The 36,000 votes that didn't count in the same-sex survey
TENS of thousands of same-sex marriage postal votes were discounted because voters didn't fill them out correctly, according to the Australian Bureau of Statistics.
While the resounding Yes result wouldn't have been different even if every invalid vote counted as a No, it does beg the question: How did so many Australians fail the seemingly simple task of ticking a Yes or No box?
ABS chief statistician David Kalisch announced the results of 12.7 million votes in a speech broadcast live around Australia on Wednesday morning.
"For the national result, Yes responses, 7,817,247 ... That's 61.6 per cent of clear responses were Yes," Mr Kalisch said.
"No responses, 4,873,987 ... That's 38.4 per cent of clear responses were No.
"A further 36,686 responses were unclear representing 0.2 per cent of total responses."
That's a lot of people who went to the trouble to post their votes only to have them not count.
"Responses were categorised as not clear if there wasn't a clear Yes or No intention of the participant," an ABS spokesperson told news.com.au.
"This included forms where both boxes were ticked, where there were no marks on the form (blank response) or where the question was changed to make the answer unclear."
The ABS did not respond to questions about whether or not penis drawings played a part.
In the lead-up to the $122 million voluntary postal survey, there was a campaign to put glitter in envelopes as a protest. But the ABS advised voters not to use the sparkly decorations because they can interfere with the running of the machine that processes all of the ballots.
The ABS encouraged voters to only use dark-coloured pens to fill out the survey and not to draw on the ballot - or risk it being invalidated during processing.
The ABS also warned that if the Yes or No answer wasn't immediately clear the ballot would be deemed invalid. Any forms with tampered barcodes would suffer the same fate.
Last month, the ABS confirmed razor blades had been found in some envelopes, along with dirt and glitter in others.
In addition to this, 14 matters were referred to federal police for investigation. None of those instances involved objects in envelopes.
ABS deputy statistician Jonathan Palmer told a Senate committee hearing in Canberra the 14 complaints were related to the attempted sale of survey forms online, while another four were referred to state and territory police and involved survey form theft.
The razor blades and other objects were found by ABS workers responsible for scanning the near 12 million forms that had been returned.
An ABS spokesperson today told new.com.au that 95 per cent of paper responses were automatically coded in a machine. The other five per cent were manually tallied by operators. The department would not allow an operator to comment on the matter, citing "privacy" reasons.
"Of the survey forms manually coded, 4.3 per cent coded to a Yes or No and 0.7 per cent coded to blank or response not clear," the spokesperson said.
"Optical mark recognition software was used to capture the single-use, anonymous barcode on each form and the response.
"Where the marks on the form were simple, for example where one response box only was clearly marked, the form was automatically coded to the appropriate response."
According to the ABS spokesperson, where the marks on the form were more complex - for example, both response boxes were marked or there were other marks on the form - it was manually coded by an operator who then determined the appropriate response.
"If the single-use, anonymous barcode on the survey form was unable to be read (damaged or defaced) the form was not counted," the spokesperson said.