Before everyone gets giddy about PA-18

[nitpicker]

Remember in 2016, the final poll average showed HRC ahead by two points in PA, but Dump won.

From a description of the PA-18 population, it may contain what in Virginia I call "Gillespie Republicans"- the sort of people who shun pollsters, but come election day, they show up to the polls and vote Republican.

The latest PA-18 poll does have Lamb ahead by six points, but with an inch of snow forecast it might deter some from voting.

So don't be too surprised if Lamb loses (but I will be happy if he does win).

[Yonnie3]

1. MOE on this latest poll was 5.1

The MOE represents the sample size and doesn't include any other types of error, so I view this poll with some skepticism.

The determining factor will be as always, getting out and voting.

[sharedvalues]

8. Monmouth poll had three turnout models

Turnout is the key and no one has any idea how to model it, so all the polls are just guesses.

[titaniumsalute]

2. Even is he loses by a few % he will have narrowed the Trump gap by about 17 points

That will scare the hell out of every Republican out there running for the House in the Fall.

[scheming daemons]

3. I am here... there is no snow

And I wish there were.

Democrats are more motivated to vote, and snow would keep more casual voters away. I.e. Saccone voters.

[rurallib]

4. any time the vote is on anything but paper ballots

is time to be wary.

[Cha]

5. No one is "giddy" that I know..

And, I personally don't count anything until it's done.

I don't know of a crystal ball, either.. so we don't know what's going to happen. those of us far away who gave money can only hope the GOTV is Spectacular!

[FSogol]

6. Gillespie lost.

Also what are the odds that many here on DU even know someone in PA-18?



An inch of snow in PA? Pssah. Not a factor.

BTW: Concern noted.

[PA Democrat]

11. I am in PA-18. There are snow flurries, that is all.

Personally, I would have driven through a blizzard to vote for Lamb. It's all down hill to my polling place. I could have worried about getting home after casting my vote!

[FSogol]

12. Awesome. GOTV!

[Demsrule86]

15. It has been snowing almost everyday in Ohio going towards Pittsburgh. I live in Ohio an

hour out of PA...in Cortland...Snow won't be a factor.

[DemocratSinceBirth]

19. Really

[NCTraveler]

7. Monmouths trend was actually pretty clear.

It’s very similar this time, just the other direction.

[zipplewrath]

9. Over 50%

The key that alerted me was that he was over 50%. It is often hard to lose when you are over 50%. Hillary was below that, and 538 warned about that. Somebody is going to win, and you can win or lose with 47%, it just depends on how much the opponent gets. Over 50% (of the vote) and you can't lose.

Lamb hasn't gotten over 50% of the vote yet. But that's the way to bet.

[DemocratSinceBirth]

10. Noooooooooooooooooo

Are you suggesting that in a competition with two or more individuals the individual I want to win might lose ?

[LanternWaste]

13. You're confusing simple hope of an amazing upset win with giddiness.

You're confusing simple hope of an amazing upset win with giddiness.

If Lamb even comes close in a loss, it will be a signal that Republicans are in danger even in districts where Trump won handily in 2016, and infuse Democratic possibilities of capturing the House and Senate this fall. The race could also serve as a model for Democrats running in deep-red districts across the country.



"So don 't be too surprised if Lamb wins..."

[Demsrule86]

14. Gillespie lost ultimately.

[Demsrule86]

16. Since this candidate picked for this district...will come close no matter what it seems...perhaps

this is how we regain a majority.

[LexVegas]

17. Hopeful is not giddy. nt

[karynnj]

18. Ahead by two, when she earlier was ahead by more.

Four things -

1) two points was within the MOE. It did mean that she was more likely to win, but it was by no means safe.

2) The MOE refers just to sampling error. Assuming that the poll was of "likely voters", the algorithm for determing who would vote is not itself error free.

3) None of the polls at the end had her above 50 percent. From the exit polls, we know that the undecided broke strongly for Trump. (My completely unscientific theory is that it might be that the undecided were disproportionately Republican leaning but were queasy about Trump. ) (I appologize for right wing site but it was what google found for me on PA 2016 polls - https://www.realclearpolitics.com/epolls/2016/president/pa/pennsylvania_trump_vs_clinton-5633.html )

4) The trend was that Hillary was losing support. Note that most polls ended a half a week to a week before the election.

As to Conor Lamb, one difference is that he DID move above 50%, which Clinton did not - though that election had Stein and Weld meaning that you could win with less than 50%. It is always better to poll higher, but the turnout model is "heuristic" and it can completely change the results of the sample. The MOE does not include any addition uncertainty to reflect potential bias in the likely voter model. I know the common wisdom is that Democrats are engaged and Republicans aren't. THAT assumption is already built in - so if turnout is flatter, the Republican will do better than the model projects. If it underestimated the Democratic enthusiasm or the Republicans are less enthused than assumed, he could do even better.

The thing to remember is that to get to this point, where he may be favored, Lamb has already far out performed what anyone could have predicted.

[okieinpain]

20. agree n/t.