Voting Machine Security Studies Show Continuing Problems

Anonymous commented on a previous post that raised questions about the security of Diebold voting machines. Diebold changed the name of its voting machine division to Premier Election Solutions in August. Anon pointed to the State of Alaska website with information on the State's ongoing study of Alaska's voting machine security. The project has been contracted to the University of Alaska Anchorage (the Department of Engineering seems to be doing the work.) A key aspect of this study is the review of studies already done by other states. The whole Phase Ia report was contracted to cover:

Overview-level evaluation of recent studies relative to existing Alaska systems,
technologies and procedures
Inclusions:
• A brief review of studies and tests that have been undertaken that might be relevant to Alaska’s situation.
• A summary of the University of California’s and Florida State University’s tests and conclusions, analyzing the recommendations that were made and are applicable to Alaska’s optical scanning technology.
• Research and assess improvements made by Premier (formerly Diebold) based on
California and Florida studies and their applicability to our systems
• Assessment of existing Alaska systems and equipment and ability to upgrade security functionality
• Research other states that are conducting similar research. Determine potential points of collaboration, partnership and leverage
• General evaluation of Alaska’s election policies, processes and procedures
• Provide repository for public input via Division of Elections website. Use this input to guide suggested approach for interactive public input/response in Phase 2. Ensure that both UAA and Division of Elections have record of public input. ....

From the Phase 1 Executive Summary we get synopsis of the overview of studies done by other states:

What did the California and Florida studies find?
As part of Phase I, we reviewed a number of election-security studies done in other states. But our reviews of the California and Florida studies were the most detailed—and those states use the same or similar electronic equipment as Alaska. Generally speaking, the studies identified a number of worrisome vulnerabilities, including:
• Vulnerability to the installation of malicious software that could allow incorrect recording or miscounts of votes.
•Susceptibility to computer viruses that could spread from voting machine to voting machine and to election management systems.
• Insufficient control of access to and management of machines, potentially making them accessible to unauthorized people.
The manufacturer of the equipment—Premier Election Solutions [Diebold]—made improvements in its software and machines, based on these studies. Follow-up studies by Florida investigators found that newer versions of Premier software and hardware corrected some but not all the flaws identified.
• Identify areas of risk in Alaska’s absentee and questioned ballot system.
• Assess vulnerability of paper ballots to tampering, and contrast with risks in electronic system.
• Determine points in the election system where there should be more redundancy in personnel or procedures.

I don't have time to get all the details at the moment. Above are the highlights from the summary. You can go to the full Phase 1a Report. The Overview of Studies begins on page 29 of the report (p. 36 on the PDF file). It looks at reports on

• Maryland
• Cuyohoga County, Ohio
• University of Connecticut Voting Technology Research Center Report
• State of California “Top-to-Bottom Review” (TTBR) Report (and Diebold's response)
• Florida Software Review and Security Analysis Summary

I would guess these are among the more objective looks at voting machine technology in use. Perhaps we can get more into this later. A giant tip of the hat to Anonymous for the link.

The Diebold Effect More Serious than Tweety or Bradley Effects

Phil mentions the Bradley effect and the Tweety effect in his comment on the earlier post on the New Hampshire primary.

Bradley effect argues that white folks will tell a pollster that they will vote for a black candidate (former LA Mayor Tom Bradley) but in the secrecy of the voting booth, they don't.

Tweety effect argues that people get so angry at smug, mysogynist pundits that they change their vote to prove them wrong.

But as I raised in the earlier post today, I want to be sure that the discussions and investigations of the New Hampshire primary also consider seriously the Diebold Effect. This, as the name implies, is the effect of people tampering with the Diebold voting machines to tilt the vote to their advantage.

[See more recent post on State of Alaska study on other state voting systems.]

Diebold, Clinton, Obama, Paul, New Hamshire, Fraud

I don't know what happened. Here's what I do know.

1. Voting machines, like those made by Diebold, have serious flaws that potentially could allow unauthorized people to reprogram them and to change the outcome.

2. Diebold has been a strong supporter of George W.

You'll notice that these links don't go to traditional media and they aren't that new. That raises other questions about why others aren't working on this now. Maybe Google pushes the traditional media back behind the blogs, but that doesn't hold up for other issues. But I trust my son's judgment on these things and he's strongly opposed to voting machines because they have too many problems.

3. Obama was leading by around 8-10% in the polls before yesterday's election in New Hampshire.

4. He was several percentage points behind Clinton after the election.

Polling of Democrats just before Tuesday's vote gave little warning of the New York senator's comeback, with most underestimating her strength. A USA Today-Gallup Poll gave Obama a 13-percentage point lead, putting her at just 28 percent. Another by CNN, local television station WMUR and the University of New Hampshire had Obama up 39 percent to 30 percent.

It wasn't just the pollsters. Journalists covering the candidates on their final full day of campaigning described larger, more energized crowds attending Obama's events than Clinton's. And from the Clinton camp came word of campaign shake-ups, as well as a moist-eyed candidate vowing to struggle on regardless — an appearance some analysts said helped humanize her and win supporters.

Exit polls conducted for The Associated Press and the television networks offered no obvious clues. Interviews with voters in the Democratic primary showed those who said they'd made their choice within the last three days — including those who said they'd decided on the final day — split about evenly between Clinton and Obama. (From Associated Press)

Reputable polls have a margin of error of 5% or less. This change is beyond that margin of error. Other candidates were predicted accurately. Though the Ron Paul camp is also grumbling apparently.

I've heard, on mainstream media, tended to blame the polls:

• It's hard to predict black candidates in biracial elections because people lie to pollsters about how they will vote.
• It was the Independents and women (Listen at least to 2:32)
  
• The pollsters need to reevaluate how they got it wrong
  

But it is only on the blogs that I'm hearing people raise the issue of voter fraud. I'm not saying it was voter fraud. I don't have omniscience. But, given what we know about the machines, and given the difference between the polls and the outcome, certainly one of the possible explanations people should be looking at is that someone tampered with the machines.

Now, it gives me some small comfort to think that the Clinton folks might have tampered with the machines, simply because it would mean when things get really dirty heading for the November election, that the Clinton camp might be the only Democratic opponent who could fight back against the Republicans in the dirty tricks department. Of course that is a pretty cynical perspective and it would be better to prevent the dirty tricks, but these, by their nature tend not be revealed until after the victor is comfortably in office. We should have laws that invalidate the election if it is proven the victor won by deceit and deception out of his campaign. Of course, that would lead to the other candidate trying to sabotage his own candidacy in the name of his opponent.


But another explanation would be that someone else messed with the machines. I still am not sure who the Republicans want to run against. Whatever they say is calculated. Truth is a strategic choice, not a moral choice. You can listen to Allen Raymond here talk about the book he wrote now that he's out of prison for tampering with the 2002 New Hampshire election. It's all so matter of fact. There's nothing wrong. He happened to get caught and paid his dues. It's not about morality, he says, it's only about winning.

Rove's advice to Obama on how to beat Clinton for instance. Is that because they hate her so much? Or because they are afraid of Clinton and want Obama to take her out? Or they want to give that impression because they really want to run against Clinton? Is it easier to defeat a black man than a white woman? Is this particular woman encumbered by enough negatives that they think it would be easier to defeat her? If they want her as the candidate, then they could have been behind tampering with the machines, if that happened.

I guess what irks me, is that when I google New Hampshire primary voter fraud all I get is blogs. Why do I have to go to a New Zealand site to get this story?

Given the huge discrepancy between the polling data and the vote, why isn't voter fraud one of the possibly explanations in the mainstream press? It isn't like they don't jump on other undocumented blog reports.

[More on this topic added here and here.]