Remember that national Reuters poll that appeared to show Hillary at 48 and Bernie at 45?

[pnwmom]

It was a pretty shocking result and didn't fit with any of the other polls released during the same period of time.

I've been trying, unsuccessfully, to find the internals on how that poll was conducted, to learn why it wasn't included in RCP's Averages of Polls, and another DUer pointed me to the answer.

IT WASN'T AN ACTUAL COMPLETED POLL.

It was the results of a single day conveniently snipped out of a Reuters rolling 3-day WEB poll. The actual completed poll put Hillary 15 points ahead of Bernie.

Any time a poll is wildly out of line with others, it's a good idea to check why that might be so, before trumpeting its results too loudly.

THANKS TO ERICH BLOODAXE BSN for emailing Huff Post with his question and posting the answer here:

http://www.democraticunderground.com/12511165494#post6


http://elections.huffingtonpost.com/pollster/polls/ipsos-reuters-23665

Ipsos/Reuters (Web) 1/30-2/3)

2) 2016 National Democratic Primary

Asked of 619 Democratic registered voters
Hillary Clinton (D) 54%
Martin O'Malley (D) 2%
Bernie Sanders (D) 39%
Wouldn't vote 5%

[NurseJackie]

2. Little wonder that so many of Bernie's fans like ONLINE POLLS.

[mak3cats]

3. Correct...

And a really good example of fair play by a Bernie supporter. We need more of this all the way around, IMHO.

[RiverLover]

4. Reuters poll?

According to Quinnipiac, “In the Democratic race nationwide, former Secretary of State Hillary Clinton has 44 percent, with Sen. Bernie Sanders of Vermont at 42 percent, and 11 percent undecided. This compares to a 61 – 30 percent Clinton lead in a December 22 survey by the independent Quinnipiac University Poll.”

http://www.politicususa.com/2016/02/05/bombshell-poll-shows-bernie-sanders-winning-the-hearts-of-democratic-voters.html

Are they the same?

[pnwmom]

10. No, they're not the same.

This Quinnipiac poll, unlike the Reuters, WAS included in the RCP poll averages. Overall, averaging all the major polls together, Hillary was 14 points ahead. But the Quinnipiac did give them a tie.

http://www.realclearpolitics.com/epolls/2016/president/us/2016_democratic_presidential_nomination-3824.html

[reformist2]

5. On noes!!! Not an online poll??!! Shun! Shun!!!!1!

[jillan]

6. Again you are looking at OLD POLLS! The poll showing them tied came out on the 5th

You can just look at what you posted - O'M is still in the race.

[pnwmom]

7. That Reuters "poll" from the 5th was JUNK. You can't pull one day's worth

of data out of an online running poll and call it a poll.

The RCP has posted the most recent results of all legitimate polls. That snippet you are calling a poll didn't qualify.

[jillan]

9. RCP has not been updated since the 4th! Again making my point!!

[riversedge]

11. Thanks so much for the update. and the work doing this.

[thomservo]

12. Reading the article...

it appears to be the rolling average from 2/2 to 2/5.
http://www.reuters.com/article/us-usa-election-poll-idUSMTZSAPEC253J02S1

[pnwmom]

15. That's what the article said but not, as you know, what actually happened,

according to the email that Erich Bloodaxe posted.

And the other confusing part was that that article had an old link at the bottom.

[thomservo]

8. The data you posted in the op is older...

The data you posted is a few days older, 1/30-2/3. I think the results showing Sanders down 3 was the rolling average from 2/2-2/5.

http://www.reuters.com/article/us-usa-election-poll-idUSMTZSAPEC253J02S1

[pnwmom]

13. Thanks for posting that. I had just checked and they were still posting

the results from the mid January poll at the bottom link -- now they've updated it.

The fact remains is that this is still an online poll, and the science behind them is still being developed. There's no way to compare its accuracy with traditional polls.

[thomservo]

14. fair enough.

thanks for the thanks.