Explanatory notes

The Crime Statistics Agency (CSA) presents statistics about the characteristics of crime recorded on the Victoria Police Law Enforcement Assistance Program (LEAP). The following explanatory notes are designed to provide additional information about the data the CSA receives from Victoria Police, how it is processed and how to interpret the summary statistics

Data source

The crime statistics produced by the CSA are derived from administrative information recorded by Victoria Police and extracted from the LEAP database. Victoria Police provides this information to the CSA 18 days after the reference period.

As the LEAP database is a live operational data system and updated regularly, the data presented reflects only the information in the database at the date and time of extraction. This means that as additional quarters of data are released by the CSA, the data relating to previous periods may change as data are updated in LEAP, investigations progress and cases are completed by Victoria Police.

Scope and coverage

The CSA recorded crime collection includes all offences that are reported to, and detected by, Victoria Police and recorded on the LEAP database. The scope and coverage of the data, however, is not representative of all crime that occurs in Victoria. Some crimes may not be recorded on LEAP, not be reported to police, or the responsibility for responding to certain offences may lie with another agency.

The following data are not available to the CSA and are not included in these statistics

  • missing person details;
  • police custody information;
  • traffic infringements;
  • regulatory activity not directly undertaken by Victoria Police, including infringement issuing and management;
  • Victoria Police staff and human resource management information (including financial and asset information);
  • information about Victoria Police operations and taskforces;
  • areas of Victoria managed by federal agencies, such as crown land and Melbourne airport, which are under the jurisdiction of the Australian Federal Police;
  • investigations managed by Australian Government agencies, such as the Australian Crime Commission; and
  • information related to prosecutions.

 

Data in the CSA Crime by location tool excludes offences that are recorded in LEAP but were committed outside Victoria, and where an LGA is not recorded. This is because these offences cannot be meaningfully rendered on the map of Victoria.

Comparisons between Victoria Police and Crime Statistics Agency statistics

The following outlines differences in the scope and counting rules of recorded crime statistics produced by Victoria Police and the CSA. Crime statistics previously produced by Victoria Police excluded the following Penalty Infringement Notices (PINs) which are now included in CSA counts:

  • 549MP - CONTRAVENE POLICE DIRECTION TO MOVE ON
  • 596A - DRUNK IN PUBLIC PLACE
  • 596B - DRUNK AND DISORDERLY IN PUBLIC PLACE
  • 599HC - BEHAVE IN DISORDERLY MANNER PUBLIC PLACE

Where a single offence has multiple weapons recorded against it, Victoria Police historically selected the first weapon to appear on the dataset for the particular offence. The CSA selects the most serious weapon that appears on the record (for example, a handgun will be selected over a knife, and so on).

Offence categories

The CSA developed an offence classification for statistical output purposes. This offence classification has been mapped to all raw offences recorded by Victoria Police. In comparison with the categories used historically by Victoria Police for statistical reporting, the CSA offence classification contains more detailed categories and reduced the number of offences mapped to Other, Missing and Unknown categories.

Due to these differences and additional changes to the calculation of rates, the CSA advises that data previously published by Victoria Police should not be compared with CSA recorded crime statistics.

 

Reference periods

The reference period is the length of time that the statistics relate to. The CSA will produce three quarterly year-to-date statistical reports a year, and one annual statistical report for the financial year. Each report is based on 12 months of data with different reference periods. This is outlined in the table below:

Report title

Reference period

Month of release

Year ending 31 March

(Quarterly statistics)

1 April to 31 March June

Annual report to 30 June

 

1 July to 30 June October

Year ending 30 September

(Quarterly statistics)

1 October to 30 September December

Year ending 31 December

(Quarterly statistics)

1 January to 31 December March

 

The ‘Latest crime data’ section of the website shows the most recently published statistics, and links to the previous three quarters. The data presented in the crime by location map covers the most recent three years of statistics.

Composition of quarterly data for statistical reporting

Quarterly crime statistics produced by the CSA are based on a rolling 12 month set of statistics that collate four quarters of data. As such, three quarters from the previous reference period are carried forward into the next 12 month period, with the addition of the most recent quarter. This means that changes that may occur within one quarter will be included in four different crime statistics releases.

 

The reference period will be different depending on the period of time that the rolling 12 months of data covers. For example, data for the January to December reference period refers to the 12 month period beginning on the 1st of January through to the 31st of December of that year. In the March to April reference period that directly follows the January to December period, nine months of data from the previous reference period (March to December) is used with three months of new data (January to March) to compile a 12 month time period for analysis. This is outlined in the diagram below:

Reference periods based on the date records are created

The reference periods are based on the date that information is created in LEAP, regardless of when the offence occurred or when it was reported to police. The date the record was created is used because it is the date most consistently recorded on LEAP.

Counting methodology

Offences recorded

Recorded offences include any criminal act or omission by a person or organisation for which a penalty could be imposed by the Victorian legal system.

For the purposes of CSA statistics, an offence is counted and included in the data where it:

  • was reported to, or detected by, Victoria Police; and,
  • was first recorded in LEAP within the reference period.

 

The exception to this is those offences that are out of scope of the data collected by the CSA.

Depending on the type of offence committed and the outcomes of investigation, police may either initiate a court or non-court legal action against an offender. Non-court legal actions comprise legal actions such as informal or formal cautions or warnings and the issuing of penalty notices, which do not require an appearance in court.

Offences that are recorded but remain unsolved at the date the data was extracted are included in the CSA dataset.

Date of record creation

Offence level data are compiled on the basis of the date that the offence was created on the LEAP database, rather than the date the offence was detected by, or reported to police. The create date may not be the date when the offence occurred, or the date when the offence came to the attention of police.

The date the record was created is used because it is the date most consistently recorded on LEAP, and cannot be edited or updated. The date the offence was reported and the date the offence was committed can both be updated and changed at any stage of an investigation.

The date the offence was reported is included on the LEAP dataset provided to the CSA, but after conducting a quality assessment, the CSA has determined that the coverage of report date information in the data is of insufficient quality to support reliable calculation of the offence population on this date. The use of report date for statistical purposes will continue and be reviewed in the future as coverage and data quality improves.

Victim reports

A victim report is counted when an individual, business or organisation reports that they have been a victim of one or more criminal offences to Victoria Police and a record is subsequently made in LEAP. A victim report count involves only one victim but can involve multiple offences and alleged offenders. One report may involve offences that occur over a period of time but if processed by Victoria Police as one report it will have a count of one in the published figures. If there are multiple victims related to a criminal event, each will have their victim report counted once in the published figures.

An individual, business or organisation can be counted as a victim more than once within the reference period, if they have made more than one separate report to Victoria Police.

Where there were multiple offences recorded within the one victim report, the report is represented for statistical purposes by an assigned offence category of the most serious offence. This is known as the principal offence (see Principal variable calculations).

Date of record creation

Victim reports data are compiled on the basis of the date that the principal offence was created on the LEAP database, rather than the date the offence was reported to police. The create date may not be the date when the offence occurred, or the date when the offence came to the attention of police.

The date the record was created is used because it is the date most consistently recorded on LEAP, and cannot be edited or updated. The date the offence was reported and the date the offence was committed can both be updated and changed at any stage of an investigation.

Alleged offender incidents

An alleged offender incident is an incident involving one or more offences to which a person, business or organisation has been linked as an alleged offender. An alleged offender incident represents one alleged offender but may involve multiple victims and offences. One incident may involve offences that occur over a period of time but if processed by Victoria Police as one incident it will have a count of one in the data presented in this section. If there are multiple alleged offenders related to a criminal event, each will have their alleged offender incident counted once in the published figures.

There may be multiple incidents within the reference period that involve the same individual, business or organisation as an offender. Where there were multiple offences recorded within the one incident, the incident is assigned an offence category of the most serious offence in the incident for statistical purposes, known as the principal offence.

Date of result

Alleged offender incidents are compiled on the basis of the date that a result was recorded on the LEAP database. The date of result is used because it is the most consistent date recorded on LEAP and directly corresponds to the status of investigation relating to the incident.

Unique alleged offenders

A unique alleged offender is defined as a person who has been involved in one or more alleged offender incidents within a reference period. One unique alleged offender may be involved in more than one offender incident during the reference period, but will have a count of 1 in the data presented in this section. If there are multiple unique alleged offenders related to a criminal event, each unique offender will be counted once in the published figures.

Date of result

Unique alleged offenders are compiled on the basis of the earliest date that a result was recorded on the LEAP database within each reference period. The date of result is used because it is the most consistent date recorded on LEAP and directly corresponds to the status of investigation relating to the incident.

Repeat offending 

Repeat offending is defined as those people who have been linked to more than one criminal incident within a 12-month financial year reference period. People who have been alleged offenders across financial year reference periods will be counted in each of the reference periods, therefore the repeat offending counts in this publication do not show repeat offending over a longer period of time.

Unique victims

A unique victim is defined as a person who has had one or more victim reports recorded by Victoria Police within the reference period. One unique victim may have more than one victim report recorded during the reference period, but will have a count of 1 in the data presented in this section. If there are multiple unique victims recorded against one criminal event, each unique victim will be counted once in the published figures.

Date of record creation

Unique victims are compiled on the basis of the date that the first report was created on the LEAP database within the reference period, rather than the date the offence was reported to police. The create date may not be the date when the offence occurred, or the date when the offence came to the attention of police.

Repeat victimisation 

Repeat victimisation is defined as those people who had more than one victim report within a 12 month financial year reference period. People who have been victimised across more than one financial year reference period will be counted in each of the reference periods that they reported being a victim. Consequently, the repeat victimisation counts in this publication do not show repeat victimisation over a longer period of time.

Family incidents

A family incident is an incident attended by Victoria Police where a Victoria Police Risk Assessment and Risk Management Report (also known as an L17 form) was completed.

A family incident can involve one or more affected family members and/or one or more other parties. For statistical purposes, these are counted as one incident but may appear multiple times in demographic counts.

The increase in the number of recorded family incidents in the past five years has in part been due to improved recording of incidents. Since 2011, initiatives such as the Family Violence Code of Practice have been put in place by Victoria Police to improve the recording of family incidents, the individuals involved and the offences committed.

Demographic characteristics of affected family members and other parties

An ‘affected family member’ is the individual who is deemed to be affected by events occurring during the family incident. The other individual involved in a family incident is referred to as the ‘other party’. The other party could be a current partner, former partner or a family member.

Where an affected family member has been affected by more than one other party within a family incident, they will be counted for each involvement. For example, where a family incident involves three affected family members and one other party, each affected family member will be counted separately, making a count of three. 

Where the other party is involved with multiple affected family members, they will be counted for each involvement. For example, where a family incident involves one affected family member and two other parties, each other party will be counted separately, making a count of two.

Where an individual is involved in multiple family incidents within the reference period they will be counted for each incident that they are involved in. 

Date of record creation

Family incidents data are compiled on the basis of the date that the incident was created on the LEAP database, rather than the date the incident came to the attention of police. The create date may differ from the date when the incident occurred, or the date when the incident came to the attention of police.

The date the record was created is used because it is the date most consistently recorded on LEAP, and cannot be edited or updated.

Aboriginal and Torres Strait Islander status 'most frequent' rule

A ‘most frequent’ rule has been adopted as the primary counting rule for the Aboriginal or Torres Strait Islander status of offenders in CSA recorded crime statistics. Under this counting rule, a person has either a yes or no response to the Standard Indigenous Question (SIQ), then the most frequently appearing response is taken as correct. If the person only has one meaningful response (‘yes’ or ‘no’), then that response stands across all records. If a person appears in the dataset two different times with a ‘yes’ and a ‘no’ in the records, a ‘yes’ response is taken over a ‘no’ response. Figures 1 and 2 illustrate how this counting rule operates in two different scenarios.

A most frequent response may be less susceptible to data entry error, as a person with a number of appearances would require multiple incorrect responses in the dataset to be incorrectly assigned. The more entries a person has in the dataset, the greater the likelihood that their data is correct. The only exception to this would be where a person wishes to change the way they identify and respond to the question over time. This would not be recognised until the majority of a person’s entries in the database reflected their current status. This rule is also dependent upon the SIQ being asked and recorded on each occasion.

Figure 1. Illustrative example of the application of a ‘most frequent’ derivation rule to output from the Aboriginal and/or Torres Strait Islander status variable relating to an offender

Figure 2. Illustrative example of the application of a ‘most frequent’ derivation rule to output from the Aboriginal and/or Torres Strait Islander status variable relating to an offender

Principal variable calculations

Some variables in the recorded crime dataset may legitimately have more than one item recorded against them. To represent this data in a summary form, the multiple responses are ordered using hierarchical classifications, which allow the CSA to select a principal response to represent each record.

Principal Offence

Offence categories presented in the alleged offender incidents, victim report and family incident tables refer to the principal offence representing the incident. Where there is only a single offence attached to a unique incident, that offence is the principal offence by default. Where multiple offences are recorded within the same incident, a principal offence is assigned using the CSA Offence Index. 

CSA Offence Index

The CSA Offence Index is a tool by which the seriousness of offence types can be ranked against each other in order to calculate the most serious offence (principal offence). The CSA Offence Index was largely adapted from the Australian Bureau of Statistics National Offence Index (cat. no. 1234.0.55.001). The diagram below describes examples of how the principal offence is determined based on seriousness.

Example Incident A: Where an incident involves one offence of Murder, one offence of Stalking and one offence of Breach of bail, the principal offence would be presented as Murder.

Example Incident B: Where an incident involves one offence of Serious assault and one offence of Offensive language, the principal offence would be presented as Serious assault.

Example Incident C: Where an incident involves only one offence of Graffiti, then the principal offence would be presented as Graffiti by default.

Location type

For offences where more than one location type is recorded, the location type is selected based on the following hierarchy:

  1. Residential location
  2. Community location
  3. Other location

For more information on the location type index, please see the location type classification.

Relationship of victim to alleged offender 

For victim reports where more than one relationship type is recorded, the relationship type is selected based on the following hierarchy:

  1. Current partner
  2. Former partner
  3. Family member
  4. Non family member
  5. Not known to victim (stranger)
  6. Unknown or not recorded relationship

For more information on the relationship type index, please see the relationship type classification.

Regional statistics

Recorded crime statistics for offences, alleged offender incidents, victim reports and family incidents are presented by Police region and Local Government Area (LGA). The CSA also presents offences data by postcode and suburb in the Crime by location tool. For more information on the geographic locations used in the CSA data please see the geographic location hierarchy (External link).

Improved location information 

The CSA has analysed the recording of geographic data in LEAP and has found that there are some inconsistencies which impact the overall quality of location-specific information. The CSA has used a combination of different location variables received from Victoria Police to improve the quality of location data, which better represents where a specific incident occurred. This work has improved the quality of location-based information to inform the public about where crime occurs across the state, and has been implemented for offences data. These changes are visible in the data published for the year ending June 2017 release onwards. 

Justice and Immigration Institutional Facilities 

For the purposes of statistical reporting, a number of facilities are now counted separately from the LGA, postcode or locality in which they are located. These include correctional facilities, youth justice facilities and immigration detention centres, and are categorised as ‘Justice institution or immigration facility’. These facilities are counted separately from the year ending June 2017 release onwards.

The CSA has identified justice institutions or immigration facilities by using their street address, location type and location description information. If there is uncertainty about where an incident occurs, the CSA will continue to show the offence in the crime counts for the area (at LGA, postcode or suburb level).

The following are included in the ‘Justice institution or immigration facility’ category:

  • Barwon Prison
  • Beechworth Correctional Centre
  • Dame Phyllis Frost Centre
  • Dhurringile Prison
  • Fulham Correctional Centre
  • Hopkins Correctional Centre (inc. Corrella Place)
  • Judy Lazarus Transition Centre
  • Langi Kal Kal Prison (inc. Emu Creek)
  • Loddon Prison (including the Middleton Annexe)
  • Malmsbury Youth Justice Centre
  • Maribyrnong Immigration Detention Centre
  • Marngoneet Correctional Centre (including the Kareenga Annexe)
  • Melbourne Assessment Prison
  • Melbourne Youth Justice Centre (Parkville)
  • Metropolitan Remand Centre
  • Port Phillip Prison
  • Tarrengower Prison

 

Incidents that occur at facilities such as Corella Place or Emu Creek are included in this category, as the CSA cannot effectively distinguish between these locations and the adjacent prison using the location recorded by Victoria Police.

The following locations have been excluded from this category:

  • Melbourne Custody Centre – This centre cannot be distinguished from the courts in the data, and is not deemed a justice institution that permanently holds prisoners.  However, convicted or unconvicted persons may be detained temporarily in these facilities.
  • Thomas Embling Hospital – This hospital is a partially secure facility that treats patients from within the criminal justice system and the mental health system, however not all patients within this facility are serving correctional sentences.
  • Wulgunggo Ngalu Learning Place –this is a transitional facility for offenders on Community Corrections orders and is used to provide services such as employment, education and life skills.
  • Police cells – as police cells are managed by Victoria Police and do not permanently hold convicted offenders, these are not considered justice institutions or immigration facilities. However, convicted or unconvicted persons may be detained for a short period of time in these facilities.

Any incidents that occur at these locations will still be included in localised crime counts.

Rates per 100,000 population

Rates per 100,000 people in Victoria are calculated for offences, alleged offender incidents, victim reports and family incidents.

Rates per 100,000 population are derived using the incident, report or offence count for the reference period and the most recent Estimated Resident Population (ERP) data.

Rates are calculated using the following formulae:

  • Offence rate = (Offence count/ERP count) *100,000
  • Alleged offender rate = (Alleged offender incident count/ERP count) *100,000
  • Victimisation rate = (Victim report count/ERP count) *100,000
  • Family incident rate = (Family incident count/ERP count) *100,000

ERPs for both Victoria and Local Government Areas are based on populations provided by the Australian Bureau of Statistics. ERPs for the data in the current reference period are based on population projection estimates developed by the Victorian Government’s ‘Victoria in Future’ program. For years prior to the current reference period, the ERP used to calculate offence rates is the ABS ERP.

ABS ERP data comes from two publications:

  • Estimated Resident Populations by age and sex are collected from the Australian Demographic Statistics (3101.0).
  • Estimated Resident Populations by Local Government Area are collected from the Regional Population Growth (3218.0).

For more information about the ABS estimated resident population, refer to the ABS website (External link).

For more information about the 'Victoria in Future' report, refer to the Department of Economic Development, Jobs, Transport and Resources website (External link).

Victorian population figures used for year ending June 2017 publication 

ABS - Australian Demographic Statistics

Victoria in Future

Jul – Jun 2013

Jul – Jun 2014

Jul – Jun 2015

Jul – Jun 2016

Jul – Jun 2017

5,733,545

5,838,110

5,946,505

6,069,636

6,157,538

24 month trend test

The trend test presented in the data tables highlights movement in data that is of a consistent and continuing nature over the previous 24 months. The CSA uses the Kendall's Rank Order Correlation statistical test (or Kendall's tau-b) to determine whether a series is trending upwards, downwards or is stable[1] over the specified time period. The procedure that the CSA uses is to conduct the Kendall’s Rank Order Correlation on the monthly total number of offences for each offence type and LGA over the previous 24 months.

From this release onwards, the CSA also applies a threshold that involves the satisfaction of one of two criteria, in order for the trend test to be conducted. If a category fails both sets of criteria, then the significance test will not be conducted.

1.Less than 30 offences in any month – This approximates to one offence per day and ensures that there is sufficient data of a sufficient quality before it is analysed.

2.Percentage Proportion threshold (<0.1% of all recorded offences) – To ensure that the data for a particular category contributes a meaningful proportion of the overall before it analysed.

This two-pronged threshold, means that Offence categories and LGA’s will only be excluded if the number of offences recorded are less than 30 in any given month and the proportion of overall criminal offences is less than 0.1%. Note that in very few circumstances, the significance test will show a significant trend, even when the yearly percentage change is very low or in the opposite direction. In other cases, the test will be nonsignificant, even when the yearly percentage change is very high. This can occur in cases where there are seasonal or non-linear variations in the data, or if extreme spikes in the data are present. Kendall’s Rank Order Correlation test is not robust against these variations, and is only sensitive to generally increasing and decreasing trends.

Confidentialisation

Confidentialising data involves removing or altering information or collapsing detail (through application of statistical disclosure controls) to mitigate the risk that a person or organisation may be identified in the data (either directly or indirectly).

Alleged offender incidents, victim reports and family incidents data contain person-based variables and include demographic information. Therefore, these datasets are subject to confidentialisation to ensure the anonymity of individuals is protected where numbers are small and there is a reasonable likelihood that a person may be identified from the data published.

The CSA will confidentialise cells in a table that are between 1 and 3. This is denoted in the tables by the value “≤ 3” appearing in cells with small numbers.

For the purpose of calculating row and column totals, each cell between 1 and 3 is assigned a value of 2, regardless of the true number of that cell. This methodology allows for totals to be calculated in tables with small cells, but this does mean that totals for certain variables may not be the same across tables within a publication or set of data cubes. This process is applied prior to the release of statistical data by the CSA.

Legislative changes significantly affecting recorded crime statistics

Breach of bail conditions

Amendments to the Bail Act 1997 which were introduced in December 2013 inserted the following sections into the act:

  • S30A Offence to contravene certain conduct conditions
  • S30B Offence to commit indictable offence whilst on bail

These amendments resulted in the introduction of two new offence codes on LEAP. There has subsequently been an increase in the number of offences recorded against the category Breach of bail conditions.

Breach of family violence orders

The Justice Legislation Amendment (Family Violence and Other Matters) Act 2012 inserted the following sections into the Family Violence Protection Act 2008:

  • S37A Contravention of notice intending to cause harm or fear for safety
  • S123A Contravention of order intending to cause harm or fear for safety
  • S125A Persistent contravention of notices and orders

Sections 37A and 123A make it an indictable offence to contravene a Family Violence Safety Notice or Family Violence Intervention Order where there was intention to cause harm or fear of safety to the person protected by the notice or order.

Section 125A makes it an indictable offence to persistently contravene Family Violence Safety Notices or Family Violence Intervention Orders.

The above amendments came into effect in April 2013 and resulted in the introduction of three new offence codes on LEAP. There has been a subsequent increase in the number of offences recorded against the category Breach of family violence orders.

Operational changes affecting recorded crime statistics

Commit indictable offence whilst on bail

In November 2014, Victoria Police changed their operational procedures for the recording of some breach of bail charges, affecting the way these offences are captured for recorded crime statistics. This change has impacted the number of offences recorded for ‘527Z Commit indictable offence whilst on bail’, and as a result the number of offences recorded in this category may be understated.

This change has not had any impact on the recording of other breach of bail offences in LEAP. The CSA is assessing the impact of this change for future releases.

Recording of ‘Fail to stop’ offences

From 13 July 2015, Victoria police changed their operational procedures in relation to ‘Fail to stop’ offences. These changes have led to these offences now being recorded in LEAP and included in the extract of recorded crime data provided to the CSA. This results in an increase in the number of offences recorded against the following Road Safety Act (1986) offences:

  • 749AUC Fail to stop vehicle on direction
  • 749XM Fail to stop vehicle on request

As a result, there has been an increase in the CSA offence category ‘E13 Resist or hinder officer’ since October 2015. For the current reference period there were only offences recorded for ‘749AUC Fail to stop vehicle on direction’.

Abbreviations used in the data

For ease of reading, some CSA offence terms have been abbreviated throughout this publication. The term 'and related offences' has been omitted from the following CSA offence category names:

  • Homicide and related offences
  • Assault and related offences
  • Abduction and related offences

In addition, the following CSA offence terms have been abbreviated as follows:

  • Stalking, harassment and threatening behaviour appears as 'Stalking/harassment'
  • Dangerous or negligent acts endangering persons appears as 'Dangerous/negligent acts'

For further information about CSA offence classifications, refer to the CSA offence classification or the glossary and data dictionary section of the website.

Revisions

Where required, the CSA may revise historical data in the most recent statistical releases to reflect the most up to date information recorded.

Additional data

Quarterly statistical releases are designed to provide a summary of recorded crime trends. Special tabulations may be able to be produced on request to meet individual user requirements. For further information contact the Crime Statistics Agency by email at info@crimestatistics.vic.gov.au (External link).

 
[1] Stable denotes that the 24 month trend was neither significantly increasing nor significantly decreasing.