Photo: RNZ / Richard Tindiller
Some police databases date from the 1970s, are outdated, siloed and vulnerable to security risk in some cases, a stocktake has found.
The Data Stocktake and Gaps Analysis was released this week as part of the wider Police-commissioned research into bias within the ranks, the Understanding Police Delivery project.
Waikato University statistician Paul Brown, who led the stocktake, said police data staff were often working in "siloes".
"A lot of participants said data was hard to find and access, and then when you could find it, it was hard to link the data, so it made it hard for effective and efficient re-use."
Gaps found:
- Data is not collected well - including on ethnicity.
- Some data not collected at all - "For example, in proceedings data, there is little to no data collected if an offender has any disabilities or mental health impairments. During roadside stops, demographic data is not usually captured."
- Data collected but flawed, incomplete.
- Not findable, not accessible - time consuming to find, difficult to use.
"Police collect a wide range of information, but it is not well documented as to why we want it, what we intend to do with it, and whether it is collected at the appropriate time or what the best way to collect it is," the report concluded.
About 40 percent of victims' information could not be collected at the time of an offence, mainly due to the complexity of the situation.
Dr Brown said that was "a huge challenge" for all involved.
Some staff saw demographic data as "nice-to-have" for policy and deeper analysis, but a safe and timely response was more critical.
"And there could be seconds between life and death in these situations, so we don't want to burden staff by making them collect high quality data - I mean, we want all that. "But there is a balance to be struck."
Databases currently in use 'old' and do not easily connect with each other
The report found some databases had been in place since the 1970s and had been built upon over the years, "creating a siloed infrastructure, making data hard to find, access, not interoperable, and therefore not optimised for both operations and research uses".
Some older software applications carried "technical and security risks", and some systems relied on the institutional knowledge of specific staff members. who could leave or retire at any time.
Dr Brown said good data was important for building fairness and equity into police service delivery.
To improve police data systems, recommendations include:
- Development of a Police Data Catalogue to record what data police currently have and where it can be found.
- A formal set of guidelines and strategy for the appropriate governance and use of Māori data within the Police data ecosystem.
- A larger "data workforce" for efficient and accurate data capture, data entry, and data quality measures.
- Building data systems to ensure data is findable, accessible, and interoperable.