NY-20: Internal Dem Model Predicts 210 Vote Murphy Win

I ran a spreadsheet myself... Here are the results (4.00 / 4)

Using data published by the Times-Union about
win percentages by county, and numbers of absentee
ballots mailed and returned so far in each county,
I ran an extremely simple spreadsheet, and found
the following... Note that this is a very blunt
(crude) method of projection. I used the rounded
percentages for each candidate in each county,
and multiplied that percentage for each times
the number of ballots returned in that county
plus 10%.

So:

Assuming that each county gets an estimated 10%
more ballots returned than they already have (on
average, they've gotten 63% back already, which
is high), Tedisco needs to perform a bit more than
1% better (+1.135% to be exact) than he did among
March 31st in-person voters in order to pull even
with Murphy.

If Murphy and Tedisco get the same percentage of
support from absentee voters as from the general
electorate, Murphy would actually pick up 158
votes in this scenario; the ballots received thus
far derive very slightly more from "Murphy counties"
than "Tedisco counties," based on the figures above.
       My model thus would put Murphy up a total of 183,
       not far from the apparent internal party projection.

Sidenote: I didn't run a breakdown of military
ballots, for the simple reason that I don't
actually assume that military voters will break
differently than the general public, despite
various assertions of that idea in the media.
(My recollection is that Obama, for example,
actually did better among military voters than
McCain.) I also don't assume that older voters,
       who may be overrepresented in the absentees,
       will vote differently from the general public;
       or, if they do swing more to Tedisco, they may
       be offset by a large proportion of NYC weekender
       voters in the Dutchess and Columbia absentees.

Take all that with a grain of salt, of course.
Generally I'd assume that absentees break very
similarly to the overall electorate. But a
very slight shift could change the result. This
is basically a coin-toss, with numbers this close.

---

by: Hudson @ Wed Apr 01, 2009 at 21:45:57 PM UTC

Beware the NOP's (4.00 / 1)

In Columbia County, the NOP's make up roughly 20% of the electorate.  Not sure if that % translates to the rest of the 20th District, but assuming that it does, then we could have something like 1600 NOP's among the absentees.  I'm predicting that they will break strongly for Murphy, perhaps on the order of 1000-600.

---

by: global yokel @ Wed Apr 01, 2009 at 22:45:51 PM UTC

In Columbia County, NOPs... (4.00 / 1)

... are often Dems or Dem-leaning folks who didn't want to register GOP but couldn't assert their true feelings for fear of job repercussions. Seriously, until very recently (and even still) the top-down machine patronage mill operation has been very strong here. I know people to the left of me who are registered REP or NOP because of their jobs. Other NOPs are often students or motor-voter registrants. And then there are a few true independents, who if anything will break 50-50. So I don't really worry about them.

---

by: Hudson @ Thu Apr 02, 2009 at 01:50:39 AM UTC

[ Parent ]

Both projections ignore affidavit ballots. (4.00 / 1)

Which will be in the hundreds, at least.

---

by: Amherst Guy @ Thu Apr 02, 2009 at 00:05:09 AM UTC

Not ignoring affidavit ballots... (4.00 / 1)

... I just haven't seen any stats on them yet. These will likely be only a tiny fraction of the votes yet to be counted.

---

by: Hudson @ Thu Apr 02, 2009 at 01:45:48 AM UTC

[ Parent ]

If by a tiny fraction (4.00 / 1)

you mean 10-30%.

---

by: Amherst Guy @ Thu Apr 02, 2009 at 02:22:22 AM UTC

[ Parent ]

Do the math. (4.00 / 1)

There are already about 6,000 absentees in. Let's say conservatively that there are 6,500 by the time the deadline hits.

30% of 6,500 would be almost 2,000 affidavit ballots. Not gonna happen.

I doubt that there will be even 500 (less than your 10%) that were filled out, and the majority of those will not prove valid.

So in the end I'm estimating in the end you'll have about 200-250 counted affidavit ballots. Maybe at best someone picks up a 10 vote advantage there. Now in theory the election could come down to 10 votes, but the absentees are where this will likely be won or lost.

But if you have some actual stats on affidavits on hand, have at it.

---

by: Hudson @ Thu Apr 02, 2009 at 02:52:31 AM UTC

[ Parent ]

I'm gonna put the number filled out (0.00 / 0)

at 2.6 per precinct, the standard in an odd numbered year election, to which this appeared to have similar turnout.

---

by: Amherst Guy @ Thu Apr 02, 2009 at 03:17:08 AM UTC

[ Parent ]

bah (0.00 / 0)

1.6 per precinct.

---

by: Amherst Guy @ Thu Apr 02, 2009 at 03:17:28 AM UTC

[ Parent ]

Agree with Amherst Guy (0.00 / 0)

(No surprise there) that affidavits could shift numbers a bit.

But I REALLY doubt the veracity of this "model." I read elsewhere that the well-meaning Democrat who cooked this up counted the number of ballots returned by county, and then allocated vote share on a proportional basis to that county's election night returns.

That normally isn't a bad way to go, as paper generally mirrors election day/night returns, but it seems this "model" overlooks partisanship, gender, income, etc. in evaluating voter choice, things that probably should be considered when predicting outcome with such a tight margin, and especially since we are trying to raise money, etc.

For example, Liz Benjamin wrote in her blog last night:

Of the 10,055 absentees, the Repubicans have a roughly 600-ballot edge, according to the board. But of the 5,906 received to date, 798 more came from Republicans.

Since then, Irene Jay Liu over at CapCon updated the counts of absentees returned to show 6381 ballots are now in, but she has not reported on partisanship.

A roughly 800 voter partisanship advantage is enough to make my stomach churn a bit, as Murphy now has a 25 vote lead.  Obviously, not "all" ballots will be counted - and our lawyers will be doing everything to knock out their voters, and vice versa, but this thing is tight as a tick.

We might have a Congressman for NY-20 by Memorial Day . . .

---

by: Lead Dog @ Thu Apr 02, 2009 at 01:04:09 AM UTC

Party affiliation doesn't mean as much as it might seem. (0.00 / 0)

As has been stated ad nauseam, the Republicans had a very large party registration advantage among the general electorate, yet the vote split 50-50 on the machine.

Arguably the GOP affidavits might be more hard-core oldschool Republicans. But my experience of intense local absentee fights is that they track closely to the machine vote. In fact, I can't think of more than a handful of cases where the winner on election day didn't prevail.

---

by: Hudson @ Thu Apr 02, 2009 at 01:48:32 AM UTC

[ Parent ]

Don't confuse enrollment with turnout (0.00 / 0)

If someone could show me the partisanship of those who actually voted at the machine, then we could create a model or at least a projection of what the paper might look like, but that information is typically not available for weeks if not months after an election.

My guess is that Democrats and Independents/Blanks slightly outperformed the GOP/CON voters as a percentage of turnout by party.

It's also helpful to know if Tedisco underperformed among REP and CON voters, because any apparent paper advantage would then mean a lot less.  But if his support matched the combined "red" turnout, his relatively large paper advantage could be problematic.

From a machine results perspective, it is likely that enough independent and unaffiliated voters continue to support the President, and voted for Murphy getting him to parity.  However, if among the paper, minor party or unaffiliated voters participated at a disproportionately lower rate to major party registrants there may be a problem.

Typically, the major parties and campaign committees run absentee voter programs, and the minor parties and blanks don't have resources (or real party infrastructure) to chase their own paper.  There may be other reasons, which others can speak to about why major party registrants do or do not participate at a greater rate.

Again, the "report" indicated a 798 voter partisan advantage for the GOP, but we don't yet know the breakdown of minor parties and unaffiliated voters among the paper universe (which has grown).  If we were able to compare the partisanship of the machine results to the partisanship of the paper, we would probably be able to compile a meaningful model.

BTW, I am not discounting your experience or the points you are trying to make, but 25 votes is a pretty small number in the context of the "reporting." I concede, however, there are too many unknowns, and we need a better read on the methodology of the modeling as well as more information on the composition of the electorate at the machines and among the paper.

---

by: Lead Dog @ Thu Apr 02, 2009 at 15:56:34 PM UTC

[ Parent ]

I Have Bad News. (4.00 / 1)

My sister saw a guy arriving from Laguardia at the Albany Airport and swears it was Roger Stone, the guy who stole Florida for Shrub in 2000.

This is the same guy who left threatening messages for Elliott Spirtzer's elderly and ailing father.

Travelling with Stone was a heavy set guy with very short cropped hair and steel rimmed glasses.

Waiting to get off of the flight, she looked at the tag of his brief case and it said Michael Caputo.

Caputo has been exposed as a Stone henchman at the Albany project.

Stone was wearing one of those suits that made him look like a gangster. His hair was kind of silver and he wore dark glasses.

My sister followed them to the curb where Caputo smoked nervously in the few minutes that they waited before a long black limousine pulled up and wisked them away.

WHICH COUNTY CLERKS OFFICE DO YOU THINK THEY'RE HEADED TO?

The people in the 20th District are not Tedisco people. Little Jim doesn't even live in the district.

These county clerks are  

---

by: Phil Rizzuto @ Thu Apr 02, 2009 at 04:18:01 AM UTC

I Have Bad News (0.00 / 0)

I have bad news.

My sister saw a man arriving from Laguardia at the Albany Airport today ands she swears that it was Roger Stone, the guy who stole Florida for Shrub in 2000 and who threatened Elliott Spitzer's elderly, ailing father on a phone message machine.

Stone was traveling with a heavy set man with closely cropped hair,  steel rimmed glasses and light stubble.

My sister looked at his brief case while they were waiting to get off of the plane and it said Mike Caputo.

Caputo has been exposed as an associate of Stone's at the Albany Project.

My sister followed them to the curb. Stone was wearing one of those suits that makes you look like a gangster. His hair kinda' silver in color, and he wore dark glasses

Caputo smokes nervously until a black limousine pulled to the curb and then whisked them away.

WHICH COUNTY CLERKS OFFICES DO YOU THINK THEY ARE HEADED TO?

The repubs in the 20th District are not Tedisco people. THETY ARE JOE BRUNO PEOPLE.

Bruno has used Stone to do his dirty work before.

These scumbags are up to their usual dirty tricks and will stop at nothing to steal victory from Scott Murphy.

The democrats need to lawyer up.

Any activists who sees Roger Stone and his corpulent sidekick should post it on line at The Albany project or another progressive web site.

If you haven't read the Weekly Standard profile of Roger Stone by Matt Labash you need to read it.

This guy is a maniac who will stop at nothing!

Don't let them steal this one from Scott Murphy.

Signed, ObamaMan

---

by: Phil Rizzuto @ Thu Apr 02, 2009 at 04:29:45 AM UTC

The Absentee/Military/Federal (AMF) ballots sent & rec'd. are... (0.00 / 0)

not as heavily concentrated in Saratoga County as the machine voters were, which may help Murphy hang on.

Whereas machine voters in Saratoga County represented about 36 percent of the total, only about 25 percent of the AMF ballots sent were from Saratoga County.

Similarly, whereas about 58 percent of the machine voters came from Counties that favored Tedisco, only about 47 percent of the AMF ballots were sent from those same counties.

Thus, the geographical origin of the AMF ballots might help to offset some of the partisanship and age characteristics of the recipient voters that will likely accrue to Tedisco.

---

by: teaparty2 @ Thu Apr 02, 2009 at 05:03:31 AM UTC