« House Floor Vote Tuesday: Domestic Partner Rights and Responsibilities Act | Main | Minimum Wage Heads to NM House Floor »

Monday, February 26, 2007

(Updated) Shocker: Undervote Rate Plummets in Minority Precincts After New Mexico Changes to All Paper Ballots

UPDATE 2: There's a Daily Kos diary up by Land of Enchantment that discusses the voting analysis and the poster's experiences working a precinct in Taos County, NM for the 2004 election.

UPDATE 1: This story is now getting national attention on its own merits plus its implications for the case being made by many prominent election reform advocates against Rep. Rush Holt's (D-NJ) new Election Reform bill, HR 811. In addition, the analysis of election data uncovers another manner in which DREs might be manipulated to alter results according to the language of the voter. This begs the question of whether minority voters who used DREs in 2004 were the victims of machine manipulation that produced the high undervote counts. The data seems to raise the possibility anyway. See the front page story on The Brad Blog for more.

From VotersUnite.org:
February 26, 2007 - 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.

The report explains: "Undervotes represent ballots on which no vote was registered for a specific contest.  Undervote rates higher than 0.5% in the major contest on a ballot, especially in presidential elections, suggest that votes may not have been counted, either through a mistake of the voter or a mistake in tabulation."

The report shows that in predominantly Native American and predominantly Hispanic precincts, undervote rates were abnormally high (7.61% and 6.33% respectively) in the 2004 presidential race, when the votes were cast on DREs.

In 2006, after the state changed to all optically scanned paper ballots, the undervote rates for Governor in those same precincts plummeted by 85% in Native American areas and by 69% in predominantly Hispanic precincts.

In Anglo precincts, undervote rates of ballots cast on DREs were about the same level as the rates for paper ballots — 2.22% and 1.75% respectively.

“We were looking for any impact the change to paper ballots may have had on New Mexico’s historically high undervote rate. When we found the dramatic drop in Native American precincts, we were shocked,” said Theron Horton, Project Manager for Election Defense Alliance. “Something was going on with the DREs in those precincts in 2004.”

“When Warren Stewart, Policy Director of VoteTrustUSA, and I did the analysis of New Mexico’s 2004 vote data two years ago, we found high undervote rates in the minority precincts,” said Ellen Theisen, President of Vote-PAD, Inc. and former Executive Director of VotersUnite.Org. “But we didn't do a complete comparison of paper ballot undervotes to the DRE undervotes in that election. When I heard of Theron’s work, I performed the comparison, and found that it’s the paper ballots that made the difference in the minority precincts.”

Read the two-page report here:

Download the data here:


Ellen Theisen

Theron Horton

February 26, 2007 at 03:05 PM in Election Reform & Voting, Local Politics, Native Americans | Permalink


The voting is so corrupt in this state.
In my mind, Madrid won. The under vote in that election was alarming.

Posted by: qofdisks | Feb 26, 2007 3:08:10 PM

My question is why weren't the Governor and Secretary of State concerned about the undervotes that were uncovered in 2004? Despite activists reporting on this repeatedly, there seemed to be no interest on the part of the governor or election officials in New Mexico.

It seems pretty clear that Kerry probably won New Mexico but maybe some parties didn't want to know that for their own reasons, you know?

Posted by: I Vote | Feb 26, 2007 3:17:38 PM

Looks like many votes may have been stolen in Hispanic and tribal areas in 2004. I can't believe the national bill by Holt doesn't see that the touch screens are bad, bad news, no matter what.

Posted by: El Norte | Feb 26, 2007 3:57:47 PM

Did they distinguish between the different type of machines used for Early Voting and for day-of voting? Did the ratios of those two kinds of voting differ in Hispanic and Native American precincts? Is this about people trying to vote straight ticket and not knowing how?

Posted by: Michelle Meaders | Feb 26, 2007 5:09:27 PM

I read the BradBlog piece, which contained this paragraph:
"The analysts of the data in New Mexico have also brought to our attention yet another new, disturbing dimension to touch-screen voting systems: they allow for simple racial profiling in elections. Whenever a voter must tell the system, at the beginning of the voting process, the language in which they want to vote, they are also opening the door for a malicious programmer to alter the behavior of the voting system based on that particular language choice."

I've been voting in New Mexico for many years. All the voting machine ballots I saw were in both English and Spanish. I think most Spanish-surnamed voters in New Mexico read and write English better than Spanish, since that's what they had in school. So even with separate ballots, they would tend to vote in English. And considering how they were educated, how many Native American voters read their native language? Do they all even have a written form? So I think that argument is bogus here.

Posted by: Michelle Meaders | Feb 26, 2007 6:35:24 PM

This report ought to shut the Repukes up after all that whining they have been doing for the last few years about voter fraud, requiring ID etc, etc.

Posted by: VP | Feb 26, 2007 7:26:17 PM

I'm not surprised at all. That's what the original litigation (Lopategui v. Vigil-Jiron) over the 2004 election was about, after all. What we now know is that the paper ballots solved the undervote problem in Native American and Hispanic precincts *this time.* However, the problem could recur, which is why we need to beef up the random audit provisions in the current law.

I'd like to point out that if the 2006 CD1 congressional race was stolen, it was probably done in the absentee precinct, not the early voting or election day voting. The hand count procedures were funky, and I also suspect some other hanky-panky, like Republican Party workers illegally dropping off a dozen or more ballots at a time to the Bernalillo County Clerk's office. Thank goodness our new Clerk, Maggie Toulouse, actually wants to fix these problems!

Posted by: Paldon | Feb 26, 2007 7:29:24 PM

The paragraph cited above by Michelle pertains to touch screen voting machines, not ballots. The point was that if you have a system that requires a voter to identify the language he or she is using to vote at the beginning of the process, stealth computer code can be used to change only those votes. It is another flaw in the very flawed touch screen system that has many openings like this one where fraudulent code can be used.

I don't think the Brad Blog article was saying it necessarily happened here, but that it was possible given how DREs are set up. Think of what could happen in certain areas where people are not as bilingual as they are here. And even here there are many elders who are not.

Posted by: No More DRE's | Feb 27, 2007 8:37:29 AM

Post a comment