Saturday, January 12, 2008

Get More Votes Next Time

The sole reason we are having this argument about the accuracy of the New Hampshire vote count is because of the media's out-of-control hype of Golden Boy Barry, both before and after the Iowa caucuses. He had high poll numbers the days before the NH primary, leading supporters (and pretty much everyone else) to expect a big win. Then the actual poll results reverted back to the numbers we had been seeing prior to Iowa, giving HRC a solid lead.

Barry's surrogates are now throwing shit at everything to see if they can prevent HRC from getting any momentum from her win the way he benefitted from Iowa. The cries of racism and vote fraud need to be understood in this light. No, your candidate lost because the majority of voters preferred someone else, not because New Hampshire voters are racist or because the scantron machines were hacked. Really.

I have now crunched numbers for two days, and though I know what they say (people in urban and very rural areas voted strongly for HRC), that will not stop the Obama campaign from squalling and throwing things to try to confuse the matter. Make no mistake - this is a campaign tactic, not a call for justice. They don't want a recount because it won't give him the votes he needs, but they do want this to remain in the political atmosphere so they can play poor widdle victim Barry having his votes stolen by the (white racist) bitch HRC.

Numbers Crunched

First and foremost, let's be very clear that every vote cast in New Hampshire was a paper ballot. The ballots themselves are not at issue. In addition, they are all paper marked with some kind of pen. There are no punch card ballots with chad troubles, and no electronic only digital votes. The only difference is whether the ballots for that precinct (yes, yes, I know NH votes by town, but I'm going to call them precincts.) are tabulated by hand or run through a scantron style machine.

In precincts where ballots were counted by hand, Obama had a vote advantage of 4.05% over Hillary. In precincts where ballots were counted by a scanner, Hillary's margin over Obama was almost exactly the same - 4.42%. However, that miniscule 0.37% HRC advantage becomes a powerful victory when you realize that 78.5% of all ballots are read by scanners. More precincts count by hand, but vastly more numbers of voters are in the precincts that scan - and that is the reason large precincts use scanning.

All countiesVotes Cast% of VotesHRCHRC %BOBO %% Difference
Hand - 13061,86521.50%21,49634.75%24,00238.80%-4.05%
Scan - 94225,91878.50%90,75540.17%80,77035.75% 4.42%

The question that has to be answered is why is there a pattern of voter preference that shows up between ballots that are counted by hand and those that are counted by scanner? The reason Barry & Co. want you to believe is that there was massive vote fraud on an unprecendented level throughout New Hampshire, involving the undetected collusion of hundreds of state polling officials. Of course, this is diametrically opposite of the original excuse for Obama's embarassing loss, which was that the vote was an accurate reflection of the racists of New Hampshire who lied to opinion pollsters and then voted against The Golden One.

However, as I pointed out in last night's post, that only looks at a single point of correlation (counting method and candidate percentage). It fails to look at whether the counting method is itself a signifier rather than a cause of a vote difference. Please note that I am not saying the scanned count is 100% accurate. It probably is not. In a few precincts, hand counts might break a tie vote, but the overall proportion of the vote will not change, even if they do a hand count of every ballot. The usual problem in those machines is failure to read a vote, so there may be undercounting. Should there be a full recount, I suspect a few hundred votes will be discovered.

One of the things I did today was tried to compare the ways in which both candidates won, rather than contrasting HRC's wins with Barry's losses, which is what is being done to try to explain away or excuse Golden Boy's loss.

There are a number of very important factors that the "VOTE FRAUD!1!1!1" screamers fail to take into account. One is geography. Hand counting and scan counting precincts are not evenly distributed around the state. They are clustered in various counties. Rural areas and exurban enclaves with small populations tend to do hand counting, simply as a matter of volume. Only four places with fewer than 500 cast votes use scannners. Conversely, precincts with large populations - the industrial areas of the state - overwhelmingly use scanners. This map on the New York Times illustrates the candidates' success. You need to scroll down the page. Be sure to click on the "Margin of victory" tab as well. The green counties are where the hand counts are concentrated. The exception to this is Merrimack County right in the center of the state - where 88.99% of the ballots were scanned (26,099) and where Obama won. If the "fix" was in, why didn't he lose that the way he lost Strafford, which also had 88% scanner read ballots?

Another factor is demographics. Large urban areas are more likely to have high concentrations of core the constituents who vote for HRC. Again, refer to the NYT map. Her votes are heavily concentrated, much in the way that the Democratic national vote on a map of the US are smaller islands of blue against the great, empty sea of Republican red. HRC won the big cities, one small upstate county, and then contested the two central counties, Belnap and Merrimack, winning the first and coming close on the second. Obama, conversely, held the western counties pretty well, but couldn't break into the coast. Belnap was a true battleground.

Confounding the claims that the scanned ballots were misread is the simple fact that Obama won decisively in many precincts where scanners were used. For example, in large precincts, with 1000 or more votes recorded, Obama won only 22 of 60 precincts where votes were scan counted. However, he won 21 of those 22 (95+%) by 40% or more. NONE of his hand-counted large precincts gave him those kinds of numbers. Conversely, HRC went into Obama territory in the big hand-count precincts, for example winning Franklin in Merrimack County by a margin of 19.36%. In large hand-count precincts, when they won, Obama won by a margin of 10.8% and HRC won by a margin of 8.8%. In scanner count big precincts, Obama's win margin was almost identical - 10.59%. HRC improved her percentages and won by an average margin of 11.58%. The point I ma trying to make here is that Obama was winning in scan-counted precincts, but he was not winning in large urban areas at the same volume and margin as HRC.

In medium sized places, Obama prevailed in the hand counts, but these counts mostly took place in his counties - Carroll, Cheshire, Grafton. His margin of victory in medium hand count was 10.79%, nearly identical to his 10.8% margin in the large precincts. HRC fell slightly, down to 6.0% win margins. Where Obama fell flat was in medium sized locations that did scan-counting, which were mostly in HRC's counties of Rockingham and Belnap. In those places HRC won by an average margin of 11.81% - almost identical to her large precinct margins - but Obama's victories were much smaller, only reaching an average margin of 5.01%, even though he won only two fewer precincts than HRC. This means that even when he beat her, her vote total was still high, adding to her state-wide victory. In short, Obama did not win large enough victories in the medium sized locales in HRC's stronghold counties. When he did win, he rarely took 40% or more of the vote.

There were 91 small precincts. In small locales, HRC did better in the hand counts than in large locales, with 63% of her victories garnering her 40% or more of the vote. Her winning margin average bounced up to 10.44%. Obama also went up in his margins, to 14.79%. These precincts were 2-1 in Obama counties, so it says to me that both candidates did better than their rivals in the more initimate locations. It was a VERY tight race, as shown by two small, hand-count precincts tieing in the vote between these two. The 5 small places that scanned their ballots split 3-2 for HRC. Ironically, HRC won only one location in her county (Coos) but took two of Obama's locations in Grafton.

In short, as soon as you start looking at the mix of hand and scan counts in the context of geography and urbanization, it is clear that there is no identifiable pattern of vote fraud. Sometimes HRC won big, sometimes Obama won big, and location was a better indicator of who would win than the counting method.

Then we get to demographics. I'm relying on the exit poll found here at CNN. HRC won the majority of women, as everyone knows. But she also won people who think she would be the best Commander-in-Chief, people who think the debates are very or somewhat important, and people who are very angry at the Bush administration. These are people for whom policy details matter, and who are uncertain about stability and security. People who are very worried about the economy support her. People who are regular voters support her. People who feel they are falling behind and are worried about health care support her. Women, union members, and lower income people support her.

Where do you find people like this? In the geographic locations where HRC won the bulk of her votes. These are the people who have always supported HRC, and they rallied to her in New Hampshire.

Instead of inventing excuses for why your candidate lost, and especially before you go inventing wild theories on the moral and political corruption of hundreds of campaign workers and polling officials (and thousands of New Hampshire voters), why not just accept the truth - Hillary Cliton won the primary by turning out her core constituency in greater numbers than any other candidate. As Big Dog said to Jon Stewart when Stewart asked him to explain how to beat Hillary - you have to get more votes.


No comments: