Quote:
Originally Posted by fatkid
God that link is so unreadably formatted and argued that it took me a while to figure out what the argument is in Nevada. It boils down to this; here's a side by side comparison of 2012 and 2016 first week early voting. These numbers are for party registration, so we don't know what people voted, it has to be inferred.
2012:
| Mail | Early | Total | % |
---|
Dem | 22956 | 139281 | 162237 | 45.3% |
GOP | 23491 | 110094 | 133585 | 37.3% |
Ind | 8613 | 53610 | 62223 | 17.4% |
2016:
| Mail | Early | Total | % |
---|
Dem | 15925 | 131076 | 147001 | 44.4% |
GOP | 17864 | 101935 | 119799 | 36.2% |
Ind | 8074 | 55998 | 64072 | 19.4% |
This shows essentially no change from 2012. Both parties have lost registered voters and turnout is lower across the board, but the numbers are similar.
So he then unskews these affiliations into actual votes, using a poll crosstab showing how many Ds are voting R and vice versa, as well as which way independents are breaking. Unfortunately he uses Axiom, which is a GOP push polling outfit that is not included in RCP or 538 averages. Here's my own unskewing effort of the 2016 numbers using a
recent Emerson poll that had Clinton +2, which is roughly the RCP average. (Bonus weirdness: in Emerson's sample there were 0% of Dems either unsure or voting Stein, but 0.2% of registered Republicans were voting Stein).
| Reg Dem | Reg GOP | Ind | Total | % |
---|
Total | 147001 | 119799 | 64072 | 330872 | 100.0% |
Vote Dem | 117748 | 13657 | 20375 | 151780 | 45.9% |
Vote GOP | 22197 | 90808 | 27935 | 140940 | 42.6% |
Other | 7056 | 6589 | 0 | 13645 | 4.1% |
Undecided | 0 | 8745 | 15762 | 24507 | 7.4% |
So a reasonably comfortable lead for Clinton, assuming the undecideds are not breaking very heavily for Trump. Note that this is a pretty generous way to do things, I could have simply noted that party affiliation turnouts are similar to 2012 and that Obama won by nearly 7 points in 2012. The analysis above tries to account for the fact that maybe more Dems/Inds are voting GOP than Reps/Inds are voting Dems, when compared to 2012. The discrepancy between polling numbers and the substantially higher early voting lead indicates this might be the case, but it could also simply be that the Dem GOTV operation is superior.
I hope you enjoyed this edition of Math You Do As A Liberal To Find Out The Actual Answers To Things.
Edit: Just checked the crosstabs of another poll (Marist) and it's broadly similar to Emerson, a bit more pro-Dem. Independents break for Trump and more Dems vote GOP than GOPers vote Dem in both crosstabs, but it's not enough to overcome the discrepancy in affiliations among voters.
Last edited by ChrisV; 10-29-2016 at 10:19 PM.