Uh. Where I went to school, that's a D+.

[Wilms]

First, I reject the notion that current technology doesn't include Hand Counts. Lever's could be included, if only for reference.

Secondly, what's so great about 68% confidence level in DRE's, and 61% for Precinct Optical Scan (that it's disturbingly less than DRE's, aside)?

Third is the slanted analysis offered.

Finally, what's this bit about subtract negative scores from positive ones? Can the ExitPoll Gurus help out?

I wonder how much InfoSentry made off of this?

NEWS RELEASE

Americans Rate Trust in Computerized Systems Highest Among Current Voting Technologies

04 February 2006

Americans have higher trust in the confidentiality and accuracy of computerized voting systems, commonly known as Direct Record Electronic (DRE) and “touchscreen” systems, than in other voting technologies being widely considered as states and counties rush to comply with the Help America Vote Act (HAVA) of 2002. National opinion survey results released today by InfoSENTRY Services, Inc. during a meeting of state election officials in Washington, DC indicated that since the passage of HAVA, computerized voting systems have recaptured a slight trust lead over scanned paper ballots being considered for use in many polling places across the USA.

Glenn Newkirk, InfoSENTRY’s President and director of InfoSENTRY’s national polling project, commented, “The trust scores for the four types of voting technology have remained generally constant over the three years in which we have conducted the surveys. While the Positive Trust Scores of DREs and polling place optical scan systems were tied statistically in our 2005 survey, DREs regained a statistically significant lead in 2005. This is a fascinating trend, given the full, head-on assault by voting critics and the media on computerized voting in many states and counties since the passage of HAVA. Trust in the so-called “touchscreen” systems’ confidentiality and accuracy is far higher than the critics of this technology would have us believe, and this level of trust is without any added features like paper records being printed by the computers. Most election officials certainly would prefer to see higher trust scores in all forms of voting. It will be up to them to see that the systems they implement, regardless of the technology, are worthy of the trust given to the systems. It will also be up to voting system vendors to improve the security and reliability of their hardware and software to meet the highest information system industry standards.”

snip

Approximately one-half the states have required that newly purchased computerized voting systems must be equipped with controversial printer technology to produce paper copies of ballots that voters can choose to inspect when they vote. Voting system critics insist that voters will have confidence in computerized voting systems only if they produce paper copies of ballots. However, the costly ballot printer add-ons have been used only in a very few elections since their introduction.

Newkirk continued his analysis by stating, “When you subtract the negative trust scores from the positive trust scores to calculate a “Net Trust Score” for the top two voting methods, Americans’ trend toward trusting the fully computerized DRE voting systems becomes clearer.”

http://www.infosentry.com/InfoSENTRY_NewsRelease_Voting-Tech-Trends-Attitudes_20060204.htm


Election Updates Blog has some comments about it.

http://electionupdates.caltech.edu/2006/02/infosentry-survey-responses-on-voting.html

[freeplessinseattle]

1. That;s a D+ where I went to school

70's is C range in WA. any way you slice it, it's unacceptable.

[Wilms]

2. You're right!

So I edited!

Perhaps my grades were worse than I thought. :blush:

[freeplessinseattle]

3. well, in bushworld,

an F would be a "heckuva job" so...

[Dogmudgeon]

6. Maybe not

70% was passing when I went to school, but I graduated in 1976.

Setting the bar at 60% makes more sense -- 10% points per letter grade. That is, if letter-grading makes sense.

--p!

[EST]

4. Either the people to be polled were picked ahead of time,

the questions were not stated clearly or the pollsters are lying.
I realize that half the people don't vote and of those that do, many have no idea whom or what they are voting for, but I cannot believe this supposed poll has anything to do with reality.

[lakeguy]

5. hmmm, ask a bunch of people a question, add and subratct some

numbers, and look....i got my answer! it's the NEW voting math.

[OnTheOtherHand]

7. seriously...

I for one would have a damned hard time interpreting people's ratings of voting technologies on a 1-to-5 "trust scale." There is no baseline for comparison.

The move to figure a "net" score that is positives minus negatives is about as reasonable as anything else one could do with these numbers (heh). If I am reading the press release right, it means that about 15% gave DREs a negative (1 or 2) rating in 2004, 20% in 2005, 16% in 2006. Something like 15-20% gave them a "3" in each year -- what that means is subject to interpretation. Who knows what people are trying to convey when they place themselves at 3 on a "trust scale"? Presumably some are conveying that they have mixed feelings, and others that they don't really have an opinion either way.

However, from what I've seen, I don't think people should kid themselves that folks are predisposed to dislike DREs. Survey data out of Georgia, and the anecdotal reports of Avi Rubin, indicate that DREs are pretty well liked when they are not obviously malfunctioning. That doesn't mean that people have embraced black boxes as a great idea -- they basically haven't thought much about it. (And this survey question completely evades the issue of paper trails.)

[Boredtodeath]

8. Hope you've looked carefully at those Georgia polls

One item undermines the poll itself:

Were you aware that these machines were instituted state-wide?
No 28%

Have you yet voted on the new machines?
No 25%

These people, clearly 1/4th of the people surveyed, have never even USED the machines they are "confident" in. Uh huh, very valid polling going on there........

http://www.cviog.uga.edu/peachpoll/2005-03-10data.php

[OnTheOtherHand]

9. umm, actually, I hope you have

Respondents were asked, "How confident are you that your vote was accurately counted?" That question was asked, of course, only of people who said they had voted -- over 90% of whom said they had voted on DREs either on Election Day or in advanced voting. (Also notice the splits among people who said they had college degrees or higher, who self-reported 90%+ turnout.)

If you want to see figures just for people who voted on DREs, I would not be surprised if the Carl Vinson Institute were willing to share them. I think it's a bit cheesy simply to insinuate that the polling isn't "valid." We can certainly talk about how to interpret the numbers, but I'll stand by my previous post.

Now, the survey results cited in the OP are pretty darn weak.

[Boredtodeath]

10. Yeah, right

Only ONE question was posed to those who didn't vote:

What is the main reason you did not vote in the November election?

Every other question in that survey was posed to all 800 participants...........look again.

So, you say people who have never used the machines should be responding to this question, right?

How comfortable are you with advanced technology?
Total Unweighted Count 800

and this......

Overall, does the new, electronic voting system in Georgia make for more accurate elections in this state, less accurate elections, or does it make no difference compared to what we had in the past?
Unweighted Count 800

[OnTheOtherHand]

13. would you care to read my post, then refer to qV6c? n/t

[Wilms]

11. But the article does mention "paper trails".

And it says they are controversial...as if DRE's aren't.

[OnTheOtherHand]

12. well, they aren't actually trying to get it right

I don't think the finding they report is exactly wrong, but they have cherry-picked one aspect of people's thinking and wedded it to marketing rhetoric, more or less. I think the data they report are not useless, but the analysis is, umm, non-authoritative.