PUBLIC ATTITUDES
TO DATA INTEGRATION
Highlight findings from the report prepared for
Statistics New Zealand
Anna Davison , Jean Beetham , Jared Thomas, Abigail Harding,
Vivienne Ivory, and Chris Bowie
Opus Research, Opus International Consultants Ltd, Lower Hutt
OVERVIEW OF PRESENTATION
• Key results of interviews, workshops, online discussion
• Highlight findings
• Trust in Statistics NZ
• Recommendations
NARRATIVE INTERVIEW TOOL
The interview schedule was an effective tool:
• Ice breaker engaging & provided personal context
• Demonstrated differences between operational and
statistical use of integrated data
• Participants’ responses were thoughtful: showed
appreciation from different perspectives
• Scenarios generated interesting discussion around
acceptability thresholds, benefits, costs, and values
NARRATIVE INTERVIEWS - OUTCOMES
• Participants considered it extremely important that data
is shared with and used by the right people and for the
right purpose – to provide benefit, rather than harm.
“Most information has the potential for a
positive and negative outcome for different
people. It depends who makes the decisions and
who has the information”
NARRATIVE INTERVIEWS – APPROPRIATE USE
• They felt strongly that information should not be shared
with non-Government or private organisations without
their consent/permission.
• “You wouldn’t want a loan shark setting up in an area where there is a high
level of poverty, people won’t turn down a pot of gold, but it will be more
detrimental to them in the long run”
NARRATIVE INTERVIEWS – AGGREGATE DATA
• De-personalised, pooled data was seen as useful, as long
as it was used by the right people and was used to have a
positive impact for either a particular group of individuals
or wider society.
• “I would be happy for the information to be used if it is going to help other
parents… if it would benefit my child.”
ACTION RESEARCH WORKSHOPS – THE METHOD
The blocks made the DI process more understandable and personal
–
–
–
–
stacking blocks
holding blocks close
bold moves
lots of debate & discussion
Lots of questions about how data could be integrated
No clear purpose so….
participants co-created the why & what
and then judged acceptability based on that
ACTION RESEARCH WORKSHOP – DATA TYPES
• Type of data was very important
in determining acceptability
They already
have that!
Normal, Personal,
routine complex,
data
sensitive
data
Too much
information!
ACTION RESEARCH WORKSHOP – PERSONAL EXPERIENCES
Attitudes to data integration can be strongly attributed to
personal experiences with and trust in
government authorities and statistics
Negative
experiences
data integration
may be:
unnecessary
not useful
misused
Positive
experiences
data integration
may be used to
produce public
benefits and
increase fairness
ACTION RESEARCH WORKSHOP – PRIVACY & SECURITY
When workshop participants discussed integrated data being held in a single
database and linked to their personal or identifying information they
tended to become increasingly less comfortable with statistical data integration.
“The important thing is how is
this confidential information
treated? Is it strictly confidential
and anonymous or can it be
misused?”
“If everything here is in Statistics
NZ, if someone hacked it they
would know everything about me,
and that’s scary.
That would have to be so secure,
otherwise it’s too much
information they’ve got about
individuals, it’s very powerful.”
ACTION RESEARCH WORKSHOP – TRANSPARENCY & CONSENT
Some participants felt that informed consent should be required for data integration
“People need to know what’s
being shared and how it’s
being used, and why they
need that,
and that you can say no to
having your data integrated”
ONLINE CONSULTATION - PROCESS
ONLINE CONSULTATION – LESSONS
• Mixed views on short timeframes – “waffling on” vs “tight timeframes”
• Rigidity of the tool
“It was not as interactive as I’d expected”
• Development of rapport with other experts
“…Felt very hesitant to edit other people's comments very much. Easier to
put in a new comment.”
• Extreme scenarios
“There were two types of concerns that people raised: privacy concerns
and concerns about whether the idea was a good one.”
“It was impossible to overcome or address the concerns”
But, valuable insights were still found
ONLINE CONSULTATION – IMPROVING ACCEPTABILITY
• Who uses it?
• What for?
• How it is protected?
Fair
Useful
Accurate
CORE
VALUES
• Some experts felt that DI
should only occur if:
– The public are informed
– The public find it widely
acceptable
Valid measure
Representative
In public
interest
ONLINE CONSULTATION – LIMITATIONS AND BENEFITS
• Experts more aware of limitations of administrative data:
– Lack accuracy, or not be in a convenient form
– Only capture those who interact with agency
• Benefits
– Longitudinal tracking
– More current, regular, accurate, detailed, and diverse data
Better data
Better
research and
information
Better
decisions
and services
DISCUSSION OF FINDINGS
• Most New Zealanders appear to have a relatively positive perception
of data integration by public sector agencies
• Attitudes to data integration do not appear to be strongly associated
with particular types of social groups
• Acceptability of data integration appears to be largely influenced by
the individual’s own personal experiences
• There also appears to be a more general value-based concern around
appropriate use
TRUST IN STATISTICS NEW ZEALAND AS A DATA CUSTODIAN
• Most of the research participants had a very high level of trust in Statistics
New Zealand as a professional, competent, and trustworthy custodian of data.
“I don’t have a problem bearing in mind it’s going to statistics
and it’s not shared, and no one else gets access to it.”
“Statistics NZ need to change their name, and it may change their attitude, and
portray a better, a more engaging type of response. Stats to me are scary, but if
they could encompass wellbeing, I’m sure people would be more responsive.”
RECOMMENDATIONS
• Statistics NZ could consider providing positive messaging
around data integration to ensure they retain and gain the
trust of the New Zealand public
– Reassure of security, confidentiality, & privacy
– Demonstrate need, value, & public benefits, particularly around
statistical data
– Consider offering the option to opt in/out of DI where personal,
sensitive, or complex data is being used
DATA INTEGRATION ACCEPTABILITY THRESHOLDS
Less acceptable
More acceptable
Ambiguous & general
Purpose
Transparent & specific need
Insecure
Security
Secure
Open access
Access
Strict restrictions, procedures & protocols
Public/ personalised
Privacy
Confidential/
anonymous
Private
Subjective & biased
Unfair & harmful
Interests
Data & metrics
Outcomes
Public
Representative & meaningful
Fair & beneficial
THANK YOU
We would like to thank the research participants
who generously gave their time and energy
to take part in this study.
Contact: Jared Thomas | Behavioural Research Manager
Opus Research, Opus International Consultants Ltd
jared.thomas@opus.co.nz
027 244 8574 | 04 587 0675