VT-GOV: Polling suggests narrow race for governor one month out

[RandySF]

BURLINGTON, VT — In recent weeks, the Vermont Democratic Party has commissioned a statewide poll conducted by Tulchin Research — the firm rated among the most accurate by the Democratic National Campaign Committee (DCCC) independent expenditure program. The poll was conducted between September 23-26 with a sample size of 406 likely Vermont voters.

After months of steady decline in Phil Scott’s popularity, it should come as no surprise that in a head-to-head match up Christine Hallquist and Phil Scott are locked in a statistical tie, with Hallquist at 42% and Scott at 50% — well within the 4.9% margin of error. Of respondents who said they would definitely vote this November, the gap narrows to Hallquist 44%, Scott 47%.

In another notable head-to-head match up, our data finds Bernie Sanders with a 55 point lead over his Republican challenger. Sanders also enjoys a 75% favorability rating.

The Lieutenant Governor, David Zuckerman, has also earned a sizable lead of 38 points over his Republican challenger: Zuckerman 65%, Turner 27%.

Not only does our polling data serve to verify that Governor Scott is losing support and is increasingly out of touch with everyday Vermonters, the numbers further highlight the success and popularity of all Democrats up and down the ballot.

We are encouraged by these numbers and look forward to sharing further analysis of this data by Tulchin Research in the coming days.


http://www.vtdemocrats.org/press-releases/polling-suggests-narrow-race-governor-one-month-out


CHRISTINE HALLQUIST





https://www.christineforvermont.com/

[Grubert]

1. If 8% is a statistical tie...

My passion for polls just ended.

[Drunken Irishman]

2. My first thought haha

I don't think they understand how MOE works.

[RandySF]

5. Yeah that was a big error on their part

But I think this is a race that could surprise everyone next month.

[Awsi Dooger]

7. Not an error

Statistical tie may not have been ideal wording but it is definitely within the margin for error

[RhodeIslandOne]

3. How is an 8% race within a 5% margin of error?

I told myself after 2016 I was done with polls.

[Awsi Dooger]

6. Margin for error is on each characteristic, not the margin itself

I remember going over this with TruthIsAll maybe 15 years ago.

The 42% for the Democrat has a 4.9% margin up or down (37.1% to 46.9%) and so does the 50% for the Republican (45.1% to 54.9%).

That's why an 8 point poll is well within a 4.9% stated margin for error.

This Pew link describes it well: http://www.pewresearch.org/fact-tank/2016/09/08/understanding-the-margin-of-error-in-election-polls/

[Grubert]

8. So a 17% lead would be within the margin of error?

?

[Garrett78]

9. No. If the margin of error is 4.9, a lead of 9.8 or less would technically...

...be within the margin of error. Margin of error is plus or minus 4.9 for each side. So, it might be a lot closer than 8 points, or it might be a blowout.

As of right now, Scott is a clear favorite.

[Grubert]

11. You said one could have 54, and the other 37

That’s therefore within the margin of error.

[Garrett78]

12. Awsi said that actually. The point being...

...you get those 37 and 54 numbers if you subtract 4.9 from Hallquist and add 4.9 to Scott. A margin of error of plus or minus 5 for a candidate with 40% means that candidate is somewhere between 35 (worst case) and 45 (best case).

[Awsi Dooger]

17. Correct...the margin for error is a doubling of stated number, so 9.8% in this example

It was spelled out in the Pew link. I should have mentioned it specifically in my earlier post.

When you apply the margin for error to each characteristic, the crossover will apply as long as the margin is < double stated margin for error. In this case the Democratic 42% can rise to 46.9% on theoretical high, and the Republican 50% can drop to 45.1% on theoretical low. That is how the article was able to say "well within margin for error."

If the margin had been 10%, then that's outside double the 4.9% so it is outside margin for error.

Again, I went over this with TruthIsAll maybe 15 years ago and he threw an absolute fit, because he had just posted a lengthy thread on the 2002 election results claiming fraud in one race after another since he was asserting they finished outside margin for error. But he didn't understand margin for error. I pointed out that the vast majority of races he was claiming outside margin for error were actually well within margin for error. I had to provide one link after another toward how margin for error is calculated, before he would accept it.

BTW, regarding this Vermont race it is considered so lopsided most of the betting markets are not even offering a price. That is partially due to lack of polling. But it is hardly one of our prime pickup opportunities.

[RhodeIslandOne]

13. Wow.

I’m sorry but that sounds like an awful poll.

[Rhiannon12866]

4. I was in Vermont this past weekend and it's awash with political signs!

And I could tell from my friend who lives in Vermont which were Republicans - she hadn't heard of them...

[OilemFirchen]

10. A 4.9% MOE on a poll with an exponentially tiny sample size is statistical malfeasance.

Tulchin Research should be embarrassed by its release.

[Garrett78]

14. The MOE is that high *because* the sample size is so low.

Vermont is a small state, and it's likely they simply weren't able to get any more responses in the allotted time frame.

[OilemFirchen]

16. We don't know what the confidence level is.

If it's as low as I suspect, then it was incumbent upon them to either increase the sample size, re-run the poll until they reached a satisfactory CL, or simply not produce the results.

We can't see the internals, so it's anyone's guess.

[brooklynite]

15. Lousy math...

IF we assume that the poll for Scott is off by the ENTIRE MOE (an extremely low probability outcome), that means he's at 45. That however DOESN'T mean that Hallquist is UP by 5% and thus they're tied.

I see nothing in the dynamics of the VT-GOV race that says this is competitive, any more than MD or MA are. A fair number of Sanders voters will be perfectly happy to cross over and vote for Scott.

Republican Gov. Phil Scott's approval rating dropped sharply in the second quarter, but he has maintained solid support among Democrats.

In fact, Scott fares 20 percentage point better among Democrats than his own party.

https://www.burlingtonfreepress.com/story/news/2018/07/30/vt-insights-gov-scott-drops-poll-but-has-democratic-support/862213002/