All eyes on 2006 - How to spot fraud hotspots

[iconoclastNYC]

I'm curious if the community here has learned enough about the possible modalities of fraud to spot suspicious results in 2006, and if so, ways in which the community can prepare for the 2006 midterms.

I'm betting that based on what is known that some sort of application could be built to show in real time as data is coming in on election where fraud might have occcured.

Data that would be included: pre-election polling data, voter registration data, past election returns, exit polling data, and vote tallies.

It would be a big effort but it would be amazing to have it in place and to be able to have this tool in place on election day so we could visually see on a computerized map where we think fraud took place.

What do you think?

[adolfo]

1. Exit Polls

We should conduct our own damn independent exit polls.

The Repugs will never release the raw data because that is how we can determine precisely where they cheated us in 2004.

[Chi]

3. We'd probably only have to do one, maybe two states.

Heya Adolfo
Could focus on Ohio, ask the exit poll experts how many precincts, and how many samples we would need to get for accuracy sake.
Maybe even make a council of people to design it.
Then figure how many volunteers we would need.

But then again, no one is gonna believe that a poll sponsored by DU members would be impartial and fair.
If it would be only used to hone in on fraud for investigation, I guess it wouldn't really matter.

[skids]

2. uscountvotes is supposed to do that...

...but they've been engaged mainly in lobbying and their project to collect elections data has been limping along very slowly.

If they aren't up to speed by the primaries, I'll see what I can pull together to archive the results as they come in. See link in my sig if you are a programmer -- help processing the results will be needed.

(EDIT: In addition, one thing I'd love to see done is the "paranoid's guide to voting" that tells voters what they can do to prevent themselves from being knocked off the rolls, prevent their ballots from being spoiled, and collect evidence during the voting process. Plus an advanced guide for pollworkers.)

[adolfo]

4. We are working on Florida Data

2004 General Election Precinct Results

http://recountflorida.com/ufed

[skids]

5. Cool. Good work.

I'll link you up when the uscvprogs webpage gets updated. Probably won't be until I gear the project up pre-primaries.

Do you guys have a list of URLs built where the "live" results will be coming in?

(I also have some primary and pre-2004 data from FL if you have any missing counties in your archive for that, because I used some FL canvasses to test the software.)

[Bill Bored]

6. One thing that needs to be done BEFORE the election is

100% bi-partisan auditing of ALL ballot definition programming.

This means making sure the candidates' names on the ballot coincide with their database records where the votes are counted, making sure that no straight party options are set up incorrectly or illegally to exclude or include any candidates in the straight tickets that shouldn't be, etc.

There needs to be REAL logic and accuracy tests of DREs and Op Scans too. This is much easier to do with Op Scans because you can use a large number of hand marked paper ballots (or a small number run through the scanners many times) and it doesn't take very long.

DREs on the other hand basically test themselves because in order for a human to test one, he/she would have to stand in front of each machine for hours at a time casting as many votes as there are likely to be on Election Day, or on all those Early Voting Days! If this isn't done, the test is simply a canned subroutine that simulates voting with fictitious candidates, ballots and totals. Useless!

But if the ballot definitions can be audited, at least we will know that votes aren't being switched in real time as they are cast using the features of the software that allow this.

IMO, if we don't catch these sort of problems before they occur, there will be little or chance to fix them later, unless there are contests of election and recounts of voter-verified paper. Candidates need to be educated about this and they must demand manual recounts whenever there is a mathematical chance that an outcome is wrong. Of course, you can't do that without a paper trail or ballot, and it will cost money. but it may be the only way to ensure that results are correct (not necessarily the exact count, but the outcome of the election, which is what really matters).

Bottom line: we shouldn't vote for or give money to any candidates that are willing to trust these machines, especially in the Primaries!