I got an e-mail today about an interesting analysis of the vote in New Mexico for 2004 & 2006. New laws meant that all precincts had optical scanners & paper ballots in 2006. In 2004, some precincts did and some didn't. Some diligent number crunchers have analyzed the numbers.
Turns out that the high undervotes in minority precincts disappear when paper ballots are used, instead of no paper trail electronic machines. The 2-page pdf this comes from can be found here.
DRE = Direct Record Electronic, i.e. "black box"
Cross posted to ePluribusMedia
A new report, based on official 2004 and 2006 New Mexico election data, shows a dramatic difference in undervotes in Native American and Hispanic precincts, depending on whether they voted on paper ballots or on Direct Record Electronic (DRE - often known as touch screen) voting machines.
Activists in New Mexico have long been concerned about how our elections have run for years. Thanks to their hard work, the legislature required all precincts to get paper trail ballots in time for the 2006 midterm election. We've got a ways to go on audits, but this was an important step forward. At least there's now something an audit can be performed on, and something to be recounted should the need arise.
The 2006 election has been run, and the bar graph above summarizes what was found. There were much higher rates of undervotes in predominantly minority precincts - especially Native Americans ones. Native Americans in New Mexico vote overwhelmingly for Democrats. Most on-reservation precincts cast c. 90% of their votes for Kerry in the 2004 presidential election. And they had the highest rates of undervotes.
The only precincts analyzed were those with 75% or more of voters of one ethnicity (including the precincts termed Anglo), based on the 2000 US Census. So ones which were more mixed were left out. For those interested in number crunching themselves, an Excel spreadsheet of the data has been published, too.
Considering Bush's winning margin was less than 6,000 votes in New Mexico, and that the undervotes were concentrated in heavily Democratic precincts, that probably accounts for the margin of victory right there.
The 2004 Presidential Election
All of this summarizes what I already knew. Because I worked Precinct 13 of Taos County in detail for the 2004 election. Precinct 13 is the Taos Pueblo Indian Reservation. Only a small percentage of the voters are non-Indian. Greg Palast has written about this precinct - he based what he wrote upon data I provided.
I worked Precinct 13 in detail during the 2004 Presidential election. Cooperating with the local Democratic party, and other "irregular" grassroots activists, I made a detailed spreadsheet of the precinct. We did a mailout, and additional work to get phone numbers and addresses current.
An aside: Addresses are a problem in this part of the country. I once lived on a road that had five different names. You had to keep track of which one to use, depending on whether you were dealing with the phone company, the electric company, cable TV, UPS, FedEx, etc. We don't get mail delivery to our homes, so it comes to PO Boxes (or area cluster boxes, which the USPS calls NDCBU - or something like that - I can never remember.) This kind of problem is even more pronounced on-reservation.
I was the Party GOTV poll worker on Election Day. We knew our spreadsheet of the precinct was the best data available. We checked off every voter who'd voted early, and then each voter who came into vote through out the day. At the end of the day 1 out of 7 ballots cast on the "black box" on Election Day showed no vote for President. However, for the early voters, who voted on optical scan paper ballots, 100% of the ballots showed votes for President. I raised an alarm about it at the time, and others took it up, and Greg Palast wrote about it in his book, and in this article (Truth Out archives).
Precinct 13 wasn't unique. On Navajo lands, indecision struck on an epidemic scale. They walked in, they didn't vote. In nine precincts in McKinley County, New Mexico, which is 74.7 percent Navajo, fewer than one in ten voters picked a president. Those who voted on paper ballots early or absentee knew who they wanted (Kerry, overwhelmingly), but the machine-counted vote said Indians simply couldn't make up their minds or just plain didn't care. On average, across the state, the machine printouts say that 7.3 percent - one in twelve voters - in majority Native precincts didn't vote for president. That's three times the percentage of white voters who appeared to abstain. In pueblo after pueblo, on reservation after reservation throughout the United States, the story was the same.
...
So we dropped in on Taos, Precinct 13. The "old" pueblo is old indeed- built 500 to 1,000 years ago. In these adobe dwellings stacked like mud condos, no electricity is allowed nor running water - nor Republicans as far as records show. Richard Archuleta, a massive man with long, gray pigtails and hands as big as flank steaks, is the head of tourism for the pueblo. Richard wasn't buying the indecision theory of the Native non-count. Indians were worried about their Bureau of Indian Affairs grants, their gaming licenses, and working conditions at their other big employer: the U.S. military.
And the story's not much different in Hispanic precincts:
Our team drove an hour across the high desert from the Taos Reservation to Española in Rio Arriba County. According to the official tallies, entire precincts of Mexican-Americans registered few or zero votes for president in the last two elections. Española is where the Los Alamos workers live, not the Ph.D.s in the white lab coats, but the women who clean the hallways and the men who bury the toxins. This was not Bush country, and the people we met with, including the leaders of the get-out-the-vote operations, knew of no Hispanics who insisted on waiting at the polling station to cast their vote for "nobody for President." The huge majority of Mexican- Americans, especially in New Mexico, and a crushing majority of Natives (over 90 percent), vote Democratic.
What if those voters weren't indecisive; what if they punched in a choice and it didn't record? Let's do the arithmetic. As minority voters cast 89 percent of the state's 21,084 blank ballots, that's 18,765 missing minority votes. Given the preferences of other voters in those pueblos and barrios, those 18,765 voters of color should have swamped Bush's 5,988 vote "majority" with Kerry votes. But that would have required those votes be counted.
BTW, for those of you who are fans of Bill Richardson for President, it's worth noting that he was on the wrong side of the New Mexico recount issue. (You'll have to read Palast's article about that, because I've come to my fair use limit in quoting Palast's story.) Nobody can explain why our Secretary of State, an Hispanic Democrat, was utterly uninterested in this. Except perhaps that, as President of the National Association of Secretaries of State, she was seduced by the corporate boys from the electronic voting machine companies. That's only speculation, but very tempting speculation.
When the 2006 results are compared and contrasted with 2004, it turns out Native Americans have the lowest undervote rates when there's a paper trail, and the highest when there isn't. I can't explain why that is, but it's good news that our state has taken an important step forward. And the numbers geeks have shown that it's made a difference for the better.