Australian Institute of Criminology

Skip to content

Methodology

The DoVE is the result of a research partnership between the AIC and Victim Services NSW. Victim Services NSW is part of the NSW Justice Department and aims to provide information and support to victims of crime. The department’s services include free counselling, financial support, specialist support groups and the Victim Access Line. It also advises the state’s Attorney General and the NSW Parliament on issues related to victims and victimisation.

In mid-2013, the AIC initiated discussions with the then NSW Commissioner of Victims Rights about the possibility of using qualitative data held by Victim Services NSW, to explore the impact of crime on victims. The AIC proposed using data from psychological evaluations of victims of crime written by authorised report writers (ARWs). The ARWs were psychologists who were contracted by Victim Services NSW to provide an independent perspective on, and an assessment of, the impact of the crime and the victim’s current situation. The ARW could not be a member of the victim’s regular treatment team. These reports were collected as part of victim’s applications to the NSW state government compensation scheme. This scheme was mandated as part of the Victims Support and Rehabilitation Act (VSRA) 1996 (NSW). This Act has since been repealed and replaced by the Victims Rights and Support Act 2013 (NSW), which outlines the victim’s compensation scheme currently operating in NSW.

Under the previous Act, compensation claims could be made by primary or secondary victims, immediate family members of victims, or any other person who had a genuine interest in the victim’s welfare (VSRA 1996 (NSW) s 25). The claim had to be lodged with the Director of Victim Services NSW within two years of the act of violence. A compensation assessor then considered claims and may have required other supporting documentation such as medical or psychological examinations to make their determination. The ARW reports were submitted to this process as an independent perspective on the impact of the crime on the victim.

During this initial discussion, the following four categories of violent crime were identified by the two agencies as areas of mutual interest:

  • physical assault (including grievous bodily harm, serious assault and common assault);
  • sexual assault (including child victims and historical sexual assault cases);
  • domestic violence (including family and intimate partner violence); and,
  • robbery (including armed and unarmed robbery).

Within these four crime types, particular areas of interest were also identified. These were:

  • repeat victimisation;
  • secondary victimisation;
  • the impact of exposure to victimisation on quality of life;
  • timeliness in service delivery; and
  • commonalities and differences in individuals’ responses to crime victimisation.

Rather than focus on discrete projects, it was decided that the AIC would create a qualitative database using ARW reports. This would provide the flexibility necessary to thoroughly explore the crimes and themes identified above.

Until 2013, awarding compensation to victims of crime was governed by the Victim Support and Rehabilitation Act 1996 (NSW). This Act has been amended a number of times and as a result Victim Services NSW requested that a sample of ARW reports be drawn from between 2003 and 2012 as reports obtained within this period were considered the most comparable. It was therefore agreed that 400 reports would be drawn from between 2005 and 2006 and a subsequent 400 reports from between 2010 and 2011. These time periods were selected based on advice from Victim Services NSW and their knowledge of the data. The method of sampling is discussed below.

Ethical considerations

This research had a number of ethical considerations that needed addressing before it could proceed. The three most important considerations were the confidentiality of subjects’ information, accessing the reports without the consent of the individual and adhering to the NSW Health Records and Information Privacy Act 2002 (no 71) (hereafter referred to as the Privacy Act (2002)).

Confidentiality

To avoid identifying individual victims, confidentiality of the data was strictly controlled and all identifying information removed from the reports prior to inclusion in the database. Deidentification took place in two stages.

An AIC researcher carried out stage one at the Victim Services NSW premises in Parramatta. After being scanned from hard copy and converted to a Microsoft Word document, all ARW reports were initially redacted. During this stage, information that could identify the victim, offender, their family or associated non-professional individuals (eg witnesses) was removed including:

  • name;
  • date of birth;
  • address/telephone number;
  • educational institutions attended; and
  • places of employment.

It was important that this process did not ultimately limit the researcher’s ability to distinguish between individuals involved in the incident, so names were replaced with repeating letter sequences. These sequences were standardised and were also used to retain information related to the redacted person’s relationship to the subject of the report. For example, the subject of the report was always identified by the sequence XXXX. If the report referred to XXXX’s mother or father, the mother’s name would have been redacted as MMMM and the father’s to FFFF. A full list of sequences is included in Appendix A. This approach did not affect the confidentiality of individuals, as it was not possible to identify XXXX on relationship alone. Where more than one person shared an acronym (for instance, in the case of multiple offenders), they were distinguished by a number at the end of the sequence (ie OOOO1, OOOO2…OOOON).

Stage two involved further redaction after the de-identified reports had been moved to a restricted access folder on the AIC server. During this stage, information pertaining to professionals including police officers, counsellors and medical professionals was also removed. This included:

  • the professional’s name;
  • business affiliation/name of business;
  • ARW number; and
  • date of report.

Maintaining confidentiality is not just limited to the reports themselves. Any publication that uses DoVE data will adhere to guidelines around confidentiality. Distinguishing information about the crime itself will also be removed from any specific case studies that are to be used in publications to illustrate key points about victimisation. This will limit the likelihood of identification through association.

Consent

Individuals who were the subject of ARW reports included in the DoVE were not contacted to provide their consent. The National Statement (2007) has clear guidelines around the conditions necessary for a waiver of consent to be granted:

  • that the involvement in the research carries no more than low risk…to participants (the National Statement section 2.3.6a);
  • the benefits from the research justify any risks of harm associated with not seeking consent (2.3.6b);
  • it is impractical to obtain consent (for example, due to the quantity, age or accessibility of records) (2.3.6c);
  • there is sufficient protection of their privacy (2.3.6e); and
  • that there is an adequate plan to protect the confidentiality of the data (2.3.6f).

Broadly, these guidelines required the AIC to demonstrate that the DoVE posed negligible risk to individuals who were the subject of the ARW reports and that it was impractical to obtain consent.

The AIC contended that the potential risk and associated harm to the individual that could have been caused through identification has been controlled, as described above, through strict confidentiality guidelines and redaction processes. It is therefore highly unlikely that an individual would be identified either through information contained within the database or upon publication. Further, the method undertaken here does not require victims to recount or remember their experiences of crime. As a result the serious risk of re-traumatisation is negated.

Two factors determined the impracticality of gaining consent from each person who was the subject of an ARW report and included in the database. The final sample was to comprise 800 reports and it would have been impossible to individually contact the subject of each report within a reasonable timeframe. In addition, the decision to sample from cases that were nearly seven years old also presented a similar challenge, given that it was likely that some individual contact details would have changed and not been updated; particularly if their involvement with Victim Services NSW had ceased.

Privacy

As Victim Services NSW originally collected the information to evaluate compensation claims, this research also had to adhere to the guidelines specified in the Privacy Act (2002). The Act states that the data can be used for research as long as individuals cannot be identified from the information provided or in subsequent publications. For example, health information can be used for a secondary purpose other than the primary purpose it was collected for if:

‘[T]he use of the information for the secondary purpose is reasonably necessary for research, or the compilation or analysis of statistics, in the public interest and;
(i) either:
(A) that purpose cannot be served by the use of information that does not identify the individual or from which the individual’s identity cannot reasonably be ascertained and it is impracticable for the organisation to seek the consent of the individual for the use, or
(B) reasonable steps are taken to de-identify the information, and
(ii) if the information could reasonably be expected to identify individuals, the information is not published in a generally available publication, and
(iii) the use of the information is in accordance with guidelines, if any, issued by the Privacy Commissioner for the purposes of this paragraph…’ (Schedule 1.10f)

This research satisfies criteria B. Steps have been taken to redact all identifying information making it highly unlikely that an individual who is the subject of a report would be identified from within the database or upon publication.

Victim Services NSW and the AIC Human Research Ethics Committee were satisfied that the research met ethical standards and the project was approved in August 2013 (protocol number P0207).

The data

Until early 2013, individuals seeking compensation from Victim Services NSW regarding violent crime victimisation had to be evaluated by an independent ARW. The information gathered would be used to determine the victim’s level of psychological harm that could be attributable to the act of violence. Some ARWs also used the opportunity to suggest treatment or support options they felt to be appropriate in these instances. The reports contain comprehensive information pertaining to the crime as well as the victim’s reaction and experiences during and after the offence. Though the level of detail varies depending on the ARW, reports also contain information on:

  • victim demographics;
  • victim’s and family’s medical/psychiatric history;
  • victim’s school/employment history;
  • psychological consequences of victimisation experience;
  • victim’s scores on psychometric tests (eg The Beck Depression Inventory, The Clinician Administered PTSD Scale, Global Assessment of Functioning); and the
  • ARW’s professional opinion regarding level of psychological harm and suggestions for treatment/services.

Victim Services NSW estimates that it had received between 2,000 and 3,000 ARW and counselling reports each year since 1996. However, due to changes in legislation, reports received between 2004 and 2013, specifically 2005, 2006, 2010 and 2011, were determined to be the most comparable.

It is important to note that this is a purposive, non-probability sample drawn from individuals who sought compensation from Victim Services NSW over two defined time periods and thus cannot be considered as able to be generalised to the wider population of victims of violent crime.

Sampling

Victim Services NSW received a total of 34,948 ARW reports from September 1997 until March 2013. The AIC was supplied with a de-identified index of all ARW reports to undertake the sampling process.

The aim was to collect approximately 200 cases from each of the four identified crime categories—physical assault, sexual assault, domestic violence, and armed robbery. However, a truly random sample was considered inappropriate for this type of research as it may have led to the exclusion of less common types of victims. For example, it was expected that for a crime like sexual assault most victims would be women. It was therefore important to ensure that where the information was available, the experiences of male victims were also included in this sample. In light of this, the decision was to select cases to include through a process of random stratification.

The full sample was stratified into three levels: crime, date of lodgement and gender of the subject of the report (Figure 1). As the focus of the database was to collect information pertaining to the four crime types of interest, the first strata separated out the cases detailing the experiences of victims of domestic violence, physical assault, robbery and sexual assault. The decision to further stratify by date was made in response to the Victim Services NSW request to sample reports received from 2005, 2006, 2010 and 2011. This was to ensure comparability between reports. Finally, it was decided to include equal numbers of male and female victims. This ensures the sample will adequately capture the experiences of men, women and children, especially in crimes where victims of a particular gender may dominate.

From within these three levels of stratification, cases pertaining to 800 victims were randomly selected. Each case was assigned a number using the random number generator command in Excel. The 50 cases within each group were then selected by including those with the 25 highest and the 25 lowest numbers. The final list of 800 was sent to Victim Services NSW so that they could request the hard copy case files for data collection.

Figure 1 Stratified sampling methodology

Data collection

Data collection took place on site at Victim Services NSW, Parramatta, in October 2013 and February 2014. An initial 619 cases were obtained over six-days in October 2013, with a final 119 collected over three days in February 2014. A total of 62 cases could not be included due to problems encountered during data collection. Some of the cases could not be found, the files presented did not contain an ARW report, or the compensation claim was currently being processed by Victim Services NSW and was therefore not available to the researcher.

The reports were stored as part of hard copy case files for each individual victim. Initially, each report was scanned and converted from a pdf to a Word file using Adobe X Pro software. In some instances this affected the quality of the report but only to the extent that a few words in affected documents were illegible. All reports were usable following the conversion.

Information identifying the subject and offender was redacted by an AIC researcher onsite at the Victim Services NSW premises under the supervision of a member of the Victim Services NSW staff. Reports were then transferred to a secure location on the AIC server where they were further redacted as required by the AIC ethics committee and as described above.

The final sample

The final sample included a total of 730 cases:

  • 182 victims of domestic violence (25%);
  • 176 victims of physical assault (24%);
  • 179 victims of sexual assault (24%); and
  • 193 victims of robbery (26%).
Table 1 Gender, age and type of victim – DoVE population sample
Domestic violence Physical assault Sexual assault Robbery
Gender
Male 79 105 91 100
Female 103 71 88 93
Type of victim
Primary 146 151 145 174
Secondary 36 25 34 19
Age
Less than 4 years 3 0 0 0
4 to 10 years 27 3 13 2
10 to 14 years 25 7 6 2
14 to 17 years 12 10 17 0
18 to 19 years 14 5 10 5
20 to 30 years 22 36 32 37
30 to 40 years 32 40 34 37
40 to 50 years 25 30 28 32
50 to 60 years 7 25 17 37
Greater than 60 years 9 10 5 26
Not specified 6 10 17 15
Total 182 176 179 193

The database

The database was constructed using NVivo 10 software. This software allows researchers to organise, categorise, and analyse qualitative information so that they can identify themes and characteristics. This made it well suited for the task of categorising 730 reports.

A framework was designed to classify each report based on the type of victim and other information represented in the report. This framework reflects the interactions between the subject, the offender, the situation and the various systems that may have involved the subject/offender. A total of 72 items are included in the framework and were selected based on AIC/Victim Services NSW discussions on data that would provide flexibility in forming and addressing research questions still to be developed (Table 2). This included items that captured characteristics reflecting a victim’s pre-crime functioning to give meaning to the analysis of impact.

To avoid the framework becoming too large and unmanageable, a number of ‘flags’ were inserted. These indicate when the report might contain information that would be of interest to future researchers, but which may be too difficult to classify in the earlier stages of development and use. The flags are basic yes/no indicators used when a particular characteristic was either present or absent. For example, if a subject reported that their mother suffered from depression, then the family history of mental illness flag would be classified as yes. These flags are also shown in Table 2.

The framework does not categorise the data; rather it creates a list of variables that researchers can subsequently use for the basis of their analysis. After each case was classified, the next step was to code them using a series of ‘nodes’. Nodes reflect the themes that researchers may be interested in exploring and were also deliberately kept broad. The structure of the nodes loosely follows the layout of the reports. A list of base nodes is presented in Table 3.

Table 2 Framework matrix used to classify reports

Victim

Offender

Situation

Systems

Victim

Type of victim
Gender
Gender of victim (if different)
Age
Age of victim (if different)
Age at crime
Occupation
Marital status
Number of children
Structure of family of origin
Pre-existing mental illness flag
Pre-existing AOD issues flag
Pre-existing physical disability flag
Pre-existing intellectual disability flag
Pre-existing medical illness flag
Prior history of victimisation flag
Family history of mental illness flag
Family history of AOD issues flag
Family history of physical disability flag
Family history of intellectual disability flag
Family history of medical illness flag
Family history of victimisation (not related to current claim)
CALD background
Indigenous status
Country of birth

Relationship
Number of victims

Relationship with victim (if a secondary claim)
Injury sustained
Medical attention required

Educational attainment
History of foster care/OOHC flag
Type of foster care/OOHC placement
Number of foster care/OOHC placements
Prior criminal history flag
Prior incarceration flag
Prior contact with government services flag (not victim services) flag
Prior contract with victim support services flag
Prior compensation claim flag
Reported crime to the police flag

Offender

Gender of offender
Age of offender
Age of offender at crime
Occupation
Mental illness flag
Physical disability flag
Intellectual disability flag
Medical illness flag
History of victimisation flag
AOD abuse issues flag
Number of offenders
Indigenous status
CALD background

Weapon
Weapon use
Level of violence

Prior criminal history flag
Prior incarceration flag
Prior contract with government services flag
Prior contract with victim support services flag
Prior history of foster care/OOHC flag

Situation

Category of crime seeking compensation for
Crime year start
Crime year end
Duration of victimisation
Location
Witnesses flag
Number of witnesses
Presence of security measures

Police involved
Outcome

Systems

Multiple support organisations flag

Table 3 Base node of the DoVE
Macro nodes Micro nodes
Level 1 Level 2 Level 3 Level 4 Level 5
ARW Observations
Opinion
Recommendations
Sources
Pre-incident Compensation
Criminal
Demographics
Education
Employment
Family
General functioning
Medical
Psychological Intellectual
Mental health
Referral
Relationship
Social
Stressors
Substance Alcohol
Drugs
Victimisation
Peri-incident Description
Post-incident Impact Functioning
Physical
Psychological Psychometric Anxiety
Depression
Functioning
Other