Trump +2 in CNN/ORC National Poll

[Adam Griffin]

It's 44-41-9-3 to Clinton on RV rather than LV. Modelling who's actually going to vote (see Reuters) is causing much of the divergence in national polls.

Wasn't the initial switch to LVs showing Clinton doing better than she was doing with RVs? I remember people making a big deal about it some weeks ago because it was so odd to see.

[Adam Griffin]

WTF?

They messed their sample up, only 28% are Democrats compared to 32% Republican and 40% Independent

No, that's the party ID numbers, which you really shouldn't try to unskew.

Meh...that's usually the line, but that's also because it is usually conservatives who fail to understand that Democrats have outnumbered Republicans in terms of party ID in every presidential election since at least the 1950s* (and likely much longer than that).

*2004 was a tie

[Adam Griffin]

WTF?

They messed their sample up, only 28% are Democrats compared to 32% Republican and 40% Independent

No, that's the party ID numbers, which you really shouldn't try to unskew.

Meh...that's usually the line, but that's also because it is usually conservatives who fail to understand that Democrats have outnumbered Republicans in terms of party ID in every presidential election since at least the 1950s* (and likely much longer than that).

*2004 was a tie

So should we unskew this poll?

No, we should throw it in the fycking garbage where it belongs. No credible poll is going to find a national electorate that is R+4, self-identified or otherwise.

[Adam Griffin]

^^^ For that matter, a 40% independent electorate among likely voters is pretty ridiculous, too.

[Adam Griffin]

No, we should throw it in the fycking garbage where it belongs. No credible poll is going to find a national electorate that is R+4, self-identified or otherwise.

28-30 Aug, Fox news Clinton +2
Among RV
D 41%, R41%, I 16%, DK 3%

I'm not sure what your point is there...? That seems like a pretty silly poll as well if those numbers are party ID numbers: registered voters are even more "independent" than likely voters. Presumably the chronic references to "40% of the country being independent" are referring to registered voters as a whole; likely voters have tended to be closer to 30%. To have them at 16% significantly over-represents Democrats and Republicans even when measuring registered voters. However, at least this poll's top-line numbers are in line with aggregate polling overall.

Having an election where party ID is tied is obviously possible (see 2004), albeit not very likely in this particular election. Having an election where the electorate is R+4 after two elections where registration has been D+6/7 in such a polarized political climate is not.