More than 70% of Nevada's voters have already voted in the now-closed two week voting period. Nevada is a swing state and could provide the margin of victory combined with Ohio and NH.
Official NV reporting gives the party registration of the voters. PPP poling has cross-tabs on the selected candidates of early voters, and the intentions of the late voters.
These two independent streams can be combined to estimate, corroborate and project actual voting results.
701,845 early votes were cast. These break down to 44% Democratic registration, 37% Republican and 19% Independent. This breakdown is very similar to the 43D-38R-20I base PPP survey result. The 2012 early votes represent 74% of the 2008 vote total. PPP reports an aggregate 74% of the respondents in its survey reported already voting.
So on registration and party affiliation and on voting behavoir, the PPP and the Nevada results are well co-ordinated. However, early voting accounts for 56% of the "active registrations" in Nevada. This indicates to me that the low-likelihood voters are not responsive to the survey, and the total vote and make-up will closely resemble the 2008 turnout. 2008 turnout was 78% or 946,000 out of 1,207,000 active registrations.
The county-by-county early voter participation shows a very similar level of activity for both Democratic and Republican registrations. Nevada can be divided in three voting zones: Las Vegas (Clark Co), Reno (Washoe Co.) and everything else. In LV, 59% of active Democrats and 61% of active Republicans have cast ballots (Independents trail at 47%). In Washoe: 57% of both parties, and 42% of Independents). This pattern of nearly identical participation rates holds for all the subdivisions-- participation rates (judged by active registration rolls) are nearly identical. Participation rates vary geographically (31% in Lincoln County vs. 65% in Humbolt) by party registrations have nearly identical rates of early voting, and Independents trail by a characteristic bracket.
The PPP poll has a distinctly different cross-tab: 81% of its Democrats reported already voting versus 73% of the Republicans (and 60% of the Independents). One explanation is that a substantial fraction of the Democratic active voter registrations are in fact unlikely to vote and did not participate in the PPP polls, are hang-ups, and so drop off the cross-tabs.
The inverse explanation is also possible: self-identified Republican PPP responders are more likely to have delayed voting, and are in fact unlikely to actually go to the polls. They are perhaps discouraged voters, who intend to vote, but will not actually cast a ballot.
There is some evidence for this in the PPP crosstabs on age versus participation and candidate. PPP reports 80% of its 65+ voters have already cast ballots and the proportion declines steadily to the 20 somethings (where only 68% report casting ballots). Among the 20-somethings 62% prefer Obama and in the 65+ cohort on 48% do. This presents the paradox: the age group least likely to have already voted is the most likely to vote Democratic, and the age most likely to be Republican is the most likely to have already voted. The consistent pattern argues that the cross-tabs are not due to random sampling error with the small groups sliced and diced out of the survey, but are a consistent trend.
If we assume that these putative conservative youth intend to vote, but have neglected to do so -- we have an explanation for the paradox. The observed voting behavior (almost exactly matched rates of participation by party) promotes the supposition that the intended action will not occur: PPP is getting low-likelihood conservative youth opinion, that will not actually convert to votes.
PPP reports votes cast for Obama by party affiliation. These are 93% Dem, 10% Republican, and 41% of Independents. These parameters can be used to estimate the current (early vote) totals and margin and possible statewide results. Strictly applied the affiliation/vote cast model calculates 367,279 (out of 701,845) votes for Obama. This puts Obama within 106,000 votes of the 50%+1 total of the 2008 turnout. However this model reports the current vote at 52% Obama, and the PPP poll represents the early cross tab as 55% Obama.
It seems the PPP poll may overstate the current Obama margin as a 55% polling proportion would require about 54% of the Independent vote. All Nevada polls show Obama trailing (sometimes substantially) in the Independent cross-tab.
In conclusion, I believe the current real-world status of Obama in Nevada is a early vote of about 75% of the total turnout with his holding about a 52-47% edge or a 43,000 vote advantage. The race will tighten as the purported day-of vote is more conservative. There appears a block of young, conservative voters that have not voted and may fail to materialize, and their absence will allow Obama to secure Nevada's electoral votes.