ABS boss on data integration: ‘Why aren’t we talking about this in the community?’

By Stephen Easton

Monday November 6, 2017

A screen-grab of Australian Statistician David Kalisch during a press conference in Canberra, Wednesday, Aug. 10, 2016. The Australian Bureau of Statistics shut down the Census website after a series of denial-of-service attacks on Tuesday night, with the strongest one coming just after 7.30pm (AEST). The federal government says the Census website was not attacked or hacked and no data was lost. David Kalisch insisted the ABS adopted a precautionary and conservative approach. (AAP Image/Adrian Muscat) NO ARCHIVING

The head of the Australian Bureau of Statistics says researchers and others who want to make the most of publicly owned data need to do more to build public trust and explain the value of their work in the wider community, not just in their own scientific and tech-savvy circles.

At the end of a seminar on data integration in the public sector on Friday, Australian Statistician David Kalisch said he was surprised there wasn’t more discussion of “the role of the citizen, and how we need to effectively engage with the public to ensure there is strong community understanding and support for how we’re proposing to use their information”.

“It’s as if we’re talking about it to ourselves within a closed environment here.”

He noted “there is some clear evidence of reduced trust in governments” as did several speakers at the conference, which was hosted jointly by the ABS and the Institute of Public Administration Australia.

While a lot of academics, public servants, nonprofits and privately funded think tanks are sold on the value of linking up various administrative databases and using them as valuable resources for research, Kalish said it was important for them to remember not everyone shares their enthusiasm and some of the public had grown “a bit wary” of what could go wrong.

The ABS chief put that down to increasing fear of data breaches like the one that made headlines last week. Regardless of the exact circumstances, each report of sensitive personal information being exposed contributes to the general sense of unease that is part and parcel of the internet age.

Earlier in the day, social science professor Janeen Baxter said she wasn’t convinced there was much concern in the community about privacy risks from public sector data integration and research using government databases, beyond a few especially “vocal” groups.

The discussions about “unlocking the value of Australia’s public data” were pervaded with an optimistic view that most people could see the wisdom in breaking down information silos in government and making data that is already sitting there more available to academics.

Kalisch, however, pointed out that a group of people who would go to a conference about public sector data integration was not a representative sample of the community.

He observed that the room was “skewed on one end of the spectrum” — they had all come with “an appreciation and understanding of the importance and utility of data to inform key decisions” and generally were comfortable that any risks to the public could be managed fairly easily. Perhaps they need to get out more.

“I think there’s a clear appreciation of the benefits to the community from more effective use of data [within the public sector] – so we’ve talked about more effective government policy, more effective government programs, benefiting the community, benefiting the taxpayer,” Kalisch said.

“But why aren’t we talking about this in the community?

“It’s as if we’re talking about it to ourselves within a closed environment here. Let’s engage with the community on these issues.”

These are fairly new topics of public concern. Increasingly people have worries about information security, privacy and the storage of personal information by companies, government agencies and other organisations in computer databases, even if they don’t know a lot of technical details about these matters.

Kalisch thinks “the techniques and technology” have gotten ahead of “the conversation that we aren’t having with the community” and sees a need to bridge the gap between “the culture of data custodians” and the wider community.

And that wasn’t all he had to say about building public trust. When the ABS was under fire from critics last year, Kalisch would have liked more support from the many other researchers and organisations in “the data ecosystem” that benefit from the information the agency provides. He was disappointed at how few backed up his attempts to reassure the public that the Census was safe and the changes would deliver even more valuable information.

Earlier, the Productivity Commission’s principal research adviser, Jenny Gordon, had shown a lot of sympathy for the ABS. Speaking on a panel, she said the privacy concerns that led to the Census boycott movement – after former Australian Statistician Bill McLennan sharply criticised the decision to retain names and addresses for a longer period of time in a major newspaper – were simply a protracted media beat-up.

In Gordon’s view, the resulting brouhaha was a storm in a teacup and privacy concerns were misplaced. She argued that news outlets prefer a negative, adversarial story, and so they stoked hysteria while refusing to give equal coverage to Census supporters.

When Kalisch wrapped up later, he was looking on the bright side. He said the ABS had collected “great data” and there were really only “a couple of days that caused a bit of grief” in about eight weeks of work.

“But I suppose one of the key lessons for me about that Census experience was the ABS actually went a little bit out on a limb there, in terms of looking at more effective use of that Census data and what we struck was a number of people that were concerned about that,” he added.

“But what did strike me most, actually, was the comparative silence of those who wanted to use that data.”

He said “the comparative silence of other key users of data was just stark” aside from a few researchers like Australian National University professor Nicholas Biddle, who spoke at the seminar.

“Now if you want data, if you want to have effective use of data, you’ve got a role, guys.

“You’ve got a role to play in terms of having that engagement, and being part of that conversation with the community.”

Kalisch made more detailed comments about “social license and public trust” in a recent speech.

The value proposition

Kalisch said it was important to remember that “public data is a public resource” paid for by taxpayers, which meant public servants and researchers had a dual responsibility: to protect it and to make the most of it for the benefit of the community.

Another issue with maximising the value of data is the large gap between the theoretical benefits and the actual impact the public sees in terms of better evidence-based public policy. We also often hear that rigorous policy evaluation can lead to better outcomes, but there is always strong political pressure against honest appraisals. Governments are increasingly bringing this to bear even on independent offices like auditors-general.

 We need to think about mutual benefits with a broader purpose for the public here, rather than someone trying to expand their own empire or expand their own influence.”

Too often policymakers misrepresent, underplay and ignore what the data is saying, or they simply avoid having much of a look in the first place. When it was Grattan Institute CEO John Daley’s turn to speak, he suggested one answer was more sharing with a wider range of trusted research organisations – the middle ground between locking up data in silos and full open access – to help inform public debate with more free flowing information.

Compared to other nations, Kalisch said, Australia’s governments had not opened up valuable public data to the extent that they could and should. He said the recently announced Data Integration Partnership for Australia would help: “I think this is a major initiative that we need to recognise as a significant step forward in terms of building some critical data infrastructure and building some essential collaborations in Australia.”

He wants to see data from all levels of government linked up with what is available from the private sector and academia a lot more in future. Collaborations to work with public data – like those between the ABS and every organisation represented at the conference – should always be based around the “higher purpose” of the public interest, he added.

“Collaborations that are premised on organisational [interests] first won’t succeed. We need to think about mutual benefits with a broader purpose for the public here, rather than someone trying to expand their own empire or expand their own influence.”

He also agreed with another point made by Daley: that “the pursuit of perfection” in terms of data quality and linkage can be unhelpful. Policy choices have to be made, after all, so an indication of where to go next in a particular area that is based on reasonably compelling evidence is a lot better than guesswork, even if there are some gaps.

“Sometimes things are good enough, and we’ve got heaps of data out there that is good enough and it can provide a good story, provide a good perspective,” said Kalisch. He urged his guests not to “get caught up in the technology and the techniques” and think about the overall aims.

At the end of the day, each area of public policy involves specific questions about what the government should do, or stop doing as the case may be. These are questions that will be answered one way or another, Daley pointed out, so why not empower as many organisations as possible — even his arch-rivals at the Institute of Public Affairs — to throw in suggestions based on publicly owned information.

Obviously there is great value in creating larger sets of data to reveal new insights into the lives of lots of people over long periods of time, and it makes sense to repurpose existing public information that is already available. This maximises the value of an existing asset and is much cheaper and easier than running new scientific surveys, which also mean further intrusion on participants. But the benefits are not always as obvious as the experts think, and the growing unease about privacy in the age of the internet can’t be dismissed as fear-mongering and hysteria.

It seems that after his Census experience, Kalisch recognises citizens are not all persuaded that data integration is safe and beneficial, and accepts that their concerns need to taken seriously by those who are already sold on the idea.

He told the conference that he still has hope of going “beyond just the analysis of data, to actually building a stronger evaluation and evidence-based culture, across the community and in public policy more generally”.

About the author
Inline Feedbacks
View all comments

The essential resource for effective
public sector professionals