At least a quarter of the total NSW council elections last year may have been impacted by the failure of the state’s digital voting system, according to a new independent report.
NSW’s iVote system crashed during the local council elections late last year, with an unknown number of people unable to access the program and lodge a vote. The system had capacity for 500,000 votes, the New South Wales Electoral Commission (NSWEC) said, but more than 671,000 votes were eventually cast via iVote.
In a report into the issue released two days before Christmas, the NSWEC said that six local government elections were potentially impacted by the tech glitch, but the remaining results can be relied on.
But researchers Dr Vanessa Teague and Dr Andrew Conway have released an alternative report on the issue, raising questions with the NSWEC’s methods to reach its conclusion, and finding that up to 39 of the 122 council election results could be questioned.
The researchers said that NSWEC’s own analysis was “not convincing”.
To determine the impact of the tech issues, the NSWEC simulated the missing votes by randomly resampling them from existing votes made via iVote, and repeated this 1000 times. If an alternative result does not appear in these tests, the result was accepted by the NSWEC.
Dr Teague and Dr Conway said that this method assumed that all iVotes are accurate, included only people who were able to register on iVote but not send a vote and that the missing votes would have been distributed the same as the existing ones.
“It’s not surprising that they infer that omitting them didn’t change much, but if that assumption is wrong then a lot more might have changed,” Dr Teague told InnovationAus.
“We simply measured what number of different votes could have changed the outcome, without making any assumptions about them.”
The researchers conducted their own analysis of the existing data and how many dropped or altered votes due to the iVote issues could have changed the election results.
“In 39 contests, the number of required additions is less than the number NSWEC acknowledges that they excluded. In many other contests, the number is only slightly more,” they said.
“The iVote results were sufficiently numerous, and in many cases sufficiently different from the paper-only returns, to alter election outcomes.”
Dr Teague said that these 39 councils could have been impacted at a minimum, but the entire election may have been effected.
These issues “put the foundations of democracy at risk”, the researchers said.
“The decision to retain the apparent outcome in all but six contests depends very strongly on their assumptions that the iVotes are accurate, and that the votes they are missing are the same as the votes they already have,” their report said.
“If those assumptions are not accepted, there is a possibility that many of the announced election outcomes do not accurately represent the choice of the people.”
Late last year the NSW government said it would be “easy” for electronic voting to be used in elections at all levels.
But Dr Teague and Mr Conway said iVote should be scrapped entirely.
“Use of iVote should be discontinued because it does not securely convey votes, and leaves the state without a rigorous way of assessing how much its problems affected the integrity of the election,” they said.
‘The same situation could easily recur if the LGE – or the 2023 state election – is simply rerun with the same unreliable, insecure and unverifiable technology.”
Do you know more? Contact James Riley via Email.