Comparing NZCASS with Police statistics

The New Zealand Crime and Safety Survey (NZCASS) and official Police statistics are both useful measures of crime. It is important to note that each set of statistics measure slightly different things and have their own strengths and limitations.

The following information will help users understand and interpret the statistics produced as part of NZCASS reporting.

Introduction to Police statistics

Traditionally Police have counted offences (“Historic Police offence statistics”) however this series of statistics ended on the 1st of April 2015. The historic Police offence statistics have now been replaced by the police ‘Recorded crime victim statistics’ (RCVS).

Comparisons between NZCASS and Police statistics have been made based on the historic Police offence statistics (rather than RCVS) for two main reasons:

  1. the recall period for the 2014 New Zealand Crime and Safety Survey was 1st of January 2013 to the date of interview (February – June 2014). The RCVS series only started from July 2014, meaning the time periods were incomparable.
  2. to look at trends with previous surveys, comparisons with Police needed to be made based on the same statistics.

In the next iteration of the NZCASS, comparisons will be made against the police recorded crime victim statistics which are likely to be more comparable with NZCASS estimates.

Read more about police recorded crime victim statistics (RCVS):

ANZSOC Classifications

The Australian and New Zealand Standard Offence Classification (ANZSOC) is a hierarchical classification of three levels.

The justice sector often uses ANZSOC to report administrative statistics (such as Police statistics). It’s harder to do this with survey data (like from NZCASS).

Read more about the ANZSOC offence classification and coding process:

Read more about why ANZSOC classifications are not always used as part of NZCASS reporting:

Back to top

How do I compare NZCASS and Police statistics?

NZCASS statistics come from survey data, while Police statistics come from administrative data. You need to ask the questions below if you want to compare survey data and administrative data. These questions will help you check that only ‘like data’ is used when drawing comparisons. They will also help you assess any differences that might impact comparability.

Is the same type of data being compared?

NZCASS is a sample survey whereas the data used in historic police offence statistics was collected through administrative processes (information on offences was collected and stored electronically in the Police National Intelligence Application – NIA).

 Police statisticsNZCASS statistics
Method Census count of all offences recorded in the Police NIA Sample survey of 6,943 New Zealand residents (aged 15 years and over)
Content Historic police offence statistics count offences recorded by Police. Counts are based on the number of offences reported by the public or detected by Police Asks people about their feelings of safety and experiences of crime (along with a range of other questions)

Comparing these different types of data will affect the completeness and precision of the statistics produced.

 Police statisticsNZCASS statistics
Sampling error Administrative data is not subject to sampling error Because not all New Zealanders are interviewed about their experiences of crime, NZCASS statistics are subject to sampling error
Respondent reporting Administrative data holds actual records of incidents that are reported to or discovered by Police, and recorded in Police systems NZCASS asks if Police came to know about the incident but there is currently no accurate and robust way of verifying this with Police records. This means a respondent may say that the incident was reported to Police but it may not have been

Are offences classified in the same way?

As a general principle, offences in the NZCASS are coded:

  • in line with current legal theory
  • in line with current Police recording procedures.

Despite this, an exact match is unlikely because Police may make different judgments when deciding:

This means that some offences may have been placed into a different offence category, which will affect comparability.

Are the same types of offences covered?

The NZCASS does not collect information on all types of offences, whereas Police statistics do. For example, the NZCASS does not cover ‘victimless crimes’ (such as drug offences) and offences where the primary victim cannot participate in the survey (such as manslaughter or homicide), while Police statistics exclude some offences that NZCASS includes, such as offences not reported to Police.

Do the statistics include the same things?

Police statistics include a range of things that NZCASS statistics do not. The most common differences that affect comparability are age and commercial targets.

 Police statisticsNZCASS statistics


Includes victimisations of those aged under 15 years

Excludes victimisations of those aged under 15 years.

Commercial targets

Includes crimes committed against commercial targets, such as business vehicles and commercial premises

Excludes crimes against commercial targets.


(Note that commercial crime may sometimes be included because descriptions of events may be hard to understand.)

Police recorded crime victim statistics (RCVS)

Police victim statistics can include a range of crimes that NZCASS statistics do not.

 Police statisticsNZCASS statistics
Death-related offences


  • murder
  • manslaughter
  • driving causing death

Excludes offences were the victim is dead

Blackmail and extortion

Includes blackmail and extortion

While collected as part of NZCASS, blackmail and extortion are not counted

Theft from retail premises

Includes crimes where a business is the victim

Excludes crimes against commercial targets

Are offences counted or recorded in the same way?

Police may make different judgments when deciding whether to record an incident as an offence. The table below provides a range of examples where this might happen and how NZCASS statistics differ.

 Historic Police offence statisticsNZCASS statistics

Was it a series of the same type of offence?

In some cases, not all incidents in a series are counted.

For example, in cases of partner assault, when offences are repeated by the same offender against the same victim over a period of time, Police may count this as a single offence depending on the circumstances (such as if they don’t have specific details on individual offences).

NZCASS counts each incident.

NZCASS estimates for offences like assault will be higher than the Police statistics.

Was it a series of different offences? Where a series of offences have been committed not all of these may be counted:
  • Where one offence is part of another
  • One offence is the means to committing another
  • Where the offences are relatively minor
  • Where there is insufficient information (who, time, place, type what happened) to clearly distinguish between offences.
The NZCASS counts multiple offences in the same way as Police do wherever possible.
Is it dealt with as a crime?

Not all incidents reported to Police end up recorded as offences.  An offence is not recorded if, at any stage after the initial report, Police judge that an offence (in law) probably did not occur.

In addition, where offences are minor and Police and the victim believe attendance is sufficient to resolve the situation, these offences may also not be recorded (for example the case of some threats).

NZCASS counts each offence based on whether it meets the legal definition of a given crime based on the information provided by victims.

Statistics are not influenced by whether a victim wants an incident to be treated as an offence or not. As such, NZCASS estimates are likely to be higher in some cases.

Is there enough evidence? Police may not record offences reported to them when:
  • there is not enough evidence to establish whether an offence happened or not
  • there is evidence that the offence didn’t happen
  • they do not consider the person reporting the offence to be credible.
Within the NZCASS:
  • where there is not enough information to establish whether an offence occurred or not, the incident is coded as out of scope
  • NZCASS only gathers information from the victims’ perspective, so no alternative views can influence count estimates
  • no judgment about the credibility of the victim is made.

Are the same categories used to report the statistics?

Survey information is sometimes grouped differently from administrative data.

Is there sufficient sample? Survey data is sometimes grouped to make sure that statistics are based on large-enough sample sizes and are robust. Administrative data is often not subject to the same restrictions, given the amount of data available.

Because of this, groupings used to report survey statistics will not always align directly with the higher-level ANZSOC classifications because the sample is not big enough.
Is the data available? In some cases, not all elements of a given ANZSOC classification are available in NZCASS.

For example, the level 1 ANZSOC classification for ‘sexual assault and related offences’ includes elements of sexual violence that are out of scope for NZCASS, such as information about ‘child pornography’ and ‘non-assaultive sexual offences against a child’.

Likewise, for some offences/categories, the Police figures include a large number of different offences that would be difficult to untangle to match the coverage of NZCASS.
How do users need information reported? NZCASS analysis and reporting needs to be useful and relevant to a wide range of users. Some ANZSOC classifications may not be the best way to report information for the majority of these users.
Is reporting consistent? In some cases, the groupings used to report survey statistics need to be consistent with previous iterations of the research. This is done because some groupings are used to monitor specific trends. These groupings may not align directly with ANZSOC classifications.

Back to top

Which offences can be compared with official Police statistics?

Because of all the reasons discussed above, we have identified a subset of offences that can be compared. This subset includes:

  • thefts of vehicles
  • thefts from vehicles/vehicle interference
  • burglary
  • robbery/theft from the person
  • assaults.

Read more about the comparable subset in Concordance – NZCASS comparable offences groupings and ANZSOC classifications.

Concordance – NZCASS comparable offences groupings and ANZSOC classifications [PDF, 67 KB]

Back to top

What adjustments are made?

To make it easier to compare NZCASS and Police statistics, the following adjustments were made in 2014:

OffenceAdjustment made to Police dataImpact
Thefts of vehicles Remove commercial vehicles Police recorded 18,016 thefts of vehicles in 2013. This was reduced by 4.5% to account for commercial vehicles as targets – down to 17,205.
Thefts from vehicles / vehicle interference Remove commercial vehicles Police recorded 36,033 thefts from vehicles and vehicle interference in 2013. This was reduced by 3.9% to account for commercial vehicles as targets – down to 34,628.
Dwelling Burglary No adjustment needed.  See concordance.  
Robbery/theft from the person Remove victimisations against people 13 years and under Police recorded 4,186 victimisations in 2013. This was adjusted to 3,829.
Assaults Remove victimisations against people 13 years and under Police recorded 37,291 victimisations in 2013. This was adjusted to 32,264.

More information about how and why these adjustments were made can be found in the NZCASS Technical Manual.

NZCASS Technical Manual [PDF, 5.3 MB]

Back to top

This page was last updated: