Background image

How we keep data private

MSD has improved the way we work with client information to ensure client confidentiality, while releasing the maximum amount of valuable data to the public.

To ensure an individual’s information is safeguarded the Ministry firstly anonymises (removes an individual’s identifying information), before grouping individual client information to make it difficult to identify a real person.

MSD then applies a range of confidentiality procedures to protect client information.

The Ministry now applies a new, cascading set of rules that test how sensitive a data table when considering its release. This test may result in a data table being withheld or the application of confidentialisation.

We’ve moved from only using manual suppression of every value less than six, to also use an approach called random rounding, which is used and endorsed by Stats NZ.

Random rounding can be automated, so removes the risk of manual error in suppression and allows significantly more data to be published, increasing the value to the public.

Two possible additional layers of protection are suppression and random rounding.

Suppression is only used for tables regarded as highly sensitive. Primary suppression removes all cells with numbers below 6. Further suppression may be required to avoid people successfully calculating the suppressed amount from other data.

Random rounding is also be applied to a suppressed table to further protect confidentiality. Random rounding to base 3 is applied to all publicly released tables ensuring there is little or no chance a table could be misused.

This may affect the data people receive and when seeking to interpret it people should note:

  • Rounding to base 3 is applied to all cells in a table including totals but no cell or total will ever be more than 2 away from the original value. A cell value of 1 may be rounded to 0 or 3.
  • Columns may not add to provided totals but the provided totals will never be more than 2 away from the original number, given totals are calculated before rounding is applied.
  • Applying random rounding to the same value across different data requests may result in numbers that are marginally different. The difference will never be more than 2 from the original figure. For large numbers the impact will be negligible but the impact will be more apparent for smaller numbers.

For those of you used to working with very technical information provided by the Ministry, a more technical explanation of the approach can be found here StatsNZ - Applying Confidentiality Rules.

Background image
Print this page.