Research project investigates people’s views of data linkage

A research project carried out by University of Auckland researchers and Statistics New Zealand investigated what people think about having their data linked up by government agencies.

In a presentation on the findings (2016), the researchers distinguish between:

• "shared" data - where data is shared within and across government agencies, often at an individual level for a service delivery outcome, and 
• "linked" data - where data from different government agencies is linked together in a database for a research or policy outcome.

When the Integrated Data Infrastructure (IDI) was raised, the research found:

• Most participants were unaware of the existence of the IDI
• Some can rationalise the benefit as long as sufficient safeguards are in place
• Some were alarmed: data quality and anonymity were key concerns

Participants wanted informed consent for their data to be linked, saying:

• Permission should be obtained for all data linkage
• Sufficient safeguards need to be in place and communicated
• The data linkage must be for a beneficial purpose (personal or public)

Participants who had concerns about data linkage identified a range of issues including: data quality/inaccuracy, identifiability as individuals, a lack of control over information, appropriate use and reuse of information, fear of surveillance, fear of discrimination and fear of disclosure.

Māori participants who had concerns expressed distrust of the research process, concern about how data could be used to further embed negative stereotypes, and concern about creating an environment for over-surveillance.

The researchers concluded:

• Many people are willing to consent - consent is seen as an important step
• Groups with negative experiences of government agencies are less likely to do so
• Data quality and accuracy matters
• Trust needs to be built - recording of information must be:
  - Relevant
  - Accurate
  - Neutral
• There needs to be safeguards across all stages of the research process
  - De-identification doesn’t remove all of the risks
  - Consider consent, storage, access, analysis, reporting
  - Clarity of purpose

Study participants included "at risk" groups including disabled persons, LGBTIQ+, mothers of young children, new migrants, young people, Māori, men and survivor advocates.

The research was funded by the Data Futures Partnership.

Related stories

Government continues work on integrated data and "social investment"

Concerns and public meeting on NGOs and provision of individual client level data