Acknowledgements | Tables | Figures | Summary | Chapter 1 | Chapter 2 | Chapter 3 | Chapter 4 | Chapter 5 | Chapter 6 | Chapter 7 | Chapter 8 | Chapter 9 | Chapter 10 | Chapter 11 | Chapter 12 | Chapter 13 | Appendix 1 | Appendix 2 | Appendix 3 | Appendix 4 |
3.1 Research design
The evaluation had seven key components:
The database and file studies provided descriptive information both nationally and in selected districts.
The national survey, key informant interviews and interviews with potential applicants, applicants and respondents contributed to an understanding of the dynamics of the operation of the Act, as well as some descriptive information, and information on strengths and weaknesses in the implementation of the Act.
3.2 Research sites
While some data was gathered at the national level, more detailed information was collected in four courts selected to include a wide ethnic mix in the population, rural and urban situations and geographical spread. The research sites were:
The research sites were chosen to give an indication of the range of experiences in implementing the Domestic Violence Act, and were not themselves under investigation. For this reason, the report does not include detailed comparisons of the four research sites.
3.3 Participants and data analysis
Database study
The national estimates on Domestic Violence Act matters given in this report come from
administrative data sources. Information on applications for Domestic Violence Act orders
was obtained from the Domestic Violence Act database held by the Department for Courts.
The Ministry of Justice provided information on breaches of the Domestic Violence Act
obtained from the Law Enforcement System and the police provided information on reported
offences under the Domestic Violence Act.
Relevant data were obtained from the above sources for the period 1 July to 30 September 1998 (key tables are given in Appendices 1 and 3). In general, the analysis in the report is based on this three-month period. This is partly so that the data is consistent with the time period used for the file study (described below). However with the data from the Domestic Violence Act database, there are also issues of data quality, where the detail of data that was needed could not be provided reliably over a one-year period.
Yearly data is provided where feasible (see Appendices 1 and 3).
Please note that:
There are limitations in respect to the police data on offenders.
First, the data are not unique numbers. Since the data are collected in relation to recorded offences, and many offences involve multiple charges or multiple offenders, it is impossible to give an accurate count of individual offenders.
For example:
a) if an offender is apprehended on one occasion in relation to multiple offences, he or she will appear in the statistics for each offence ie. there will be multiple apprehensions involving the one offender;
b) if an offender is apprehended more than once in the 12 month period, that same person will appear in the statistics multiple times;
c) a single recorded offence may be resolved by apprehending multiple offenders.
Second, the statistics only relate to those offenders apprehended by police in relation to recorded offences. Police have no data on the age, gender and ethnicity of offenders who have committed the remainder of offences reported to police but whom the police have not managed to apprehend, or all other offenders whose crimes go unrecorded and undetected.
File study
The file study reviewed a selection of cases in each of the four research sites. The
cases were drawn from applications made between 1 July and 30 September 1998. Each file
was reviewed for eight months from the date it was opened. Key tables are given in
Appendix 2.
Table 1 below shows the number of cases available for each court and the number of cases selected. Reasons why files were not available include those unable to be found [Footnote 6] and those transferred to other courts. Overall, 95% of the files that were eligible for the study were available for selection.
Where possible, 100 cases were selected (systematically) from each court (101 were selected in Christchurch). For Whangarei and Lower Hutt this meant that all the available cases were selected. For Auckland Central and Christchurch, nearly all the available cases were selected.
In summary, 5% of cases were not available for the file study and a further 11 cases were not selected. Strictly speaking the file study cannot be seen as a census of the cases over the 1 July to 30 September 1998 period in the research sites. However, in practice, it does capture the great majority of cases over this period. It may be that the 5% of cases that were not available to the file study were different from those that were available, but the relative size of this group is such that it is unlikely to have a substantial impact on the patterns observed in the data. However the reader should be cautious when considering absolute numbers as these may be an under-count of the true values for the 1 July to 30 September 1998 period.
Table 1 Number of cases selected
The analysis of the file study data has raised concern about the quality of data on court files. Sometimes different pieces of information on a file are contradictory. Where these contradictions have been observed, advice has been taken from the court staff about which pieces of information are likely to be correct. The reader should see the results from the file study as a guide and should exercise appropriate caution.
Survey
A nationwide postal survey was conducted of all Family Court Judges, all Family Court
Coordinators and a sample of lawyers who have experience with the Domestic Violence Act
(1995). The survey was adapted for each group of recipients and had three main objectives:
The survey was used to gather a range of information from respondents. Table 2 shows the proportion of Family Court Judges and Family Court Coordinators and lawyers who answered the survey. Note that only a proportion of lawyers answered the survey and furthermore the survey of lawyers did not use a representative sample. Consequently, information provided from the lawyers should be seen as only representing those lawyers who responded and cannot be generalised to any wider group of lawyers.
Family Court Coordinators were asked to supply the names of lawyers in their area who were engaged in Family Court work. A sample of practitioners was drawn from this list.
Judges, Family Court Coordinators, domestic violence clerks and lawyers who were individually interviewed were not asked to complete a questionnaire. The questions were covered as part of their interview. The questionnaires are included as Appendix 4. The response to the survey is summarised below:
Table 2 Response to survey
Despite their names having been provided by Family Court Coordinators, several lawyers returned surveys saying they had not done any or enough domestic violence cases recently to warrant completing the questionnaire.
The surveys required participants to both select from options and to elaborate with comments. The researchers coded the comments and entered the data onto an Excel spreadsheet for analysis. As responses were received from relatively small numbers (although high proportions) of judges and Family Court Coordinators, percentages have not been used when discussing this data.
Key informant interviews
Key informant interviews were carried out at each research site and at the national
level with Family and District Court Judges (10), Department for Courts staff, including
bailiffs (20), lawyers (10), police (9), private process servers (1), programme providers
(17) and community groups (8). The interviews discussed issues raised in the file study
and survey as well as those identified in the scoping study completed in 1998 [Footnote 7].
Potential applicant, protected persons and respondent interviews
Finding people willing to share their experiences with interviewers was a complicated,
time-consuming and delicate process. For both safety and privacy reasons, it was essential
to work through intermediaries and to ensure that the interviews were carried out by
experienced and qualified interviewers. One person in each of the four research sites was
responsible for arranging the interviews. This group of four included a former social
worker, a mental health nurse with a social science background, a psychologist and a woman
with both a legal background and experience in the health area. Additional interviewers
were used as appropriate. These included Maori and Pacific interviewers and some men. All
were experienced, and received additional training for the job.
Interviews were carried out with 27 potential applicants, 41 protected persons and 43 respondents in total (see Table 3). Because the aim was to get as wide a cross-section of participants as possible, potential interviewees were sought through a variety of sources, including:
In the event, most protected person interviews were obtained through programme providers, women's refuges, marae-based services and community groups. Despite efforts to do so, it was not possible to interview any male applicants for this report.
Most respondent interviews were obtained through programme providers. This was not entirely satisfactory as only about a third of respondents complete programmes, meaning the sample was biased towards those who had complied with their order, and excluded respondents who were excused from programmes or who did not attend.
Finding potential applicants to interview was the most difficult aspect of the project. Interviewers in all sites reported difficulty in finding people who had experienced domestic violence but had not applied for a protection order. Interviewees were both difficult to locate and unwilling to talk about their experiences. In several cases, arrangements were broken because the potential interviewee was too nervous to speak out or actively feared for her safety. Because of these issues, the report does not include the experiences of the most frightened potential applicants or protected persons.
The safety of participants was the prime concern of interviewers. Intermediaries were asked to identify any potential safety issues and all interviewers had a protocol to follow should they fear for the safety of protected persons or children or for their own safety. Interviews were carried out at a place chosen by the interviewee and no interviews were carried out with a protected person and the respondent to the same order. With these precautions, no incidents occurred, although one field manager did arrange for support for a protected person with a young child.
The researchers discussed the nature and purpose of the interviews with intermediaries, stressing safety issues and the need for confidentiality and fully informed consent. Intermediaries then obtained consent for interviews from protected persons, respondents and potential applicants.
As the sample of potential applicants, protected persons and respondents was neither random nor systematic, the information gathered represents the range of experiences of these groups. It cannot be generalised to any wider group of potential applicants, protected persons or respondents.
Table 3 Characteristics of respondents,
Case studies of 'best practice'
This component of the research was designed as a means of capturing some of the good ideas
and innovation that Family Courts demonstrate in implementing the Domestic Violence Act
1995. The surveys and interviews asked judges, Family Court Coordinators and a sample of
Family Court lawyers for any examples of particularly good practice in elements of the
Domestic Violence Act process in the courts in which they worked. The intention was to
seek more detail from court staff in that area about the good practice which had been
highlighted by their colleagues. Perhaps reassuringly, informants felt overwhelmingly
positive about the performance of their local courts. However, this was not a great deal
of help in narrowing the field to select some examples of best practice for more detailed
examination.
In order to focus the search for examples, key processes of Domestic Violence Act implementation were identified for further examination. In processing Domestic Violence Act applications and subsequent protection orders, five points stood out as being areas in which several courts were dissatisfied with their own performance. These were:
It was decided to concentrate on these points in the process in the hope that dissatisfied courts might pick up ideas from others' practices. To help identify courts that performed well, the researchers met with the Domestic Violence Regional Advisors to seek their advice and consulted the Family Jurisdiction Team at Department for Courts National Office.
The various strategies led to 15 sites being identified, a number of them noted as having good practice in more than one of the areas of interest. This, and the subsequent interviews, confirmed what is already widely known -that it is important to have good selection processes to ensure that appropriate people are appointed. It must be stressed that the somewhat ad hoc method used to determine best practice sites will not have identified all sites in which high quality work and innovation is evident.
Phone interviews about best practice were conducted with 12 Family Court Coordinators. These courts were selected with a view to including a mix of larger and smaller courts, courts with a Domestic Violence clerk and those without. With a couple of exceptions the Family Court Coordinators interviewed were asked about their court's practice in each of the areas of enquiry.
1. In New Zealand, most applications under the Domestic Violence Act are from women and most respondents are men. For this reason, and to improve the readability of the report, the term 'he' is generally used in relation to respondents and the term 'she' when describing a potential applicant or protected person, in preference to the more correct 's/he'. The researchers acknowledge that some applicants are male and some respondents are female. The female respondents included in the study are clearly identified.
2. Applicants for a protection order become protected persons as soon as an order is granted. The report uses this terminology, although most key informants continue to describe protected persons as 'applicants', even after an order has been granted. This is apparent in some of the comments quoted in the report.
3. Respondents are persons against whom an application has been made for an order under the Act, and includes people (other than an associated respondent) against whom an order is made under the Act.
4. The court may direct that a protection order apply against a person whom the respondent is encouraging or has encouraged to engage in behaviour against a protected person, where that behaviour, if engaged in by the respondent, would amount to domestic violence. This person is called an associated respondent.
5. The terms 'withdrawal' and 'discharge' are used interchangeably by some people in relation to protection orders. In this report, the term 'withdrawal' is limited to applications that applicants seek to withdraw before an order is granted. Once an order is granted, any application to have it revoked is in effect an application for a discharge.
6. The report refers to 'undertakings'. This is a written undertaking by a respondent not to engage in violent behaviour. In return, the applicant withdraws the application for a protection order before the order is made. Undertakings have no legal standing and cannot be enforced by the police or the courts.
7. A protection order is first issued as a 'temporary' order. The respondent has three months in which to give notice of his or her intention to defend the order. If there is no defence, the order automatically becomes a 'final' order and stays in place until the protected person or the respondent applies for and is granted a discharge by the court.
8. The court may grant a variety of orders in association with a protection order. These include a 'furniture order' and an 'ancillary furniture order'. Under a 'furniture order', the protected person may remove the furniture from the house she previously shared with the respondent. An 'ancillary furniture order' is always made in conjunction with an occupation or tenancy order. In this case, the protected person retains possession of the house and the furniture.
9. Percentages are given only for data drawn from the national domestic violence database study, police statistics, the Law Enforcement System of the Ministry of Justice and from the file study. Actual numbers are given for information from the surveys of judges, Family Court staff and lawyers.
10. This report provides a number of descriptive statistics on issues related to the Domestic Violence Act. Because there are quality concerns about the data, the statistics should be treated with caution (see methodology section for more discussion on data quality).
11. No one data source provides all the information needed on applications under the Domestic Violence Act. Consequently in this report it has been necessary to repeatedly move between one data source and another. National information on applications comes from the Domestic Violence Act database while more information from the files in the pilot courts comes from the file study. Note that only limited information is available nationally. For example, reliable national information is available on the types of orders applied for but not on the type of orders made.
Footnotes
6. Including files with the Judge. Files at the High Court on appeal were specifically tracked down and selected.
7. Barwick, H. , Gray, A.& Macky, R.(1998) Domestic Violence Act 1995: Scoping Report Department for Courts and Ministry of Justice, Wellington
