 12, 16 @ 3:24 pm:
I bet the young man is from Oswego
- Ron - Wednesday, Oct 12, 16 @ 3:26 pm:
Lol@DuPage Saint!
- Not quite a majority - Wednesday, Oct 12, 16 @ 3:33 pm:
Note to self: ignore LA Times polls FOR EV ER.
- Federalist - Wednesday, Oct 12, 16 @ 3:41 pm:
Talk about bad polling! Wow!
Makes you wonder about other polls as well as if I already didn’t.
- Anonymous - Wednesday, Oct 12, 16 @ 3:41 pm:
Too funny. Moral of the story: when reading polls, don’t ignore the footnotes.
- Responsa - Wednesday, Oct 12, 16 @ 3:51 pm:
==Too funny.==
I’m glad you think it’s funny. The people allegedly overseeing and authenticating this poll should lose their jobs for this.
- Ron - Wednesday, Oct 12, 16 @ 3:55 pm:
http://projects.fivethirtyeight.com/2016-election-forecast/?ex_cid=rrpromo
- Tommy - Wednesday, Oct 12, 16 @ 3:59 pm:
Interesting stuff. The folks running the survey obviously know they’ve been stuck with outlier results for a few weeks, why not change their methodology on the fly??? Is mantaining the same survey group in a daily tracking poll so sacrosanct that they can’t change weighting or drop a participant in midstream? Fako, you out there?
- Someone you should know - Wednesday, Oct 12, 16 @ 4:01 pm:
What do you expect from a TRONC Product?
- Lol @ IL - Wednesday, Oct 12, 16 @ 4:03 pm:
“Trump will hand the Democrats the Senate and the House. The Republican party is a joke.” Senate maybe, house, absolutely not. Too many states are too gerrymandered.
- muon - Wednesday, Oct 12, 16 @ 4:34 pm:
As train111 mentioned, we only know this because this poll is completely transparent. They provide raw data and demographic assumptions. That let Upshot first duplicate the poll results, then recalculate them without the overweighted point. They should be commended for that level of transparency. I wish all polls were as open about their internal weighting.
- wordslinger - Wednesday, Oct 12, 16 @ 5:26 pm:
That LA Times/USC poll has been beyond trippy outlier.
You’d think an honest newsman like Ferro would have gotten a handle on that.
- weltschmerz - Wednesday, Oct 12, 16 @ 6:50 pm:
When Ike was told that half of the American people were of below average intelligence, he was appalled.
- sal-says - Thursday, Oct 13, 16 @ 3:36 pm:
Wonder why sal-says is not a poll junkie? A lot of polls are junk.