[EM] Legal Analysis of HR811 & Brennan Center Think-Tank Analysis of HR811

[Abd ul-Rahman Lomax]

I reviewed, on the Range Voting list, a blog that criticized HR 811. 
On that blog, it was claimed that the requirement of HR 811 that 
voters must be, on request, provided with a paper ballot to fill out, 
would not take effect until 2010. I could find nothing about this in 
the text of the bill, the latest I could find, as reported out of 
committee, which would otherwise be effective for November 2008 elections.

Does anyone have more information on that point?

HR 811 appears to be very good legislation. It does not resolve all 
potential problems with DREs, for sure, but for the first time makes 
it possible to independently verify the vote with paper records.

Once there are paper records, public ballot imaging becomes possible. 
I have so far only seen state law in one state on the subject of 
ballots being photographable public records, in Florida, where law 
specifically permits it. So in Florida, at least, it becomes 
possible, in 2008 if not before, for the public to fully count the 
election, thus bypassing all dependence upon DREs for vote counting; 
the DREs then become only a convenience for election officials, not a 
crucial function because counting errors *would* be discovered. All 
of them, not just those from the statistical checks.

However, there does remain the problem of DREs printing ballots which 
are altered from voter intent and expression on the machine. It is 
argued that votes will not carefully examine the ballots, and that 
therefore the precedence should be reversed: ballots should be 
hand-marked by the voter and then the data entered from them. 
Certainly some opportunity for influencing the vote corruptly 
remains, then, with the use of DREs to routinely prepare the ballots 
for the voter. Voters, however, can interdict this, and voter 
volunteers could actually extensively check the machines.

For example, a voter could go in and vote many times, viewing the 
receipt, and then cancelling the vote as incorrect in some way, thus 
presenting herself with many opportunities to view transcription 
errors in a context where she was being careful about how to vote. 
Likewise, it is not necessary that every voter verify the printed 
ballot in order to be able to detect systematic failure to properly 
record votes.

What has come to me as the ultimate measure to be taken with this is 
to emphasize to voters that the paper ballot *is* the vote, that the 
machine is only a device for preparing it. While it is true that the 
machine's copy of the vote may be used for counting, it is also true 
that it might not. In the end, it is the paper ballot that is the 
legal vote and which will prevail if, for example, the machine vote 
is lost or destroyed, or differs from the paper ballot. I think that 
studies that showed that voters weren't carefully examining the paper 
receipts was that the conception was that they were merely receipts, 
something extra and not important. HR 811 makes it clear that this is 
not the case.

The paper ballot is the vote, whether it is prepared by a DRE, or is 
a paper ballot filled out by the voter. And I'd suggest that voters 
who don't have disabilities making it difficult be encouraged to 
request and use paper ballots. If enough do this, the alleged 
economic advantage of counting with the DREs would evaporate, and 
what are really more efficient and cheaper and more difficult to 
corrupt systems which scan paper ballots would replace them.

Indeed, I've suggested that equipment for automatically counting 
ballots is lying about unused, or readily available for use, in 
abundance, it need not be purchased. All it takes is a fax machine 
and a computer, any standard computer of modest capability, the kind 
that you presently have to pay someone to take away for recycling. If 
ballots are imaged by multiple independent observers, whether by fax 
or by using digital cameras, the counting process becomes much less 
critical, since different processes can be used by different 
analysts, and corrupting the images is likewise impossible, since 
they reside in totally independent locations and under the 
independent control of the observers.

This would free election officials to use whatever process is most 
efficient, since the stringent security requirements of, for example, 
manually counting physical ballots are eliminated. Ballot images can 
be copied and counted by independent teams, with each count 
proceeding far more rapidly due to the free handling that becomes 
possible. However, I'd expect software to be developed that would 
take images, recognize the votes on them, and automatically count 
them, presenting ambiguous ballots to an operator for resolution. 
Observers could observe this, but it is far more efficient to allow, 
indeed, each observer to do their own counting! You can do that with 
ballot images, not with the physical ballots.

Ballot imaging should make it possible to immediately sequester the 
physical ballots. The mandatory audits under HR 811 would, of course, 
be done with the physical ballots, by hand, thus providing another 
independent validation of the vote, albeit only statistically because 
of the reduced sample size.