[EM] untraceable receipts

Rob Speer rspeer at MIT.EDU
Fri Nov 14 21:13:01 PST 2003


On Fri, Nov 14, 2003 at 01:29:55PM -0800, Ernest Prabhakar wrote:
> >option of paying you $20 for your verifiable vote receipt. Or
> >threatening you if you don't get a receipt and give it to them.
> 
> Actually, there's a trivial way to avoid the problem of 'coerced 
> ballots', even with plaintext receipts.  All you need is "easily forged 
> receipts."   You could even have a "sample machine" set up in the same 
> room, where users could 'test' their voting, and get a fake (but 
> indistinguishable) receipt.  The electronic display (and booth signage) 
> should clearly indicate that this is a sample, but the receipt should 
> be indistinguishable from a legitimate one.    A coercer would have no 
> way of telling the difference.

If the machine only prints out a receipt, then it's not much of a
verification system. It verifies that you pressed the right button, but
not that the machine recorded your vote, or that the votes from your
machine were tallied correctly. In this limited case, "fake receipts" do
work.

But if the receipt has some sort of verification number which you can
check against a database of votes later, to ensure that your vote was
actually counted, fake receipts defeat the whole system because you
cannot tell if you were given a fake verification number to start with.
(Suppose the machine fails to record votes from 1 in every 100 voters
who are registered with Party X, and gives them fake "verification"
receipts!)

-- 
Rob Speer




More information about the Election-Methods mailing list