National horse race polls are meaningless, and robo-call polls are the worst, but if you must talk

[Attorney in Texas]

about the new national robo-call horse race poll by Emerson College, you should at least talk about this:

Sanders is up 6% and Clinton is down 3% since the last meaningless national robo-call horse race poll by Emerson College

[JaneyVee]

1. All polls matter.

[Scootaloo]

6. Except the ones that put Sanders ahead on NH, because white people

and of course except for the ones in Iowa that nudge Sanders in the lead, because caucuses make Iowa a fascist dictatorship hellhole that doesn't count.

And the polls conducted online, even if they're not the self-selection variety. because the internet doesn't matter. Not like landlines and lonely retirees who - apparently - make up 42% of the voting public.

[JaneyVee]

8. Those matter too, as they reflect one state polling.

I'm not the one denying polls.

[Scootaloo]

9. Not in this series of posts, at least.

[JaneyVee]

11. Which one did I deny?

Not including internet polls.

[Scootaloo]

17. The point is towards Clinton supporters at large

I've seen New Hampshire dismissed as meaningless "because of demographics" (i.e., because it's predominantly white.) I've seen Iowa dismissed for the same reason, and also because 'caucuses don't count." And yeah, the endless panning of every internet poll, no matter the methodology (by the same people who think sampling elders 65+ at 42% is representative and strong methodology)

[Thinkingabout]

2. Robocalls are bad, online are not scientific

[upaloopa]

3. If you search you'll probably find some other

subsets that are in Bernie's favor.

Most likely Bernie wins in the most click poll clickers polled.

[Attorney in Texas]

4. Iowa and New Hampshire likely voter polling with a good sample size and a reliable methodology

will start to become closer to predictive (but still as much as 10% or more divergent from the actual results) as we progress into the latter half of January.

Until then, the polls are a snapshot in time capturing a moment when the voters are not voting (and in the case of national polls, capturing a moment in time for a hypothetical vote that will never occur).

If you must fixate on the national polls, you should ask why does Clinton do so much better in robo-call polls (like Emerson's) compared to how she does in live-phone (cell and landline) polls conducted by pollsters who conduct the most polling according to the most well established methods:



Clinton has almost a 40% lead in the Emerson robo-call poll but she has only a 17% lead in the most frequent national horse race pollsters who do live phone polls of likely voters.

This is a fairly consistent phenomenon (and it applies to Trump and Clinton).

What is behind the phenomenon that Clinton consistently does better in robo-call polls as compared to more traditional live phone polls?

[upaloopa]

5. Hillary's lead is similar in all national polls

I think if all polls are similar then they all are pointing in the same direction.

Nationally Hillary is around 55% to 60% and Bernie is around 30%. It's been that way for months now.

[Attorney in Texas]

7. Except ... no that's not correct at all. Her lead in the NBC/WSJ poll is LESS THAN HALF her lead in

the Emerson robo-call poll.

A 20% spread between one poll and the other means that both polls cannot be even close to correct. One poll (or perhaps both) are certainly garbage (and both are meaningless in that they are taking a snapshot of a national vote that will never occur).

[JaneyVee]

10. This poll was post-debate.

[Attorney in Texas]

12. Are you deliberately misrepresenting or are you unable to interpret what you are commenting about?

The Emerson robo-call poll is based on data gathered beginning on December 17 (mostly before the debate).

[JaneyVee]

13. And ending on December 20th.

Oopsie!

[Attorney in Texas]

15. And when was the debate? When was most of the data gathered? What were you representing?

[JaneyVee]

16. The debate was on the 19th.

[Attorney in Texas]

18. So 3/4 of the data was gathered before the debate, right?

[NCTraveler]

14. K&R

Love the enthusiasm.