Did Democrats Massacre Republicans on First Day of Early Voting in Nevada? Maybe Not

Update: An earlier version of this post contained erroneous data.  The corrected version is below.

Nevada had its first returns of early voting posted last night and the headline numbers looked very good for Democrats.  Here was the headline and post from Nevada expert Jon Ralston:

Democrats massacre Republicans on first day of early voting in urban areas

The numbers show a 26 percent edge in Clark County and 10 percent in Washoe County. That’s about a 10,000-vote lead in Washoe and Clark combined.
The Clark numbers: 55-29, or 18,388-9,588 (Actual registration: 46-31)
The Washoe numbers: 48-38, or 4,604-3,619 (Actual registration: 38-38)
Yes, it’s only one day. But if it continues like this, Nov. 6 is going to be a very bleak day for Republicans in this state.

I’m not the biggest fan of reading too much into early voting especially on only one day’s data. There is an inherent uncertainty in what the submissions actually mean (how many voters crossed over parties with their vote?) and the sample sizes are really small.  This makes drawing conclusions not much better than just plain old guess work until a critical mass of votes have accumulated. But a lot was missing from Ralston’s report that tells a very different story, namely a comparison with 2008. It is impressive to have a 10,000 vote lead on day 1 of early voting in a state won by 121k votes in 2008. And Clark County is the motherload of votes in Nevada accounting for 67% of the state’s votes between the two major candidates in 2008.   Barack Obama won Clark County by 124k votes, accounting for his entire margin of victory.  So it would be very instructive to see what Clark County looked like in 2008.  Our commenter Paul provided the daily vote total for Clark in the last election:

The Democrats started with a large lead and steadily built on that lead throughout the early voting achieving an aggregate advantage over Republicans by election day of 86,607, approximately 70% of their final margin.

What is harder to see is that Republicans steadily closed the gap with Democrats throughout the early voting days before nearly catching the Democrats on the final day.  On day 1 their vote total was 37% of the Democrats vote and by the final day their vote was 89% of the Democrats vote:

10/18/2008 10/19/2008 10/20/2008 10/21/2008 10/22/2008 10/23/2008 10/24/2008 10/25/2008 10/26/2008 10/27/2008 10/28/2008 10/29/2008 10/30/2008 10/31/2008
Rep % of Dem Vote 37% 45% 56% 51% 54% 56% 61% 59% 58% 68% 69% 72% 68% 68%

The rate of change over those days was a 2.4% increase on average in favor of the Republicans closing the gap with Democrats but unfortunately they were starting from a very low base (37%…yuk).

Day 1 of 2012 was very different despite the misleading reports like Ralston’s above.  It is true that Democrats enjoyed a whopping 55 to 29% lead among ballots returned or voted on Day 1 — a 26 percentage point (pp) margin.  But in 2008 the Democrats enjoyed a 62 to 23 advantage — a 39pp margin.  So the GOP is outperforming by 13 points.  Extrapolating from one day’s data is inherently risky and unreliable but the inferences Ralston is drawing are completely without merit. For fun we are going to construct a 2012 early voting turnout based on the first day’s data and restricting ourselves to the trends in the 2008 results.

In 2008 looking at every day subsequent to the first, the Democrat turnout averages 15,224. With Democrat turnout in the subsequent days averaging 97% of the day 1 result we insert that as a constant going into election day. We then take the Republican turnout and increase it at a rate of 2.4% every day, consistent with the 2008 results.

The final aggregate lead for Democrats is 81,786, nearly 5000 below the 86,607 lead in 2008. This is again with conservative assumptions using the 2008 GOP ground game which was nothing like today’s operation. Obama won Nevada by 121k and based on these assumptions Republicans would eat into that lead by 5k votes, a 4% reduction. That’s not an enormous amount on Day 1 but it is marked improvement for a state party that has struggled over the last few years.  Obviously Team Nevada, the Romney campaign and Senator Heller’s campaign deserve the credit for the improvement but the main point is this was by no means a massacre as reported.

Based on the 2.4% rate of change we see by the Friday before the elections, Republicans would nearly catch up to the Democrats vote total on the Friday before the election, substantially better than their 65-70% average in the final days of 2008.  The key here, again, is minding the gap.

The rest of the state trends heavily Republican with the exception of Washoe County which is a battleground County that President Obama carried in 2008.  The Republican gains in Clark County are even more pronounced in Washoe County where Republicans reduced a 34pp gap in 2008 into a 10pp gap. Republicans increase in turnout in that County was 3x the increase of Democrats on Day 1. Yesterday was no reason for either party to pop bottles of champagne or even start putting it on ice.  The aggregate turnout numbers show both parties are very active on the ground this year achieving a 32% increase in overall turnout. The Democrats are enjoying their typical early vote advantage but a great change is happening this year as it appears Republicans are more than up to the challenge of “minding the gap” even in Democrat bastion Clark County.


  1. damien
    Posted October 21, 2012 at 12:25 pm | Permalink | Reply

    and Ralston tries to pull a “crowley” and it works out just as good

  2. Tom
    Posted October 21, 2012 at 12:37 pm | Permalink | Reply

    It really shouldn’t be a surprise that almost everything from the left is skewed and anti-Romney. Romney will win NV and this election is going to be a mini-landslide for him. Romney will have 350+ electoral votes.

  3. Posted October 21, 2012 at 12:38 pm | Permalink | Reply

    exceptional analysis here 🙂

    There is also the unknown that while unions are undoubtedly aiding in member turnout, we won’t know how even their own folks have actually voted until lection day … given the messages from the likes of Wynn and others about the future of the economy in the US and Nevada in particular.

  4. housebroken dad
    Posted October 21, 2012 at 12:45 pm | Permalink | Reply

    Keith, are you know believing what the rest of us already know, that Ralston is nothing more than a leftist hack?
    Anyway, good start for R/R in NV, let’s see how each day progresses; it will certainly be interesting and entertaining to monitor.

  5. Posted October 21, 2012 at 1:10 pm | Permalink | Reply

    Romney up by 7 in Gallup LV poll, 3 in RV poll.

    • Ron
      Posted October 21, 2012 at 2:13 pm | Permalink | Reply

      Exactly the split in the 2010 mid-term–52% republican, 45% democratic.

  6. margaret
    Posted October 21, 2012 at 1:56 pm | Permalink | Reply

    In 2008, a reasonable assumption was that Republican early voters would vote McCain, but it turns out that was faulty when a lot voted for Obama. This election, we have to again assume that Republicans and Democrats voting early will vote along party lines but I think this will again end up faulty. There will be few Republicans who vote Obama and a surge of Democrats voting Romney. Obama is losing chunks of his voting blocs from 2008. Romney has consolidated most of McCain’s vote, gained a majority of the independents as well some Democrats. If this continues for the next 2 weeks, Romney will win, guaranteed.

    • EpiphoneKnight
      Posted October 21, 2012 at 2:52 pm | Permalink | Reply

      While we’d like to see Mitt win everywhere he can, if he loses Nevada, that may be ok given trends in PA, IA, etc.

    • valleyforge
      Posted October 21, 2012 at 9:38 pm | Permalink | Reply

      Romney may well pull a larger cross-over vote (Republicans typically do), but that doesn’t necessarily apply to the early vote. These folks may be highly-motivated partisans rather than the type who would cross-over on Election Day.

  7. Brian
    Posted October 21, 2012 at 4:30 pm | Permalink | Reply

    Nevada isn’t a vital piece of the puzzle, but I find myself REALLY wanting Mitt to win it just so this smug jerk Ralston shuts up. Being an “expert” doesn’t mean anything when you’re so blatantly slanted toward one side.

    (That and Mitt winning Harry Reid’s state would be awesome.)

  8. rcl_in_va
    Posted October 26, 2012 at 3:48 pm | Permalink | Reply

    Keith, I’m enjoying your blog very much. Lots of great stuff here. In the FWIW category the CC numbers for 10/20/2008 must be incorrect. CC offical results website shows early vote total that day 23,270. Your model shows 42,784. I don’t have the REP/DEM/OTHER numbers but it kinda looks like all three are about 2X what they should be. Now irrelevant, but thougt you might like to know. I’m folowing NV closely and think NV and IA are tough for RR, but are possible and could be huge.

    • Posted October 26, 2012 at 4:27 pm | Permalink | Reply

      can u provide a link? I’m not seeing what you’re seeing

      • rcl_in_va
        Posted October 27, 2012 at 2:45 pm | Permalink

        Sorry for the delayed response; must have missed your reply. It is http://www.clarkcountynv.gov/Depts/election/Pages/ElectionHistory.aspx For the “select an election” dropdown box choose “2008 GENERAL ELECTION (11/04/2008)” and for the “select the type of information” drop down choose “EARLY VOTE TURNOUT SUMMARY”. The selected page shows Daily Turnout Week 1, and the total at the bottom of 10/20 is 23,270, not your 42,784. This does make a 5k delta in the DEM advantage total (I get 81,965 and you have 86,607).

  9. rcl_in_va
    Posted October 26, 2012 at 4:06 pm | Permalink | Reply

    It seems the DEM number is actually the total for that day. If so the numbers s/b D=12655, R=7118, O=3497, Total=23270. Like back in school, working the lab results backwards to get the correct initial conditions.

One Trackback

  1. […] Nevada only started a couple days ago and after the first day of returns I created a rather crude model for how the rest of early voting could go based on the 2008 historical information.  Although the model was basic, assumptions were honest […]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: