Help salvage the #CensusFail by taking yourself back to Tuesday

ONCE you eventually log onto that Census website, take yourself back to Tuesday ... but with internet access, is the latest advice.

Pretend it’s still Tuesday ... but with internet access, is the advice on getting the Census finally complete. Picture: Supplied

IF you’re having trouble remembering how you got to work, who was at your house or where you parked you car on Tuesday, it might pay to write it down if you’re serious about saving this Census.

Tuesday night’s #CensusFail debacle saw millions unable to complete the national survey online at the very time the nation should have been taking its five-yearly real-time statistical selfie, and has left the whole operation exposed to the potential inaccuracies and dodgy figures — not to mention an incensed Prime Minister.

But with reports the site is now apparently up and working, it’s important to pretend it’s still Tuesday night when you eventually log on.

That way, says demographer Mark McCrindle, there’s a greater chance the figures the ABS eventually collates will reflect a real-time picture of Australia in 2016, rather than one the way we remembered it.

Which could be important for future policy decisions on issues like public transport and social demographics, says McCrindle, who was among the millions who tried fruitlessly to fill in the questionnaire in real time on Tuesday.

Just as important, McCrindle says, is not filling your details twice.

The ABS was hoping for a 98 per cent completion rate, but there are fears of over-completion among the online confusion, he said.

The danger of that is slightly skewed results if the double-ups aren’t picked up.

While the answers to many of the Census questions posed won’t have changed since Tuesday — unless in the past few days you’ve had a religious conversion, changed jobs, moved house, McCrindle says there a few critical questions that require Australians to cast their mind back to where they were and what they did — aside from wrestling with a website that was down.

“They include: How you got to work on Tuesday, 9 August 2016?; who was present in the dwelling on Tuesday night?; were there any people away who usually live there and how many cars owned or used by residents of the dwelling were garaged or parked at or near it.

McCrindle reckons there’s a window of about a week to have the Census complete before data will be compromised through forgetfulness or apathy.

The message you might get logging onto your mobile this morning.

“It won’t affect some elements of the data, because the answers will remain the same,” he said.

“And the trickiest groups to get are those who are in non-traditional accommodation, like boarding houses, visiting hotels, in transit, sleeping rough, and many of those are captured by specific teams who go out to physically capture the data, so that should have been covered,” he said.

“But the people who were trying to get online — and that’s mainstream Australian households have to try again.

“Definitely there’s an irritation and pushback factor, and that’s why fines are there for non-completion, so I think people will come back on board.

“But double-counting is a potential problem. Did a paper form get filled in by a member of a household out of frustration, and now someone else has logged in and filed the same data? That’s going to need significant paper versus e-census crosschecks.

“Or you might have people coming back to Australia that weren’t in Australia on Census night who fill it out anyway. There will be an overcount because of that.”

McCrindle said the whole point of the Census is to capture a moment in time.

“If it happens on the one night at the one time, the data is done, it’s finished,” he said.

“When it spills over a week or two, that’s when the potential for multiple completes rises, and also when questions like how they got to work must be what they actually did, not what they usually do.

And it’s back, two days on. For now.

“The point is some people might drive to work most days, but maybe on August 9 they didn’t, so a snapshot in real time is the most robust way of collecting data.

“The danger is people will answer with what they usually do, when in some cases, what they actually did in real time is different.

“Wait another couple of weeks and there’s a risk it becomes a theoretical exercise rather than a real-time snapshot.

“At this point people still remember what they did last Tuesday, but wait another couple of weeks and the risk is it becomes more theoretical exercise rather than an actual recording of what they did.”

While it’s become a painfully drawn out operation, McCrindle suspects in the end, embarrassment and costs aside, the figures that come out of Census 2016 won’t be compromised data.

“In the end, I suspect no,” he said.

“This has been such a big issue that when people talk 2016 Census data they’ll put a mental asterisk beside it.

“But the feeling is, in the end not much will be compromised. If we get the completion rate they want, it won’t be a major problem.”

dt home focus future homes 30jan16 Mark McCrindle