[LINK] New proposal for e-voting - Turnbull

[Roger Clarke]

>On 10/09/2013 4:28 PM, Jan Whitaker quoted The Age:
>
>>  Mr Turnbull, who was easily elected to his Sydney seat of Wentworth
>>  on Saturday, said he thought there was also a large number of people
>>  who voted fraudulently, "in the sense that they go to the polling
>>  place and say they're someone else".
>>
>>  He said he thought many people who did so were voting for a friend or
>>  relative who was away or sick - and that this was based on anecdotal
>>  evidence he had received since first running for Parliament in 2004.
>>
>>  Impersonating another voter in a polling place is a serious offence
>>  and carries a jail term of 6 months.

At 16:47 +1000 10/9/13, Bernard Robertson-Dunn wrote:
>And how does e-voting fix this?
>Or is he conflating e-voting with a national ID system?

Possibly.

However the report continues:
>Mr Turnbull said that electronic voting could be done in a closed
network in the polling booth so that it could not be hacked from the 
internet.  ...
>He suggested that an electronic system could point out to voters if
they were about to cast an informal vote and give them the
opportunity to correct it.

So he could be proposing no change to the elector-authentication 
process, and only a change to the means whereby the elector's vote is 
recorded.

It would presumably cost quite a bit to kit a very large number of 
locations, once every 1-3 years, with kit that's 
certified-hardware-secure, and runs software that is effective, 
sufficiently flexible, and certified-software-secure (whatever that 
means) and certified-data-secure (which isn't easy to spec).

And maybe the hardware would have to have a screen 2 metres wide ...

It would be interesting to know how much lower the informal vote 
would be if a user-friendly application were in place.

A quick stab in the dark is below.  Sensitivity testing is always fun.

________________________________________________________________________


Blank   28.9    No effect?  Or would that be rejected??
Protest 16.9    No effect?
Just 1  27.8    Some effect?                            Say 20%
Ticks   11.8    Some effect, perhaps considerable?      Say 50%
Non-seq  9.2    Some effect, perhaps considerable?      Say 50%

TOTAL   94.6    (I wonder what the 5.4% do)

So maybe 17% would become formal, of 6% of 15 million = 153,000
Across 150 electorates, that's 1,000 votes per electorate.

If the votes are distributed much the same as the already-formal 
votes, then there's no change in the outcomes.  There might be some 
bias towards Labor, if error-making is correlated with lower-socio-ec 
demographics, and if people in that demographic still think Labor is 
what it's name suggests.

But even if the difference in distribution were considerable, very 
few seats would change hands as a result of the investment.

So don't look for payback from that quarter.


-- 
Roger Clarke                                 http://www.rogerclarke.com/

Xamax Consultancy Pty Ltd      78 Sidaway St, Chapman ACT 2611 AUSTRALIA
Tel: +61 2 6288 6916                        http://about.me/roger.clarke
mailto:Roger.Clarke at xamax.com.au                http://www.xamax.com.au/

Visiting Professor in the Faculty of Law            University of N.S.W.
Visiting Professor in Computer Science    Australian National University