Executive summary

The purpose of this audit is to identify, describe and disseminate information about Australian research projects (including program evaluations) undertaken during the 1995-2010 period on topics relating to the protection of children. This audit covers a broad scope of topics, ranging from universal preventative initiatives through to statutory child protection and out-of-home care. The scope of the topics for the audit reflects the six supporting outcomes outlined in the report Protecting Children is Everyone's Business: National Framework for Protecting Australia's Children 2009-2020 ("National Framework"; Council of Australian Governments [COAG], 2009b).

The specific aims of the audit, in keeping with the aims of two previous audits on similar topics (Cashmore & Ainsworth, 2004; Higgins, Adams, Bromfield, Richardson, & Aldana, 2005), were to identify:

• research projects (published and unpublished) undertaken in Australia between the 1995-2010 period on topics of relevance to the National Framework (COAG, 2009b);
• outcomes and progress since the Audit of Australian Out-of-Home Care Research (Cashmore & Ainsworth, 2004) and the National Audit of Australian Child Protection Research 1995-2004 (Higgins et al., 2005);
• gaps, duplication and areas for development in relation to the outcomes and national priorities identified in the National Framework; and
• priorities for future research and data collection on the basis of the audit results, outcomes of the Towards a National Agenda forum (October 2009) and priorities identified in the National Framework.

Building upon the methodology employed by Cashmore and Ainsworth (2004) and Higgins et al. (2005), this audit asked researchers working in relevant fields to submit information via an online audit form about projects they have been, or are currently, undertaking. However, in recognition of the limitations of this approach as a means of providing a comprehensive picture of the research field, and in order to ensure we captured as many projects as possible, we also conducted a literature search to identify research projects that met the audit criteria.

In keeping with the approach of the Higgins et al. (2005) audit, the following types of projects were excluded:

• literature reviews, unless they were systematic literature reviews or research audits;
• program descriptions where there were no evaluation components per se; and
• process evaluations or quality assurance procedures where the impact or outcomes of the program were not assessed.

Furthermore, in order to effectively manage the scope of the audit, detailed inclusion criteria were developed, based upon the key issues highlighted in the National Framework.

In total, 1,359 research projects that met the audit criteria were identified. Analysis of available information about these projects showed that just over half addressed issues specifically relating to child abuse and neglect and just over a quarter related to out-of-home care.

Overall, a much larger proportion of projects focused upon child abuse and neglect after it had occurred and statutory services (i.e., supporting outcome 4 of the National Framework), compared to the proportion that addressed issues relating to prevention and early intervention (i.e., supporting outcomes 1, 2 and 3).

Projects focusing upon sexual abuse were more common than projects that focused upon family violence (as a form of maltreatment), and much more common than projects that focused upon physical abuse, neglect and psychological maltreatment.

As Cashmore and Ainsworth found in 2004, foster care continues to be an area that receives more attention in research than other types of out-of-home care, such as kinship and residential care. Outcomes for children in out-of-home care is a common topic for out-of-home care research, while topics relating to permanency planning (e.g., reunification, adoption from care) constitute only a small proportion overall of this area of research.

Although the proportion of projects focusing upon issues relating to Indigenous Australians has increased over the past 16 years, very few of those projects focus upon issues such as Indigenous community-led solutions to child welfare issues or partnerships between Indigenous communities and government/services. There is some indication that out-of-home care research is more likely than child abuse and neglect research to address issues relating to Indigenous Australians. Very few projects focus on issues relating to disabled children, adolescents and families from a culturally and linguistically diverse background.

While detailed information about methodological approaches for these projects was not always available, there is some indication that research in this area continued to rely upon small sample sizes (i.e., less than 100) and was largely qualitative. Few projects utilised longitudinal data and there were very few projects employing a randomised control trial design. The challenges of utilising these methodologies within the child welfare field is likely to influence the extent to which they are employed; however, leading researchers and advocates in the field have consistently identified the critical need for this type of research.

Organisations conducting child protection projects (such as government departments, universities and non-government organisations) appear to be more likely to have worked alone than with other organisations, although the information available suggests that almost one-third of projects were conducted by more than one type of organisation. From the limited data available from respondents to the audit, it would appear that there continues to be low level of funding for research in this field.

The projects identified through this audit are now available via an online database. It is hoped that this online database, and the results of this audit, will provide researchers with a valuable tool to build upon the strengths of this broad research field and fill critical "knowledge gaps". The continued development of research within this field, along with the availability and continuous updating of the database, will ensure that practitioners and policy-makers have access to an evidence base from which they can plan and deliver services that improve outcomes for Australia's children.

Note: ^a Project numbers do not include those for which type or topic of project could not be identified. Projects may also fall into more than one category.

1. Objectives

The specific aims of the Protecting Australia's Children Research Audit (1995-2010) are to identify:

• research projects (published and unpublished) undertaken in Australia between the 1995-2010 period on topics of relevance to the report Protecting Children is Everyone's Business: National Framework for Protecting Australia's Children 2009-2020 ("National Framework"; Council of Australian Governments [COAG], 2009b);
• outcomes and progress since the Audit of Australian Out-of-Home Care Research (Cashmore & Ainsworth, 2004) and the National Audit of Australian Child Protection Research 1995-2004 (Higgins, Adams, Bromfield, Richardson, & Aldana, 2005);
• gaps, duplication and areas for development in relation to the outcomes and national priorities identified in the National Framework; and
• priorities for future research and data collection on the basis of the audit results, outcomes of the Towards a National Agenda forum (October 2009) and priorities identified in the National Framework.¹

In addition to this audit, an online resource that can be used for the purposes of a "rolling audit" has been produced. This resource includes information about all of the research projects identified through this audit and enables an ongoing monitoring process of research being undertaken in these fields.

An audit of research conducted in these areas is critical, considering the issues facing the child protection and out-of-home care (OOHC) sectors in Australia. These issues include increasing notifications of child abuse and neglect, and increasing numbers of children and young people being brought into and remaining in state care.

In order to make the best use of research resources, it is necessary to have a clear picture of the research that has already been conducted and the work that still needs to be done. Without this information, there is potential for wasteful duplication and gaps in knowledge.

Since 2004, four key projects have sought to provide a clear picture of Australian research about child abuse and neglect, child protection and out-of-home care. Taken together, these projects cover a period from January 1995 to October 2006. They are:

• Audit of Australian Out-of-Home Care Research (Cashmore & Ainsworth, 2004): an audit of Australian research about out-of-home care undertaken between 1995-2004;
• National Audit of Australian Child Protection Research 1995-2004 (Higgins et al., 2005): an audit of Australian research about child abuse prevention, the protection of children and the care of children subject to a child protection intervention undertaken between 1995-2004;
• Out-of-Home Care in Australia: Messages From Research (Bromfield, Higgins, Osborn, Panozzo, & Richardson, 2005): an in-depth review of out-of-home care research undertaken between 1995-2004 and a summary of key messages from research identified through the Cashmore and Ainsworth (2004) audit, and additional research identified; and
• Getting the Big Picture: A Synopsis and Critique of Australian Out-of-Home Care Research (Bromfield & Osborn, 2007): a review of Australian out-of-home care research identified through the Cashmore and Ainsworth (2004) audit, and research undertaken after the release of that audit (2005 to October 2006).

In addition to these projects, another piece of work, entitled A Systematic Map of the Out-of-Home Care Literature 1995-2008 (Mayfield, 2009), has identified Australian research undertaken in the field of out-of-home care.

The development of a national research agenda as part of the three-year action plan (COAG, 2009a) for the National Framework (COAG, 2009b) highlights the need for an up-to-date audit of research in these fields. In recognition of the contributing influence of universal prevention and early intervention strategies for reducing child abuse and neglect (as outlined in the National Framework), such an audit needs to include these topics.

The current audit encompasses the topics of child abuse and neglect, child protection, out-of-home care and the broader topics of universal prevention and early intervention. Together, these fields reflect the six supporting outcomes outlined in the National Framework. By documenting Australian research conducted over the past 16 years in these fields, this audit brings together a significant body of knowledge about protecting Australia's children. It documents areas of strength and concentrated research, and identifies areas where further research evidence is needed to support actions under the National Framework.

2. Methodology

The methodology for this audit drew upon those used in similar previous audits conducted in the fields of child abuse and neglect and child protection (Higgins et al., 2005) and out-of-home care (Cashmore & Ainsworth, 2004). In keeping with Higgins et al. and Cashmore and Ainsworth, this audit sought to identify the number of projects (not publications) that have been completed since 1995.

The first step was to identify the audit criteria (i.e., what should be included or excluded). The second step was to identify projects that met the inclusion criteria. This process consisted of two main methods: a literature search and collecting information (using an online audit form) about projects that researchers or organisations are currently undertaking or have completed since 1995. The third step was to integrate the findings of previous audits. All of these steps are described below.

The audit team undertaking these tasks consisted of staff from the Australian Institute of Family Studies (AIFS). Staff from the Social Policy Research Centre (SPRC) at the University of New South Wales contributed to the analysis of the findings from the audit.

2.1 Audit criteria

The first step for the audit was to decide upon the audit criteria; that is, which projects would be included and which excluded from the audit. There were two key questions that needed to be answered at this stage:

• What types of project will be included in the audit?
• What topics are relevant to the audit?

Both of these two key questions are explored below.

Project types

In order to remain consistent with previous audits, the definition that Cashmore and Ainsworth (2004) used to set the parameters for research was also used for this audit. They defined "research" as:

the systematic gathering of information involving data collection and analysis, using either original data or administrative data sets. (p. 9)²

In short, if a project involved the analysis (i.e., interpretation) of data, it was included in the audit, and therefore encompassed program evaluations.³ However, the following types of evaluation projects were excluded:

• program descriptions where there was no evaluation per se;
• process evaluation or quality assurance procedures where the impact or outcome of the program was not assessed.
• Literature reviews, unless they were systematic literature reviews or research audits, were also excluded.

Project topics

On the face of it, identifying the topics that might be relevant to the audit would appear to be relatively straightforward. Projects on the following topics were clearly relevant:

• child abuse and neglect;
• child protection (e.g., service systems, assessment, intervention); and
• out-of-home care.⁴

However, the above topics do not cover the entire scope of the National Framework, which is specifically based on the premise that in order to bring about a substantial and sustained reduction in child abuse and neglect in Australia over time, the vulnerability of families can be reduced through universal prevention and early intervention strategies. Hence, three of the six supporting outcomes (outcomes 1, 2 and 5) in the National Framework have a preventative focus.

A methodological challenge of this audit was that including all research that addressed some aspect of universal prevention and early intervention could have led to the identification of a very broad range of projects that said little about the protection of children. On the other hand, such research could be relevant because it could inform policy-makers about the broader issues that need to be considered in order to prevent child abuse and neglect.

The decision regarding which topics to include was guided by the initial literature searches, which provided an indication of which search terms were likely to produce a significant number of projects that were not relevant. For example, in supporting outcome 3, only those risk factors most commonly associated with child abuse and neglect (i.e., family violence, parental substance and abuse, and parental mental illness) were included unless, in the publication abstract, another risk factor (e.g., poverty, homelessness) was explicitly referred to in relation to child abuse and neglect.

Table 1 provides a detailed description of those topics deemed to have been relevant to the audit, and maps these topics against the supporting outcomes in the National Framework. The descriptions of the supporting outcomes outlined in the National Framework were used to develop this table (COAG, 2009b).

Notes: ^a The category of service availability/service access/service improvements was included because of the focus in the National Framework on "right services at the right time" (p. 17). For this specific inclusion category, all the risk factors identified in the National Framework were included. They are: poverty and social isolation, unstable family accommodation and homelessness, poor child and maternal health, childhood disability, mental health and behavioural problems, young people disconnected from their families, schools and communities, past experiences of trauma. ^b Research that specifically mentions risk factors for abuse and neglect was included (e.g., relationship between homelessness and abuse and neglect); however, research that addressed an issue that may be associated with abuse and neglect but did not specifically mention it as a risk factor was not included (e.g., families at risk of homelessness). ^c These risk factors were included because they are most commonly associated with the occurrence of child abuse and neglect, and identified in families involved with child protection services (COAG, 2009a). ^d Childhood mental health and/or behaviour problems are identified as a risk factor for abuse and neglect in the National Framework; however, as with other risk factors in the National Framework, only research that related children's mental health or behaviour difficulties to child abuse and neglect was included (see note ^b above).

2.2 Phase 1: Literature search

In order to identify projects that met the audit criteria, extensive searches of the literature were made, utilising the following two databases:

• Australian Family and Society Abstracts (AFSA): bibliographic database that indexes abstracts from published and unpublished material of relevance to Australian families;⁵ and
• Australian Research Online (ARO): online database that includes information from public research repositories of Australian universities and the Australian Digital Theses Program.

In order to identify projects from the literature search, a filtration process consisting of three key stages was followed:

1. A judgement was made, based upon the abstract, as to whether the research met the criteria (as outlined in section 2.1).
2. These records were then scrutinised more thoroughly and analysed according to whether the type of project (e.g., research and outcome/impact evaluation) and topic of project (as outlined in section 2.1) met the inclusion criteria.
3. In order to capture projects (rather than publications) the records were checked, and where multiple publications were identified as belonging to the same project, these records were "collapsed" into a single record.⁶

The literature search identified a total of 636 projects that met the inclusion criteria for the audit (see Appendix A for a full list of search terms used for the literature searches).⁷

This literature search did not incorporate publications or theses on the topics of child abuse and neglect and child protection published in the period 1995-2004, as this search had already been conducted as part of the Higgins et al. (2005) audit. We therefore undertook the same filtration process above on all the publications and theses identified by Higgins et al. (2005) and identified a total of 368 projects from among these.⁸ Where abstracts could not be found for some of these projects but the report/article title indicated the reference met the inclusion criteria for the audit, the reference was included; however, the potential for detailed analysis in these cases was limited.⁹

2.3 Phase 2: Online audit form

To capture information about other projects that fit the inclusion criteria, we developed a web-based survey (the "audit form"; see Appendix B). A similar method was utilised for the both the Higgins et al. (2005) and Cashmore and Ainsworth (2004) audits. This audit form was made available through the National Child Protection Clearinghouse website from mid-September 2010 to the beginning of November 2010.

The benefit of collecting information about projects directly from researchers and organisations ("stakeholders") through an online survey was that projects that had not been reported in published form and any published research not captured by AFSA and ARO could also be identified.

Inclusion criteria

It was expected that overly complex inclusion criteria (as outlined in Table 1) could potentially have discouraged potential participants. Therefore, in order to clearly highlight to stakeholders what topics of research were within the scope of the audit, the inclusion criteria were simplified and organised into eight themes. These themes reflect the first four supporting outcomes of the National Framework:¹⁰

• building safe and supportive communities for children and/or families (supporting outcome 1);
• support for vulnerable and at-risk families (supporting outcomes 2 and 3);
• the prevention of child abuse and neglect (including primary, secondary and tertiary prevention) (supporting outcomes 2 and 3);
• child abuse and neglect (including sexual abuse, physical abuse, psychological maltreatment, emotional abuse/neglect, physical neglect and witnessing family violence) (supporting outcome 4);
• child protection services and/or systems and/or practices (supporting outcome 4);
• other child protection issues (including but not limited to: reporting abuse and neglect, approaches to risk assessment, identification of risk factors for child maltreatment, treatment of adolescent or adult offenders) (supporting outcome 4);
• support for children who have been abused or neglected (supporting outcome 4); and
• out-of-home care (including leaving care) (supporting outcome 4).

By simplifying the audit criteria, however, it was possible that some entries received via the online audit form would not meet the inclusion criteria for the audit. Therefore, each audit form entry was manually checked against the inclusion/exclusion criteria before it was included.

In addition to asking stakeholders to identify the theme of their project, we also asked them to provide the following information:

• the project title;
• the year of project commencement, whether the project was completed and, if so, the year of project completion;
• the chief investigators on the project;
• the discipline(s) of the chief investigators (e.g., social work, psychology, law);
• the type of organisation(s) undertaking the project (e.g., university, government, non-government organisation [NGO]);
• whether the project was conducted as part of the requirements of a degree program (e.g., professional doctorate/PhD);
• the aim/purpose of the project;
• the project type (i.e., research and/or evaluation);
• the type of program evaluated (if it was an evaluation project);
• the research question/hypothesis;
• whether the project addressed the issue of child abuse and neglect and, if so, what type(s) of maltreatment were addressed;
• the major findings and implications of the project;
• the data sample (i.e., what sample was used and the size of the sample);
• funding amount and funding organisation; and
• products generated from the project (e.g., journal articles, conference papers, reports).¹¹

It is important to note that while every effort was made to clearly define terms in the audit form, some inconsistencies may exist between the way in which the audit form respondents categorised their projects and the way in which the audit team categorised projects identified through the literature search. For example, a number of audit form respondents identified their project as addressing the topic of the "prevention of child abuse and neglect", yet when asked whether their project addressed the issue of child abuse and neglect, they responded "No". When coding the projects identified through the literature search, however, all projects that mentioned child abuse and neglect in the title or abstract were coded as addressing child abuse and neglect.

For those audit responses where there were obvious coding anomalies, these were amended.¹² However, some potential anomalies (e.g., a project that does not mention child abuse and neglect in the aim/purpose, findings or implications section but is identified as addressing that issue) may not be as straightforward as they seem. In these cases, the information provided was taken as being correct.

A total of 267 projects were entered through the online audit form,¹³ 198 of which were identified by the audit team as meeting the inclusion criteria.

After the deadline for submission of projects via the online audit form had passed, we were alerted to a number of hitherto unidentified projects that may have met the inclusion criteria for the audit. However, as the analysis of the data had already been completed at that stage it was impossible to include those projects in the audit (see section 4.5). These projects can be added to the online database (see section 5).

Promoting the online audit form

The Higgins et al. (2005) audit involved an intensive process of contacting relevant networks, agencies, organisations and government departments in order to encourage people conducting research in the area to submit the details of their projects via the online audit questionnaire. Nevertheless, Higgins et al. noted that: "the number 135 in total of audit responses was lower than expected" (p. 7), and acknowledged that this is one of main limitations of relying upon a self-report audit form to gather information about research. In their review of out-of-home care research, Bromfield and Osborn (2007) also identified a number of publications via a literature search that had not been captured by the Cashmore and Ainsworth (2004) audit (which identified a total of 94 projects), suggesting that Cashmore and Ainsworth's methodology, which also relied partly upon self-report, also did not capture all the projects undertaken during the 1995-2004 period.

In recognition of the limitations of the 2004 and 2005 audit process, in the current audit we used the library search as the main method for identifying projects for the audit. The online audit form was used as a supplementary method for collecting information about unpublished research and research that was not identified through the literature search.

Nevertheless, an intensive process of promoting the audit questionnaire was undertaken for this audit in order to capture as much information as possible about projects undertaken during the 1995-2010 period. The various aspects of this promotion are outlined below.

Services and agencies

All of the child and welfare agencies approached for the Higgins et al. (2005) audit (30 in total) were contacted by phone and asked to distribute information about the audit questionnaire to all relevant staff within their organisation. They were also asked to disseminate information about the audit questionnaire to their own networks (e.g., affiliated researchers, other agencies etc).

Universities

The research offices of all universities across Australia (38 in total) were contacted by phone and asked to distribute information about the audit questionnaire to all relevant faculties and postgraduate student networks within their university.

An email about the audit questionnaire was also sent to all heads of school, deans of departments and faculties that were identified through the Higgins et al. (2005) audit (248 in total).

Governments

Representatives from each state/territory government were contacted by phone and/or email and informed of the research audit.

Other key contacts

Based on stakeholder lists and informal networks from the Australian Institute of Family Studies and the Social Policy Research Centre, other key people and organisations contacted to promote the audit and encouraged to contribute details about relevant research were:

• Australian Institute of Health and Welfare;
• Australian Institute of Criminology; and
• invitees to the Towards a National Research Agenda forum (69 in total; see footnote 1).

Other forums where the audit was promoted included:

• Australian Research Alliance for Children and Youth Update (October 2010);
• online Early Childhood News bulletin (EdNA);
• National Child Protection Clearinghouse e-alert subscriber list;
• National Child Protection Clearinghouse childprotect discussion list members;
• Communities and Families Clearinghouse (CAFCA) e-alert subscriber list;
• Closing the Gap Clearinghouse website;
• AIFS e-alert; and
• AIFS seminar on Safeguarding and Protecting Children Across the United Kingdom, presented by Dr Sharon Vincent (13 October 2010) (approximately 40 attendees).

These organisations or groups of researchers were asked to complete the online audit form and/or forward the information about the audit to any organisations or researchers working in the relevant fields and encourage them also to complete the audit form.

2.4 Integrating findings of previous projects

As discussed previously, the current audit covers the topics and time periods of four other key project reports (see section 1). All of these projects sought to identify and/or analyse the quality of Australian research on the topics of (a) child abuse and neglect and child protection; or (b) out-of-home care.

In addition to these projects, we also identified a fifth report that was a useful source of information on Australian research: A Systematic Map of the Out-of-Home Care Literature (Mayfield, 2009).

Mayfield's (2009) report identified available Australian and international literature on the topic of out-of-home care published during the 1995-2008 period. It also provided a "map" to illustrate which areas have a substantive evidence base and where there were gaps requiring further research. Only the Australian literature identified by Mayfield (2009) was included in the filtration process described in section 2.2.

The results of the Higgins et al. (2005) literature search were used to identify research relating to the topics of child abuse and neglect and child protection published in the period 1995-2004 (see section 2.2). In addition, we reviewed the reference lists of the other four reports to determine whether they met the inclusion criteria for the current audit. It was expected that by doing so the current audit would provide a more comprehensive picture of projects undertaken in the relevant fields.¹⁴

This process of review involved:

• cross-checking each publication in the reference list against the projects identified through the current audit's literature search and online audit form;
• identifying online abstracts for those publications that had not been captured via the literature search or audit questionnaire (where abstracts could not be found but the report/article title indicated the publication was relevant to the audit, the reference was included as a project; however, the potential for detailed analysis of items identified in this way is limited); and
• applying the three-step filtration process (see section 2.2) for the literature search.

A total of 58 projects that were identified from the four other key reports had not been captured via the literature search or audit form.

All of the projects collected through the Higgins et al. (2005) audit form (135 in total) were included in this audit.¹⁵ As detailed information about the projects collected for the Cashmore and Ainsworth (2004) audit (94 in total) was not available for the purposes of this audit, only projects identified in the Cashmore and Ainsworth reference list (58 publications in total) were checked against the inclusion criteria for the current audit.

2.5 Cross-checks between sources

Once all of the projects had been identified from the three key sources (library search, audit form and the five projects described in section 2.4), they were cross-checked to ensure that there was no duplication of projects. As discussed previously, although every effort was made to ensure individual publications were "matched" to projects, in some cases this was difficult, especially for publications where abstracts were not available.

3. Results

Table 2 provides a summary of the projects identified in this audit. Each item is discussed in more detail in the following sections.

Notes: ^a Project numbers do not include those for which type or topic of project could not be identified. Projects may also fall into more than one category.

3.1 Number of projects identified

A total of 1,359 projects were identified over the 16-year period (1995-2010). The number of projects completed in three roughly equal time periods (1995-2000, 2001-05 and 2006-10) is provided in Table 3.

This table shows that the completion/publication dates for projects collected for this audit were fairly evenly spread across the 16 years, with only slightly more in the 2001-05 period (32.2%), compared to the 1995-2000 and 2006-10 periods.

Notes: ^a Year of completion was either the year identified by the audit form respondents or that of the publication from which the project was identified. Percentages may not total 100.0% due to rounding.

3.2 Project sources

The source from which these projects were identified, are listed in Table 4. There are two important points to make about findings presented in this table.

Firstly, it is important to note that even with comprehensive library searches, cross-checks and the use of the audit form, this may be an underestimation of the total number of research and evaluation projects undertaken relating to protecting Australia's children during the 1995-2010 period. This is because not all projects have associated publications (and would, therefore, not have been captured via the library searches) and, although widely publicised, the audit form was not completed by all researchers/evaluators working in the area, either because they were not aware of it or did not have time to complete it.

Secondly, as discussed in the methodology section, although every effort was made to identify individual projects from the library searches, there may be some instances where projects have been listed more than once as a result of the difficulties associated with "matching" publications to projects.

Notes: ^a This number includes publications and theses identified in the Higgins et al. (2005) audit. Although Higgins et al. (2005) identified a total of 518 publications (353 publication and 165 theses), not all of these met the inclusion criteria for the current audit and some were "matched" to projects identified through other sources. ^b 135 projects were received via the audit form for the Higgins et al. (2005) audit; however, some were "matched" to projects identified through other sources. Percentages may not total 100.0 due to rounding.

3.3 Project type: Research or program evaluation

Projects were categorised according to whether they were a research, evaluation or research and evaluation project (see Text Box 2).¹⁶

A total of 1,207 projects (88.8% of the total number of projects) undertaken during the 1995-2010 period were identified as fitting into one of these categories or an "other" relevant category.¹⁷ The results are presented in Table 5.

The most common project type was research (79.5%), followed by evaluation (13.5%). Only a relatively small proportion were both research and evaluation projects (6.1%).

Projects identified through the audit forms (both 2005 and 2010) were much more likely to be evaluations compared to those identified through library searches. A total of 232 projects that identified project type (i.e., research and/or evaluation) were identified through the audit forms, and of those, 81 (34.9%) were either evaluation or research and evaluation projects.

This suggests that although formal publication of results may not be a high priority for organisations undertaking evaluation projects (as noted by Higgins et al., 2005), these organisations are still keen to disseminate information about those projects (as evidenced by their completion of the audit form).

3.4 Types of programs evaluated

In total, 235 evaluation projects (including both evaluation only and research and evaluation) undertaken during the 1995-2010 period were identified (17.3% of the total number of projects).

Table 6 shows the number of evaluation projects undertaken during each specific time period. This table shows that while there was a slight decrease in the proportion of evaluation projects between the 1995-2000 period and the 2001-05 period (14.4% and 13.9% respectively), there was a notable increase during the 2006-10 period (20.8%).

These evaluation projects were categorised according to what type of program was being evaluated (some programs fell into multiple categories, hence a single evaluation project could be coded for more than one program type). ¹⁸ In those cases where the type of program evaluated was not specified (n = 54), the program type was coded as "not specified". All the types of evaluation programs that Higgins et al. (2005) identified were included in this analysis, and a further four categories were added, namely: out-of-home care programs; collaborations, cross-sector partnerships and improving links between services; early intervention and prevention programs; and family violence support programs).

Table 7 shows the number of projects that evaluated a specific program type as a proportion of the total number of evaluation projects (not including 54 projects coded as "not specified"). In the evaluations, the most common program type being evaluated was family support programs (30.9%), followed by prevention programs for children (19.3%) and tertiary intervention programs for child victims (16.6%). The least common program types evaluated were collaborations, cross-sector partnerships and improving links between services (3.9%); early intervention and prevention programs (3.3%); and family violence support programs (2.8%).

Notes: ^a "Out-of-home care programs" include mentoring programs for young people leaving care, supporting and training for carers, support programs for parents whose children are in out-of-home care. ^b The category "other" includes those projects that belonged to a single subtopic (i.e., the project did not address more than one topic) with fewer than five projects. Examples of these subtopics include: evaluation of assessment tools (n = 2), implementation of legislation (n = 2), programs to support high-risk adolescents (n = 2), and school-based violence prevention programs (n = 1). Percentages total more than 100.0% as projects could evaluate more than one program type.

Table 8 shows the number of projects that evaluated a specific program type as a proportion of the total number of evaluation projects completed during each specific time period (including those projects where the completion year was not identified or the project was not yet completed).

Notes: ^a Year of completion was either the year identified by the audit form respondents or that of the publication from which the project was identified. Percentages total more than 100% as projects could evaluate more than one program type.

Excluding projects that were not yet completed, or for which the year of completion was not indicated, Table 6 provides an indication of changes over time. An analysis of these projects suggests that from 1995-2000 to 2006-10 there were increases in the both the actual numbers and the proportion of completed projects that evaluate programs that related to:

• family support (from 21.4% of projects to 37.7%);
• child abuse prevention (9.5% to 28.3%);
• parent education (from 2.4% to 11.3%);
• community education (from 2.4% to 9.4%);
• child-focused programs within adult-oriented services (from 2.4% to 9.4%); and
• community development (from 0% to 15.1%).

On the other hand, there were decreases between the 1995-2000 and 2006-10 periods in the proportion of completed evaluation projects that evaluated tertiary intervention programs (from 23.8% to 17.0%). However, in actual numbers of projects, it was a decrease of only one (from 10 in 1995-2000 to 9 in 2006-10). This suggests that while the total volume of child protection related research in Australia has increased over that time period, the evaluations of tertiary intervention for children remained stable, and that other comparatively neglected areas started to receive more attention. In addition, 10.0% of evaluation projects completed in the 2001-05 period were specialist or innovative child protection programs, compared to only 3.8% of evaluation projects completed in the 2006-10 period and none in 1995-2000. Moreover, the proportion of completed projects that evaluated out-of-home care programs increased in the 2001-05 period (from 9.5% to 24.0%), but decreased in the 2006-10 period (from 24.0% to 7.5%).

A comparison of the findings from this audit, for the 1995-2005 period, with the findings from the Higgins et al. (2005) audit (which looked at the 1995-2004 period) provides an insight into the changes over time in the focus of program evaluations.

Higgins et al. (2005) found that the three most common types of programs evaluated during the 1995-2004 period were prevention programs for children (21.7%), family support programs (16.9%), and tertiary intervention programs, parent education programs and statutory child protection services (each 12.4%). This audit similarly found that during the 1995-2005 period the three most common types of programs evaluated were family support programs (30.9%), prevention programs for children (19.3%) and tertiary intervention programs (16.6%).¹⁹

Although the order of most to least common of the three most common types of programs evaluated is different for this audit, when compared to Higgins et al. (2005), the most common types identified are similar, except when considering the fact that Higgins et al. (2005) found that the proportion of evaluations that were about tertiary intervention programs was equal to the proportion of projects that evaluated statutory child protection services or parent education programs (12.4%). In the current audit, we found that during the 1995-2005 period, the proportion of projects that evaluated tertiary intervention programs (14.5%; n = 17) was higher than the proportion that evaluated parent education programs (2.6%; n = 3) and statutory child protection services (2.6%; n = 3).

The difference in proportion of the programs evaluated relating to each type between this audit and Higgins et al. (2005) could be attributed to a number of factors, including the issue noted previously regarding the validity of categorisations assigned by some respondents to the 2005 audit form. As a result of including projects from which only an abstract was available, a larger number of projects was identified in the current audit (1,359), compared to the Higgins et al. (2005) audit; however, we were unable to categorise many of these according to the evaluation program type based on the information contained in the abstract. It may also be, however, that because Higgins et al. (2005) did not identify how many individual publications related to each project, this resulted in the identification of a larger number of project types (due to the potential for double counting).

3.5 Project topics

Project topics were analysed according to three categories:

• supporting outcomes under the National Framework;
• child abuse and neglect; and
• out-of-home care.

The results of this analysis are presented below.

Supporting outcomes under the National Framework

As discussed earlier, in order to clarify the scope of this audit we used the National Framework to identify which topics were relevant to each supporting outcome (see Table 1). We used the information in this table to then categorise all of the projects according to the supporting outcomes they addressed (projects could fall into more than one supporting outcome).²⁰ There was insufficient information available to categorise 120 (8.8%) of projects.²¹ The results for the remaining projects during the 1995-2010 period are presented in Table 9.

Notes: ^a See Table 1 for topics included in each supporting outcome. ^b The proportion of projects that addressed issues relating to supporting outcome 5 differs from the proportion that addressed issues relating to Indigenous Australians (see section 3.9). This is because supporting outcome 5 has a much narrower focus than those projects that address issues relating to Indigenous Australians (see footnote 49). ^c This includes all the projects that addressed sexual abuse specifically as a form of maltreatment (see Table 12), plus two other projects that addressed sexually abusive behaviours of child and young people but did not identify it explicitly as a form of child abuse. Percentages total more than 100% as projects could address more than one supporting outcome.

It is important, when reviewing these results, to take into account the breadth and scope of each supporting outcome. For example, only a comparatively small proportion of the total number of projects addressed supporting outcome 1 (Children live in safe and supportive families and communities), as the topics deemed relevant to this particular topic were fairly limited in scope (i.e., child abuse and neglect education; community education programs; and children's participation in decisions regarding child protection, juvenile justice, family court and out-of-home care). Therefore, while there may be many research projects that could be considered to relate to the fairly broad topic of the safety of children in communities and families; the scope of this topic, within the context of the National Framework, is likely to exclude many of these projects.

In comparison to supporting outcome 1, supporting outcome 4 incorporates a broad range of relevant topics (i.e., child abuse and neglect, child protection - including child protection services - and out-of-home care); hence, comparatively, a much larger proportion of projects were identified that relate to supporting outcome 4.

The results in Table 9 show the following:

• Comparatively few projects have investigated issues regarding community education about child abuse and neglect, parenting, child development and children's needs (supporting outcome 1 included all these topics and totalled only 2.7% of the total number of projects). This is supported by the data in Table 7 that showed that of all the evaluation projects identified in this audit, only 6.6% evaluated community education programs.
• The National Framework states that children's participation in decisions regarding child protection, juvenile justice, family court and out-of-home care is a "key signal of valuing and supporting children" (p. 15). Compared to research about most other supporting outcomes, the low proportion of projects addressing supporting outcome 1 suggests that this topic is also under-researched.
• A much higher proportion of projects (83.9%) have looked at issues relating to child abuse and neglect after it has occurred and/or out-of-home care (i.e., supporting outcome 4) rather than those factors relating to prevention and early intervention, and related issues such as children at risk, area-based disadvantage and service access for vulnerable/at-risk families (i.e., supporting outcome 2 and 3) (28.1% total). This reflects the nature of the child welfare service system that historically has also had a higher concentration of resources at the "pointy end" of the system (i.e., statutory services), compared to resources targeted towards vulnerable/at-risk families who fall below the threshold for statutory intervention (Bromfield & Holzer, 2008).
• Those projects that addressed service access for vulnerable/at-risk families (i.e., a subtopic of supporting outcome 2) tended to focus upon a range of issues relating to supporting vulnerable/at-risk families or programs, rather than focusing specifically upon service access.
• Most of the projects that addressed area-based disadvantage (i.e., a subtopic of supporting outcome 2) were associated with the former Stronger Families and Communities Strategy. The Stronger Families and Communities Strategy was a national initiative specifically designed to support children and families, with one strand (Communities for Children) focusing on children and families living in disadvantaged communities (Edwards et al., 2009; Muir et al., 2009).
• A sub-analysis of supporting outcome 3 projects shows that of those projects that looked at a specific risk factor (i.e., domestic/family violence, substance abuse and/or mental illness) (n = 76), 81.6% looked at domestic/family violence, while only 10.5% looked at substance abuse and 7.9% at mental illness.
• Comparatively, very few projects were related to addressing Indigenous disadvantage; community-wide strategies addressing alcohol misuse and family violence; community-led or community-identified solutions to protecting children; Indigenous children in out-of-home care maintaining connection to family, community and culture; and partnerships between government, service providers and Indigenous families/communities that relate to protecting children (supporting outcome 5, total: 1.3%). Higgins (2010) highlighted the importance of community-led solutions for preventing and responding to child abuse and maltreatment of Aboriginal and Torres Strait Islander children.
• Comparatively, the topics of child sexual abuse, the prevention of child sexual exploitation and support for child sexual abuse survivors (supporting outcome 6) is a fairly well researched topic (16.6%, the third most common topic). This is not surprising considering the fact that sexual abuse has received a significant amount of attention from researchers internationally (Higgins, 2004).

Child abuse or neglect

The topic of child abuse or neglect was addressed in 778 projects²² identified through the audit (57.2% of all projects identified).²³ Projects that addressed the issue of child abuse or neglect ranged from those that focused solely upon this topic (e.g., the prevalence of child abuse and neglect among a specific subgroup of the population) to those that referred to the issue of child abuse or neglect (e.g., prevention of child abuse and neglect, children in out-of-home care and educational outcomes that included a consideration of how trauma resulting from abuse and neglect may impact upon educational outcomes).

Table 10 shows the projects that addressed child abuse and neglect as a proportion of the total number of projects identified.

Notes: ^a Year of completion was either the year identified by the audit form respondents or that of the publication from which the project was identified.

The results in Table 10 show that:

• for completed projects where time of completion was specified, the highest proportion of projects addressing child abuse and neglect was during the 1995-2000 period (62.7%); and
• the proportion of projects addressing child abuse and neglect during the 2001-05, 2006-10 and for those projects not yet completed was relatively similar (52.1%, 53.1% and 53.7% respectively).

A qualitative analysis was conducted of the available information for all 778 projects that addressed the issue of child abuse and neglect in order to determine what topics they were addressing.²⁴ Of those 778 projects, there was insufficient information to identify subtopics for 118 of them.²⁵

The qualitative analysis included most of the topics used by Higgins et al. (2005) to sort projects.²⁶ A number of other categories that emerged frequently were also added to the list that Higgins et al. (2005) devised. A total of 29 topics were identified. The results of this analysis are presented in Table 11.

As outlined in Table 11, the three most common topics for projects addressing child abuse and neglect were: support and/or treatment for children who have experienced or are at risk of abuse and neglect and their families/parents/carers (17.6%);²⁷ workforce, practice and system issues (17.4%); and family violence  (14.8%).

There are several notable findings from this analysis:

• Projects addressing the topic of support and/or treatment for children who have experienced or are at risk of abuse and neglect and their families/parents/carers included a range of research and evaluation projects ranging from early intervention programs to therapeutic interventions for young people who had experienced abuse. As an extremely vulnerable group, it is encouraging to note that a significant proportion of research is addressing how to support these children and their families/parents/carers.
• However, client experiences of programs and services is a comparatively under-researched topic, suggesting that while priority has been given to researching support and treatment for children (17.6%), less priority has been given to evaluating how clients experience the programs designed to support them (3.9%).
• Projects addressing the topic of workforce, practice and system issues included a range of sectors that work with children (including child protection, child welfare and education), issues in practice at both the micro (e.g., parent-worker relationships) and macro levels (e.g., collaborative work practices between sectors), and broad systemic issues (e.g., research utilisation in child protection). Research on workforce, practice and system issues is important as these factors are critical to the protection of children. Child protection services have, in recent years, become overwhelmed by increasing demand (i.e., notifications); therefore improving approaches to workforce, practices and system issues could contribute to improving this situation (Higgins & Katz, 2008). Factors such as working in partnership and collaborating between sectors can improve responses to child protection (Higgins & Katz, 2008).
• A more extensive qualitative analysis of these specific projects could determine the proportion that address workforce issues that were identified as research priorities in the Towards a National Research Agenda forum (AIFS, 2009), such as technology in human service organisations; leadership in human service organisations; funding issues; and recruitment, retention, education, training and remuneration.
• Although projects relating to workforce, practice and system issues were common, just over 1 in 10 projects related more specifically only to statutory child protection work, that is, reporting abuse and neglect, identification of risk factors and approaches to risk assessment (11.1% in total).
• A significant proportion of projects addressed the topic of the impact of child abuse/neglect on children/young people (14.7%). The amount of research on this topic has been acknowledged as extensive (Lamont, 2010).
• While providing support and/or treatment of children who are risk of abuse and neglect could be seen as a form of prevention, it is important to note that only a small proportion of projects that addressed the issue of child abuse and neglect explicitly referred to prevention or early intervention programs/services (6.4%).
• Of all types of child maltreatment, sexual abuse has received the most attention in the research field internationally (Higgins, 2004). This trend is reflected in Australia, with enough projects to warrant the inclusion of the separate topic of perpetrators/offenders of child sexual abuse. There were not, however, enough projects addressing perpetrators/offenders of any other forms of maltreatment to warrant a separate category. A small proportion of projects addressed the issue of characteristics of abusive families/parents.

Notes: ^a The category "other" includes those projects that belonged to a single subtopic (i.e., the project did not address more than one topic) with fewer than five projects. Examples of these subtopics include: multi-type maltreatment (n = 4), relationship between poverty and maltreatment (n = 3), service access (n = 2) and violence against children in homelessness accommodation (n = 1). Percentages total more than 100% as projects could address more than one topic.

Maltreatment type

Those projects that addressed the issue of child abuse and neglect were coded according to the type or types of maltreatment they addressed (a project could address more than one type of maltreatment).²⁸ The results for the 1995-2010 period are presented in Table 12. This table shows the projects that addressed a specific maltreatment type, all maltreatment types, or those where maltreatment type was not specified as a proportion of the total number of projects (778) that addressed the issue of child abuse and neglect.

More than one-third of projects did not specify a maltreatment type. These projects included those that referred to child abuse and neglect in generic terms or where insufficient information was available to determine whether specific maltreatment types were focused on. Over the entire 1995-2010 period, the most common specific type of maltreatment addressed was sexual abuse (26.2%), followed by witnessing family violence (16.1%), other maltreatment types (e.g., medical child abuse, homicide, child death) (6.0%), physical abuse (5.8%), neglect (4.6%) and psychological maltreatment (3.6%).

Notes: ^a The category "witnessing family violence" includes projects where "family violence" (or a similar term) was explicitly mentioned. In some jurisdictions, witnessing family violence is categorised as psychological maltreatment. For this audit, projects categorised as addressing "witnessing family violence" were not categorised as addressing "psychological maltreatment". ^b The category "type not specified" includes those projects that used generic terms such as "child abuse or neglect" and "child maltreatment" but did not specify which types of child abuse or neglect were the focus, and those where insufficient information was available to identify maltreatment types. Projects that referred to a specific type of maltreatment and also used a generic term (e.g., "the project focuses on child abuse and neglect and how sexual abuse particularly impacts upon children") were coded as that specific maltreatment type only. A number of projects referred to child abuse and neglect and family violence (i.e., family violence as a separate category); these projects were coded as "family violence" only. ^c The category "all five maltreatment types" refers to projects that specifically stated that they addressed all types of abuse and neglect or those that individually listed all five maltreatment types. Percentages total more than 100.0% as projects could address more than one maltreatment type.

Of those projects that addressed sexual abuse, witnessing family violence, physical abuse, neglect or psychological maltreatment (n = 340), most (85.0%) focused solely upon one type of maltreatment. Of those projects that addressed sexual abuse (n = 204), a significant proportion (81.9%) focused only on sexual abuse and no other form of maltreatment.

These findings are consistent with previous research, which has shown that the topic of sexual abuse dominates child abuse and neglect research (Higgins, 2004).²⁹ Although there are some inconsistencies between these findings and the findings from the Higgins et al. (2005) audit, a qualitative analysis of the data collected for the 2005 audit shows that research on single maltreatment subtypes tended to focus upon child sexual abuse and family violence and that, "there is a lack of research investigating issues specifically associated with child physical abuse, psychological maltreatment or neglect" (p. 26).³⁰ The current audit shows that this has remained true when looking at the entire 1995-2010 period (though see Table 13 for a more detailed analysis of changes over time).

Although this audit identified witnessing family violence as the second most common specific form of maltreatment addressed, it is important to note that as part of the search strategy for this audit, all projects that addressed the topic of family violence and children and/or parenting were included, regardless of whether the project specifically focused on the effect of witnessing family violence on children, or its relationship to child abuse and neglect (see inclusion criteria for supporting outcome 2 in Table 1). In other words, the number of projects addressing family violence is indicative of the number of projects that focus upon the topic of family violence and children and/or parenting but not indicative of the number of projects that address family violence specifically as a form of child maltreatment in itself or as a risk factor for other forms of abuse/neglect.

A significant proportion of projects did not specify maltreatment type. This was because insufficient information was available, or "abuse" or "maltreatment" was referred to in a generic sense. Table 13 outlines changes over the 1995-2010 period in research conducted on the maltreatment types. Changes over time are represented according to the year in which the project was completed.

Notes: ^a Year project completed was either the year identified by audit form respondents or year of the publication from which the project was identified. ^b The category "witnessing family violence" includes projects where "family violence" (or a similar term) was explicitly mentioned. In some jurisdictions, witnessing family violence is categorised as psychological maltreatment. For this audit, projects categorised as addressing "witnessing family violence" were not categorised as addressing "psychological maltreatment". ^c The category "type not specified" includes those projects that used generic terms such as "child abuse or neglect" and "child maltreatment", but did not specify which types of child abuse or neglect were the focus and those were insufficient information was available in the abstract or the abstract was not available to identify maltreatment types. Projects that referred to a specific type of maltreatment and also used a generic term (e.g., "the project focuses on child abuse and neglect and how sexual abuse particularly impacts upon children") were coded as that maltreatment type only. ^d The category "all five maltreatment types" refers to projects that specifically stated that they addressed all types of abuse and neglect or those that individually listed all five maltreatment types. Percentages total more than 100% as projects could address more than one maltreatment type. For the total number of projects addressing child abuse and neglect during these periods, see Table 10.

The findings outlined in Table 13 illustrate the following:

• The proportion of projects within each of the three periods that were focused on sexual abuse decreased over time, from 30.3% of the total number of projects completed in 1995-2000 to 29.4% in 2001-05 and 24.6% in 2006-10. This suggests that compared to other forms of maltreatment, the focus upon sexual abuse in research has declined over the past 16 years.
• The proportion of projects that focused upon witnessing family violence increased between the 1995-2000 and 2006-10 periods (from 13.9% to 18.4% respectively).
• Only a comparatively small number of projects (n = 45) referred specifically to physical abuse. However, over the three time periods there appears to have been a slight increase in the proportion of projects that focused on physical abuse (2.9%, 5.7% and 8.2% respectively).
• Only a small number of projects referred specifically to neglect (n = 36); however, over the three time periods there appears to have been a slight increase in the proportion of projects that focused on neglect (2.0%, 4.4% and 6.3% respectively).
• Only a small number of projects referred specifically to psychological maltreatment (n = 28); however, between the 1995-2000 and 2001-05 periods there appears to have been a slight increase in the proportion of projects that focused on psychological maltreatment (1.2% and 3.9% respectively, remaining at 3.9% in the 2006-10 period).
• Between the 1995-2000 and 2006-10 periods, there was a notable increase in the proportion of projects that addressed all five maltreatment types (7.0% to 21.3% respectively). This may reflect a move towards a more sophisticated understanding of the field of child abuse or neglect.

Out-of-home care

Out-of-home care was a topic addressed in 371 projects identified for this audit (27.3% of the total number of projects identified).³¹ Relying on personal responses from researchers/organisations, a search of key reports and documents, and a search of funding databases, Cashmore and Ainsworth (2004) identified 94 projects undertaken during the 1995-2004 period. The methodology for the current audit, using literatures searches supplemented by an audit form, identified 172 projects during that same period (i.e., 1995-2004).

Table 14 shows the projects that addressed out-of-home care as a proportion of the total number of projects completed during those time periods.

Notes: ^a Year project completed was either the year identified by audit form respondents or year of the publication from which the project was identified.

The findings suggest that between the 1995-2000 and 2001-05 periods, there was a notable increase in the number and proportion of projects addressing the issue of out-of-home care (from 17.7% to 29.9% respectively). This increased again during the 2006-10 period (to 34.6%). Overall, the findings suggest that during the past 16 years there has been an increase in the focus by researchers on out-of-home care.

Although an increase in the proportion of projects about out-of-home care could be a positive phenomenon for the protection of Australia's children (reflecting an increased interest and an increased body of knowledge), Cashmore and Ainsworth (2004) noted that there was a need for large-scale, multiple-site cross-state projects (p. 11), as a significant proportion of projects they identified were based on "small non-random samples, often in single agency services" (p. 10). If a larger proportion of projects about out-of-home care is attributed to a larger number of such small-scale research projects, this is not necessarily a positive shift.

A qualitative analysis was conducted of all 371 projects that addressed the issue of out-of-home care in order to determine what topics they were addressing. All the categories that were used to sort projects were drawn from Mayfield's (2009) systematic map of out-of-home care projects. Mayfield (2009) categorised research according to 17 content areas, 10 of which were covered by Cashmore and Ainsworth (2004). Seven other themes that were identified during the coding process were added to the topic classification framework: out-of-home care statistics and trends, abused children in care, parents/families of children in out-of-home care, emotional and/or behavioural difficulties of children/young people in out-of-home care, abuse of children/young people in out-of-home care, analysis/comparison/review of policies relating to out-of-home care; and meeting the needs of children/young people in out-of-home care. Of the 371 projects, there was insufficient information to identify subtopics for 14 (3.8%) of them.³² The results of the analysis are presented in Table 15.

Notable findings from Table 15 include:

• For projects that addressed the issue of out-of-home care, the most frequent topics were: outcomes for children and young people in care (23.2%), foster care (21.7%) and out-of-home care statistics and trends (10.1%).
• Studies that included outcomes for children and young people in care were the most frequent type of out-of-home care research. This focus is a promising sign, given that Cashmore and Ainsworth (2004) identified outcome studies as an urgent research priority. However, few studies undertook primary research on child outcomes of the kind recommended by Cashmore and Ainsworth (2004) - specifically, large-scale quantitative studies across multiple jurisdictions. Only two studies were longitudinal.
• Studies on foster care were the second most numerous type. These include studies on the recruitment and retention of foster carers, support and training for carers, the impact of fostering on different family members, and the experiences of Indigenous and culturally and linguistically diverse (CALD) carers. A small number of these projects focused upon options for improving the recruitment and retention of carers.
• In 2004, Cashmore and Ainsworth noted that research about out-of-home care has an "almost exclusive focus on foster care" (p. 24). The current audit, which captured a larger number of projects than the Cashmore and Ainsworth audit, supports their findings to the extent that there is a strong focus upon foster care (21.8% of out-of-home care research), especially when compared to residential and kinship care (7.3% and 6.4% respectively).
• Kinship care is the fastest growing form of out-of-home care in Australia and internationally (McHugh, 2009), but the relative lack of Australian research on both statutory and informal kinship care has been identified as a gap (McHugh, 2009; McHugh & Valentine, 2010). For this audit, kinship care was a central theme of 23 of the studies, relating to both formal/statutory care and informal care. It is important to note, however, that most of the projects categorised as "outcomes for children and young people in care", and all of the large-scale quantitative studies, relate only to statutory care. The Australian Institute of Health and Welfare (AIHW) annual reports on trends in out-of-home care (part of the Child Protection Australia series) are also limited to statutory out-of-home care (i.e., not including informal kinship care) (AIHW, 2011).
• Thirty-six studies include out-of-home care statistics and trends as a theme. Of these, 13 are the annual AIHW and individual state reports on child protection statistics.³³ The remainder are mostly analyses of data collection systems and/or reviews of the reasons behind the increase in number of children living in out-of-home care and ways in which this increase could be arrested or reversed. Three studies include a history of out-of-home care systems and substitute care provision in Australia, as a means of providing contextual information on these trends.
• The participation of children or young people in decisions affecting their lives is one of the topics identified as a key issue in the National Framework (as part of supporting outcome 1). Within the field of out-of-home care research, this topic has been addressed in 29 projects. The majority of these projects (n = 13) were undertaken during the 2001-05 period.
• In addition to research about kinship care, another possible research priority identified by Cashmore and Ainsworth (2004) was permanency planning. A small number of projects addressed issues relating to reunification and permanency planning (n = 15) and a very small number addressed adoption from care (n = 3). Overall, these two topics constituted 4.9% of the total number of out-of-home-care projects. Half of the 18 projects that addressed reunification and permanency planning or adoption from care were completed before 2005 (9 of the 18), suggesting that even with Cashmore and Ainsworth's call for further research in this area, research activity on this topic has not increased significantly over the past six years.
• A further qualitative analysis of these projects could determine the extent to which another of Cashmore and Ainsworth's (2004) possible research priorities has been addressed by research; that is, engaging children and young people in education.³⁴

Notes: ^a The category "other" includes those projects that belonged to a subtopic with fewer than five projects. Examples of these subtopics include: adoption from care (n = 3), therapeutic foster care (n = 3), research audits (n = 3) and preventing children from entering out-of-home care (n = 2). Percentages total more than 100.0% as projects could address more than one topic.

3.6 Research infrastructure and funding

Discipline area

The discipline area of the chief investigators was identified in a total of 315 projects (23.2% of the total number of projects).³⁵ The results are presented in Table 16.

The findings outlined in Table 16 illustrate that, for those projects where the discipline area of the chief investigators was identified, the most common discipline area was social work (43.2%), followed by psychology (30.8%), social policy (23.8%), other (13.0%), sociology (10.5%), education (8.9%), health (8.3%) and law (3.8%).

Note: Percentages total more than 100% as chief investigators could have more than one discipline area.

Higgins et al. (2005) also found that social work was the most common discipline area of researchers. However, in comparison to the Higgins et al. audit, the current audit found a higher proportion of researchers in the area of psychology compared to social policy, and a higher proportion of researchers in the area of sociology compared to education. In keeping with Higgins et al., the current audit found that the least common discipline area of researchers was law.³⁶

Table 17 outlines changes over time in the discipline areas of the chief investigators, where identified. The findings illustrate the following:

• Between the 1995-2000 and the 2006-10 period, the proportion of projects undertaken by researchers from the discipline of social work increased substantially (from 28.6% to 53.8%).
• The proportion of projects undertaken by chief investigators from the following discipline areas has increased gradually over the three time periods:
  • sociology (2.4%, 8.2% and 14.0% respectively); and
  • education (4.8%, 6.1% and 10.8% respectively).
• Although the actual number of projects undertaken by chief investigators from the discipline of psychology has increased over the three time periods, the proportion of projects has gradually decreased (35.7%, 32.7% and 31.2% respectively), due to an increase in activity in other areas, particularly social work and social policy.

These findings regarding changes over time in the discipline area of chief investigators need to be considered in light of the fact that this information was available for less than one-quarter of the total number of projects (23.2%). The findings may have been different had the discipline area of the chief investigators of all the projects been available; however, this information was generally not available in the abstracts of publications, therefore it was difficult to obtain more representative data.

Notes: ^a Year project completed was either the year identified by audit form respondents or year of the publication from which the project was identified. Percentages total more than 100% as chief investigators could have more than one discipline area.

Types of organisations conducting projects

The types of organisation conducting the projects were identified for a total of 361 projects (26.6% of the total number of projects).³⁷ The frequency of each organisation type, as a proportion of the total number of projects where organisation type was identified, is outlined in Table 18.

Note: Percentages total more than 100% as projects could be conducted by more than one type of organisation.

The findings outlined in Table 18 show that of those projects where organisation type was identified, the most frequent organisation type was university (48.5%), followed by government (48.2%), non-government organisations (32.7%), research institutes (8.0%) and "other" organisations, such as consumer organisations, private companies and hospitals (2.8%). The relatively small proportion of projects undertaken by research institutes (8.0%) is likely to be a consequence of the comparatively smaller number of research institutes when compared to, for example, universities and non-government organisations.

Table 19 illustrates changes over time for the types of organisations conducting projects over the 1995-2010 period. Changes over time are represented according to the year the project was completed.

Notes: ^a Year project complete was either the year identified by audit form respondents or year of the publication from which the project was identified. Percentages total more than 100% as projects could be conducted by more than one type of organisation.

The findings outlined in Table 19 illustrate the following:

• Between 2001-05 and 2006-10, there was an increase in the proportion of projects completed by universities (34.4% and 47.9% respectively), but the proportion of projects completed by government decreased slightly (49.2% and 45.7% respectively).
• Between the 1995-2000 and 2006-10 periods, the proportion of projects completed by non-government organisations increased (32.4%, and 40.7% respectively).

As with the findings regarding the discipline area(s) of chief investigators, these findings regarding the type of organisations conducting projects need to be considered in light of the fact that this information was only available for just over one-quarter of the total number of projects (26.6%). The findings may have been different if information regarding the type of organisations conducting projects was available for all projects; however, this information is generally not available in the abstracts of publications, therefore it is difficult to obtain more representative data. There was also a limited response to the audit tool, which identified only 297 projects out of the total of 1,359 (including projects identified through literature searches).

Cross-organisational projects

Table 20 shows the number of cross-organisational projects (i.e., projects that were undertaken by more than one organisation type) and the number of single organisational projects (i.e., those projects conducted by one organisation only) as a proportion of each organisational type.³⁸

Note: Percentages total more than 100% as more than two types of organisations could conduct a project.

The results outlined in Table 20 show that:

• the most common cross-organisational project type for government was government and university (32.8%);
• the most common cross-organisational project type for universities was university and non-government organisation (44.0%), followed by university and government (32.8%); and
• non-government organisations were most commonly undertaking cross-organisational projects with universities (44.0%).

Level of funding

Audit form respondents for the current audit were asked to provide details of the level of funding they had received and the type of agency that had provided that funding for their project. The funding amounts for the 80 projects (5.9%) with available information are included in Table 21.³⁹ The table also shows the range, mean, median level of funding for these 80 projects.

The total amount of funding was $10,120,183; however, it should be noted that this amount is an underestimation of the full extent of research expenditure on projects that address the topic of protecting Australia's children as the figure is only based on 5.9% of the total number of projects identified from this audit.

Table 22 provides a comparison of funding levels between those reported in the Cashmore and Ainsworth (2004), Higgins et al. (2005) and current audits. Keeping in mind the caveat discussed above, these data suggest that the average amount of funding for projects relating to the broad topic of protecting Australia's children was higher across the entire 1995-2010 period than the average amount for child protection and child abuse and neglect research in the 1995-2004 period, and much higher than the average amount for out-of-home care research for the 1995-2004 period. Note though that the current audit identified four projects with funding of $700,000 or more, which increased the average amount overall.

The median level of funding also increased when compared to the two previous audits. Nevertheless, the difference between the mean and the median for the current audit indicates that although a small number of projects were receiving a significant amount of funding (i.e., $700,000 or more), half were undertaken with less than $65,500. The relatively low median level of funding may provide a partial explanation for why there is an abundance of small, qualitative studies in this field. In other words, large-scale, longitudinal studies require significant resources when compared to small-scale, qualitative research projects.

Notes: ^a Cashmore and Ainsworth (2004) did not specify the average amount of funding. This amount was calculated by dividing the total amount of funding identified by the total number of projects identified (n = 94). The funding levels for three of the projects identified by Cashmore and Ainsworth were not available; however, an estimate for these three projects was made, based upon the average amount of funding for the other 91 projects. ^b For both the Higgins et al. (2005) and the current audit, not all projects identified provided information about project funding.

Source of funding

It was possible to identify the source of funding for 200 projects (14.7% of all projects) (Table 23).⁴⁰

Notes: Percentages total more than 100% as more than one type of organisation could fund a project.

The most common type of organisation funding projects was state/territory governments (29.5%), followed by the Commonwealth government (26.5%), universities (23.0%), self-funding (18.5%), other (13.5%) and philanthropic organisations (10.5%). The significant proportion of projects funded by state/territory and/or Commonwealth governments is not surprising considering their extensive responsibility for the protection of children.

It is interesting to note that almost one in five projects were self-funded. The proportion of projects that included some level of self-funding almost doubled the proportion that were funded by philanthropic organisations (18.5% and 10.5% respectively).

The Higgins et al. (2005) audit, which covered the 1995-2004 period, identified a much larger proportion of projects funded by the Commonwealth government when compared to state/local government (26% and 18% respectively). In contrast, the current audit, covering the 1995-2010 period, identified a similar proportion of projects funded by state/territory and Commonwealth governments (29.5% and 26.5% respectively). If there had been an increase in the proportion of state/territory government funded projects during the 2006-10 period, compared to Commonwealth funded projects during this period, this may have explained the discrepancy between the findings of the current audit and the Higgins et al. (2005) audit. However, the information regarding changes over time for funding sources (see Table 24) suggests that the proportion of both state/territory government funded projects and Commonwealth funded projects increased significantly during the 2006-10 period; hence it is most likely that this particular discrepancy relates to the differing methodologies of the two audits.

Cashmore and Ainsworth (2004) identified Australian Research Council funding specifically (categorised for the purposes of the current audit as "Commonwealth research granting agency") and found an equal proportion of funding for ARC and state government funding. In as far as ARC funding represents funding from a Commonwealth research granting agency, Cashmore and Ainsworth's findings are similar to those of the current audit.

Table 24 illustrates changes over time for sources of funding over the 1995-2010 period. Changes over time are represented according to the year the project was completed.

The findings from Table 24 show:

• The number of projects funded by all funding sources except philanthropic organisations increased significantly between the 1995-2000 and 2006-10 periods.
• The number of projects funded by philanthropic organisations remained steady over the three time periods (between 3-4 projects).
• In the 2006-10 period, the number of projects funded by all funding sources was relatively similar (state/territory government funded = 17 projects; Commonwealth government = 20; universities = 14; self-funded projects =  16; and other funding source = 16), while the number of project funded by philanthropic organisations was significantly lower (4 projects only).

Notes: ^a Year project completed was either the year identified by audit form respondents or year of the publication from which the project was identified. ^b During this time period no projects had more than one source of funding. Percentages total more than 100% as projects (except in 1995-2000) could have more than one source of funding.

As with some of the other findings discussed above, it is important to note that this information regarding funding sources needs to be considered in light of the fact that data were only available for a small proportion of projects, in this case 14.7%.

Projects conducted as part of a degree program

In total, 253 projects (18.6% of all projects) were identified as having been completed as part of a university degree program, of which 236 indicated the type of degree program involved (Table 25).

The most common type of degree being completed by researchers who were undertaking projects as part of a degree program was professional doctorates/PhDs (47.4%), followed by Masters degrees (30.8%) and Honours degrees (13.8%).⁴¹ This is in keeping with the findings of the Higgins et al. (2005) audit, where professional doctorates/PhDs were also the most common category.

3.7 Project methodologies

Qualitative and/or quantitative methodology

The audit team categorised the projects according to whether they utilised a qualitative and/or quantitative methodology. Due to time constraints, only those projects that explicitly stated which type of methodology was utilised were coded. The audit team identified 112 projects (8.2% of all projects) that explicitly stated which type of methodology was utilised. Of those 112 projects, the majority used a qualitative methodology, just over one quarter utilised both qualitative and quantitative methodologies and just over one in five used a quantitative methodology only (21.4%).

The comparatively high proportion of qualitative studies is consistent with Cashmore and Ainsworth's (2004) audit, which found that 52% of the research projects they identified were qualitative.⁴² It is also consistent with the findings of the out-of-home care research studies by Bromfield & Osborn (2007) and Bromfield et al. (2005). Both reports noted an over-reliance on qualitative techniques among the out-of-home care projects they identified and that, as a consequence, the "evidence base is rich in detail, but has limited generalisability" (p. 104).

It is important to note that as only a very small proportion of projects (8.2%, n = 112) had the relevant information available for determining the type of methodology used, these findings should not be viewed as being representative of the entire field of research that relates to the project of Australia's children. In terms of actual proportions, however, the findings do support the conclusions of similar projects that analysed the types of methodologies used within this research field (e.g., Cashmore & Ainsworth, 2004; Bromfield et al., 2005; Bromfield & Osborn, 2007).

Randomised control trials and longitudinal studies

Of the 1,359 projects identified in this audit, only four explicitly stated the use of a randomised control trial design. All four projects were completed during or after 2000; two were by the same author; and two were PhD projects. The topics of those randomised control trials were:

• a social support intervention on the health of mothers in the year after birth (Bogossian, 2003);
• home visiting as an early strategy for the prevention of abuse and neglect (Fraser, 2000);
• a home visiting intervention for vulnerable families with newborns (Fraser et al., 2000); and
• a treatment program for sexually abused children with posttraumatic stress disorder (King, 2000).

It has long been argued that it is difficult to introduce randomised control trials for interventions in the child welfare and child protection system due to ethical and practical dilemmas (Bromfield & Arney, 2008; Lamont, 2009), which may reflect why so few studies used this method.

The audit also found that 20 projects were longitudinal studies that collected data over a significant period of time, and three of these are ongoing. This is an encouraging number, given the extensive costs and resources involved in tracking participant developments over an extensive period of time.  It is encouraging also to note the commencement in 2011 of the NSW Pathways of Care study of children and young people in out-of-home care. The aim of the study is to provide the knowledge needed to strengthen the out-of-home-care service system in NSW in order to improve the outcomes for children and young people in out-of-home care.  These outcomes include children's and young people's permanency, safety and wellbeing (including their physical health, socio-emotional and cognitive/learning development).

Longitudinal studies are important for observing changes in participants over several years rather than just at one point in time.

It is interesting to note that of the 20 longitudinal projects identified, 11 focus on out-of-home care. Therefore, while they only constitute 27.3% of all projects, out-of-home care research accounts for more than half of the longitudinal studies identified in the current audit.

Sample type

In 443 projects (32.6% of all projects), there was sufficient information to identify the sample type (i.e., the source of data). Table 26 indicates the frequency of each sample type (one project could have more than one sample type) as a proportion of the total number of projects where sample type was indicated.

Notes: ^a This included people who had contacted a foster care agency, ethnic community leaders, adults who had been in care as children. Percentages total more than 100% as more than one sample type could be included in a project.

Notable findings from Table 26 include:

• The three most common sample types were professionals (40.4%), followed by children (31.4%) and parents (23.5%).
• Projects that included a sample of out-of-home carers were more likely to examine foster carers than kinship carers (10.8% and 7.4% respectively).

Of the 182 projects that used a sample of professionals, 166 (91.2%) indicated the type of professionals that were included in the sample. The results are outlined in Table 27.

Note: Percentages total more than 100% as more than one type of professional could be included in the sample type for a single project.

The three most common types of professionals were child protection workers (49.4%), followed by welfare workers (47.6%) and health professionals (26.5%).

The Higgins et al. (2005) audit found a slightly higher proportion of projects with welfare/family support workers as the sample type (23%), compared to child protection workers (21%). Apart from these two specific professional types, Higgins et al. (2005) did not identify other professional types, therefore it is not possible to compare the findings of the two audits relating to types of professionals.

Of those projects that used a sample of parents, 87 indicated the "status" of those parents. The results are outlined in Table 28.

Notes: Percentages total more than 100% as more than one "type" of parent could be included in the sample type for a single project.

As outlined in Table 28, the most common status of parents in the parent sample were those at risk of abusing or neglecting their children or who had abused or neglected their children (37.9%), followed by other parents (33.3%), those parents whose children were in out-of-home care (29.9%), and parents from the general community (23.0%).

The Higgins et al. (2005) audit found a higher proportion of projects using a "community sample of parents" as the data source when compared to a "high-risk" individuals/families. However, the current audit found the opposite; that is, there were a higher proportion of projects using a sample of parents who were at risk of abusing or neglecting their children or who had abused or neglected their children.

Sample size

For 288 projects (21.2% of all projects), there was sufficient information to identify sample size. Table 29 illustrates the frequency of the sample sizes, from very small (1-19) to large (500+).

These findings regarding sample size are very similar to the findings of the Higgins et al. (2005) audit; with the largest proportion of projects having a small sample size (36.5%), followed by medium (28.8%), very small (20.1%) and large (14.6%) sample sizes.

Table 30 illustrates changes over time for the sample sizes for the 288 projects where sample size was identified.

The findings outlined in Table 30 illustrate the following:

• Between the 1995-2000 and 2006-10 periods, although there was an increase in the overall number of projects with a very small sample size (from 8 to 21), there was a decrease in the proportion of those projects (from 22.2% to 17.5%).
• Between the 1995-2000 and 2006-10 periods, although there was an increase in the overall number of projects with a small sample size (from 15 to 44), there was a decrease in the proportion of those projects (from 41.7% to 36.7%).
• The number of projects with a medium sample size increased (from 11 to 21), although the proportion of those projects remained very consistent over time (30.6%, 30.4% and 28.3% respectively).
• Both the number and proportion of projects using a large sample size increased over the three time periods (from 2 projects (5.6%) to 7 projects (10.1%) to 21 projects (17.5%).
• Despite a tendency towards small projects, there was a good variety of sample sizes, including those projects with a large (i.e., 500 samples or greater) sample size. These data are based, however, on the small number of projects with available information about sample size.

Notes: ^a Year project completed was either the year identified by audit form respondents or year of the publication from which the project was identified. Percentage may not total 100.0% due to rounding.

3.8 Productivity/output

A total of 1,181 projects (86.9%) were recorded as having produced at least one "output"; that is, a report, journal article, conference paper, thesis, book, book chapter or other product (e.g., newsletter article). This suggests that in most cases researchers in this field undertook some form of dissemination.⁴³

Table 31 shows the total number of outputs for each output type and the number of projects that produced one of those output types. The table shows that of those projects that produced at least one output, the most common output was a report (29.7%), followed by a journal article (28.0%), a conference paper (18.8%), a thesis (11.7%), a book or book chapter (4.3%) and other (3.2%).

The information presented in Table 31 provides an indication of the extent of dissemination of project research. Information regarding nearly one-third of projects was disseminated via a report, although some of those reports may not be publicly available. In addition, information regarding more than a quarter of the projects was disseminated via a journal article. Almost one in five presented information relating to their research at a conference.

3.9 Issues relating to Indigenous Australians

The audit identified 214 projects (15.7% of all projects) that addressed issues relating to Indigenous Australians, which range from projects that focus solely upon Indigenous Australians to those that discuss the implications of the research for Indigenous populations.⁴⁴ Table 32 shows the change over time in the projects that addressed Indigenous Australians as a proportion of the total number of projects identified for this audit.

Note: ^a Year project completed was either the year identified by audit form respondents or year of the publication from which the project was identified.

These findings suggest that over time there has been an very notable increase in the proportion of projects relating to the protection of Indigenous children, from 5.1% of all projects completed in the 1995-2000 period to 26.2% in the 2006-10 period (more than a five-fold increase).⁴⁵

In 2004, Cashmore and Ainsworth noted a lack of out-of-home care research that addresses Indigenous children. Of the 371 projects that addressed out-of-home care, 89 (24.0%) addressed issues relating to Indigenous Australians. Of the 778 projects that addressed child abuse and neglect, 104 (13.4%) addressed issues relating to Indigenous Australians. This data suggest that issues relating to Indigenous Australians are more commonly addressed in research relating to out-of-home care rather than that relating to child abuse and neglect.

3.10 Other population groups

In order to determine the extent to which research is addressing the diversity of the Australian population, the audit team coded projects according to whether they referred to any of the following population groups:

• children with a disability;
• families from CALD backgrounds; and
• adolescents.

The results are presented in Table 33.

As outlined in Table 33 fewer than one in ten projects (8.6%) referred to one of these population groups, suggesting that the extent to which research relating to protecting Australia's children addresses the diversity of Australia's population is limited.

Considering the fact that children with a disability have a higher risk of being abused or neglected than other children (Sullivan & Knutson, 2000; Westat, 1993), it would seem that there is a need for more research on this specific population group. In 2004, Cashmore and Ainsworth identified research about this particular population group as a content gap in the field of out-of-home care research (p. 24).

There is limited research regarding cultural issues within statutory child protection services in Australia, including the needs of children and families from CALD backgrounds within child protection services (Sawrikar, 2009), and the assessment of families from CALD backgrounds (Kaur, 2007). Considering these issues, a focus upon this specific population group would appear to be justified. Cashmore and Ainsworth (2004) also identified research about this population group as a content gap (p. 24).

Adolescents are an important subgroup to consider in this context, especially in relation to child protection services and out-of-home care. Adolescents in the child protection system may have experienced repeated exposure to trauma and, for many, the trauma of maltreatment is compounded by experiences such as the death of a parent or separation from siblings (Robinson & Miller, 2010). Only a small proportion of projects addressed issues relating to adolescents (3.6%).

4. Conclusion

This audit identified a total of 1,359 projects undertaken in Australia during the 1995-2010 period on topics of relevance to the National Framework for Protecting Australia's Children. Some key characteristics of this body of research are:

• there was a much higher proportion of projects that were research compared with those with an evaluation component (79.5% and 19.6% respectively);
• there was a much higher proportion of projects addressing National Framework supporting outcome 4 (children who have been abused or neglected receive the support and care they need for their safety and wellbeing) (83.9%), compared to the proportion that addressed any of the other individual supporting outcomes;
• of those projects that addressed child abuse and neglect, a much higher proportion of projects addressed sexual abuse when compared to those that addressed physical abuse, neglect or psychological maltreatment (26.2%, 5.8%, 4.6% and 3.6% respectively);
• a significant proportion of projects were funded by state governments, a Commonwealth research granting agency, government or universities;
• the majority of projects had a small or medium sample size (65.3%);
• the number and proportion of projects that addressed issues relating to Indigenous Australians increased over time; and
• very few projects focused on adolescents, children/parents with a disability or families with a CALD background (only 8.6% of the total number of projects).

Whereas the earlier discussion focused on the characteristics of the research identified, the following discussion synthesises these findings in order to explore what analysis of these audit data can tell us about:

• research gaps and areas for development; and
• outcomes and progress since previous audits.

4.1 Research gaps and areas for development

One of the key functions of this audit is to identify those areas where there are significant gaps in knowledge regarding the protection of Australia's children. Based on the findings of the audit, the following discussion incorporates four key types of gaps:

• content gaps: subjects that are under-researched, with particular consideration given to those topics addressed in the National Framework;
• population group gaps: population groups that are under-researched;
• methodological gaps: methodologies that are under-utilised; and
• data gaps: data that are not available.

Content gaps

Community education

There is a content gap in relation to the roles and effectiveness of community education regarding child abuse and neglect, parenting, child development and children's needs. Community education regarding this topic is identified in the National Framework as a key means of protecting children from abuse and neglect, supporting families and encouraging a shared understanding and responsibility for the problem of abuse and neglect. International research demonstrates that community education, through methods such as social marketing, can lead to improvements in knowledge of child abuse and neglect and behaviours and attitudes towards children (Horsfall, Bromfield, & McDonald, 2010). However, the findings from this audit suggest that there has been minimal research on this topic in Australia in the past 16 years. There have also been few evaluations of community education programs (6.6% of all evaluation projects), which therefore limits our understanding of the effectiveness of these types of programs.

If, child abuse and neglect is "everyone's business", as researchers, child welfare advocates and policy-makers consistently argue, then people need to learn about what child abuse and neglect is, how to identify it, how to respond to it and, perhaps most importantly, how to support children and families in order to ultimately prevent child maltreatment. Research evidence is needed to understand whether (and how) community education can assist with this.

Risk factors

Domestic violence, substance abuse and mental illness have been consistently recognised as the key risk factors for abuse and neglect (Higgins & Katz, 2008; Bromfield, Lamont, Parker, & Horsfall, 2010). They are the drivers of the increased demand for statutory child protection services over the past decade. This audit found that of all these risk factors, domestic violence has been by far the most commonly researched topic; that is, research pertaining specifically to children, families and parenting and domestic violence has been much more common than research pertaining to children, families and parenting and substance abuse or mental illness.

It is perhaps not surprising that a search for projects relating to domestic violence and families would uncover a significant number of projects considering the way in which domestic violence is conceptualised (i.e., as a "family" problem, and often referred to as "family violence"). Nevertheless, considering the strength of evidence regarding the risks that parental substance abuse and mental illness pose to children's safety and wellbeing, the relatively few projects addressing these two issues - compared to domestic violence - suggests a significant knowledge gap.

The research relating to risk factors primarily focuses upon a single specific risk factor (e.g., substance abuse or mental illness or domestic violence) rather than multiple risk factors. This is significant in itself, as research demonstrates that these risk factors often co-exist (Bromfield et al., 2010; Dawe, Harnett & Frye, 2008). Furthermore, most of the research relating to risk factors pertains to the impact of those risk factors on children, and supporting children and families who are experiencing these difficulties (including improving service responses), rather than the actual identification or prevalence of risk factors.

The annual reports of the Victorian Child Death Review Committee (e.g., Victorian Child Death Review Committee, 2008, 2009) are an exception to this rule, as they provide information about the prevalence of risk factors among children and young people who died who were current or recent clients of child protection services.

Maltreatment types

As far as Australian research that focused upon a specific type of maltreatment is concerned, sexual abuse is a much more common focus than any other specific maltreatment type. Compared to sexual abuse, the other forms of maltreatment (physical abuse, neglect and psychological maltreatment) continue to receive less attention from researchers. Although there is some indication that the proportion of projects addressing these topics has increased slightly, they still do not reflect the prevalence of those maltreatment types within the Australian community.

In 2009-10, in all jurisdictions except Western Australia and the Northern Territory, the most common type of substantiated notification was emotional abuse (AIHW, 2011). In Western Australia and the Northern Territory, neglect was the most common type of substantiated abuse (AIHW, 2011). Yet, over the 1995-2010 period, these were the least likely of all types of maltreatment to be the focus of research (emotional abuse: 3.6%; neglect: 4.6%). Similarly, in 2009-10, in all states and territories except Western Australia, sexual abuse was the least common type of substantiated abuse (AIHW, 2011), yet over the 1995-2010 period, sexual abuse was the type of maltreatment that received the most individual attention (26.6%).

Although there are some problems with using child protection service data as a proxy for child maltreatment prevalence (e.g., the data only includes cases of maltreatment that were detected and reported and, as such, are likely to be an underestimation of the actual number of children who are abused or neglected), these data are the best available indicator that we currently have of the extent of the problem of child abuse and neglect in Australia (Bromfield & Horsfall, 2010). Therefore, the lack of focus on emotional abuse, neglect and physical abuse within the research is of significant concern.

A recent audit of child protection research in Ireland reported similar findings. Irish child protection research has a strong focus on sexual abuse (Buckley, Corrigan & Kerris, 2010). While neglect is the most commonly reported form of abuse in Ireland, the proportion of research on neglect is low compared to the proportion of research on sexual abuse. In Ireland, physical and emotional abuse are reported marginally less than sexual abuse but, the authors concluded, physical and emotional abuse topics still "appear to be under-researched" (p. ii).

In addition to sexual abuse being the most common type of maltreatment addressed in research, perpetrators/offenders of sexual abuse also appear to be a common theme in research addressing the issue of child abuse and neglect. Of all projects that addressed child abuse and neglect, 4.0% addressed this topic. On its own, this figure is not significant, but when considering that research looking at the perpetrators of other maltreatment types were not numerous enough to justify inclusion as a specific category, this seems to support the argument that there is a gap in the research on other maltreatment types.

While there is increasing recognition internationally that maltreatment types do not occur in isolation, many Australian researchers continue to focus on single maltreatment types, without considering the co-existence of other forms of maltreatment and their potential contribution to the issues observed (Higgins, 2004).

Those projects that do address all types of maltreatment are primarily annual government reports (such as the annual AIHW child protection reports; e.g., AIHW, 2011), or they relate to support for children who have been abused, or workforce, practice and system issues. Few projects that address all types of maltreatment focus upon the impact of the abuse on children, risk assessment or prevalence of abuse. This suggests a significant content gap in relation to the prevalence, impact and assessment of multiple forms of maltreatment. Few studies specifically address the overlap in children and young people's experience of multiple types of maltreatment and its unique cumulative or interactive impact.

Issues relating to Indigenous Australians

The audit findings suggest that research addressing issues relating to Indigenous Australians has increased significantly over the past 16 years; from 5.1% of total research projects in 1995-2000 to 26.2% in 2006-10.⁴⁶ Overall, it appears that issues relating to Indigenous Australians were more commonly addressed in research relating to out-of-home care than that relating to child abuse and neglect. This reflects the reality that Indigenous children are over-represented in the out-of-home care system (Berlyn & Bromfield, 2010).

Although 15.7% of projects overall addressed issues relating to Indigenous Australians, only 1.3% directly addressed the topics that are the focus of supporting outcome 5 (see Table 1):

• how to address Indigenous disadvantage (e.g., housing);
• community-wide strategies to address alcohol misuse and family violence;
• community-led or community-identified solutions to protecting children;
• Indigenous children in OOHC maintaining connection to family, community and culture; and
• partnerships between government, service providers and Indigenous families/communities that relate to protecting children.

This suggests that there are significant gaps and areas for development specifically in relation to these topics in research relating to Indigenous Australians. These findings are consistent with other reports noting the evidence gaps in this area (Higgins, 2010).

These specific gaps are significant for a number of reasons. Firstly, entrenched social problems in some Indigenous communities (especially poor housing) may play a role in the over-representation of Indigenous children in child protection and out-of-home care services (Berlyn & Bromfield, 2010); therefore research on how to address Indigenous disadvantage (rather than research that simply identifies or measures the disadvantage) is critical to reducing the over-representation of Indigenous children in these services.⁴⁷

The Closing the Gap initiative represents a positive step towards identifying how to address Indigenous disadvantage (COAG, 2010). Furthermore, the Closing the Gap Clearinghouse has developed a research and evaluation register specifically for the purpose of identifying research relating to this issue.

Secondly, the ability of Indigenous communities to function in a self-determining way is widely viewed as a critical factor in protecting Indigenous children from maltreatment (Cuneen & Libesman, 2000; Higgins, 2010). Partnerships between Indigenous communities and governments and/or service providers are critical also, in recognition of the fact that a "top-down" approach is not effective (Higgins, 2010). More research is required to bridge the gap regarding this specific content area. Research that demonstrates the effectiveness of self-determination and partnerships appears to be an area that could be addressed.

Lastly, maintaining Indigenous children's connection to culture when they are in out-of-home care is critical to effectively meet the needs of Indigenous children (Higgins, Bromfield, Higgins, & Richardson, 2006). Establishing and maintaining these connections through initiatives such as the Aboriginal Child Placement Principle recognises the damaging impact of previous government policies that undermined them (Berlyn & Bromfield, 2010).

Out-of-home care

In 2004, Cashmore and Ainsworth highlighted the "almost exclusive focus upon foster care" (p. 24) within Australian out-of-home care research. Noting that this lack of attention is "understandable in the sense that there are few options other than family based foster care in Australian and very little residential care" (p. 24), the authors nevertheless identified kinship care as a "possible priority" for a future research agenda.

The findings from the current audit demonstrate that foster care continues to dominate research about out-of-home care when compared to research addressing kinship care. As kinship care is the fastest growing type of out-of-home care in Australia (McHugh, 2009), there is an obvious knowledge gap regarding this specific type of care situation. Issues such as whether kinship care leads to better outcomes compared to foster care (e.g., greater placement stability) are critical considering the increased reliance upon kinship care in Australia.

Another "possible priority" identified by Cashmore and Ainsworth (2004) was permanency planning. In the current audit, only a small number of projects addressed issues relating to reunification and permanency planning (n = 15), and a very small number of projects addressed adoption from out-of-home care (n = 3). Combined, this constitutes just 4.9% of the total projects that addressed issues relating to out-of-home care. Permanency planning, therefore, also appears to still be a significant knowledge gap.

Evaluations

Overall, the proportion of projects that were evaluations of programs, services or strategies, when compared to research projects, was small (19.6% to 79.5% respectively).⁴⁸ This is a concern considering that program evaluations, especially impact and outcome evaluations, can determine whether or not interventions are of benefit to participants and are cost-effective (Holzer, Higgins, Bromfield, & Higgins, 2006).

Although some types of programs are more commonly evaluated than others (e.g., family support programs, prevention programs for children, and tertiary intervention programs for child victims), overall, the low proportion of projects that were evaluations suggests that there is a knowledge gap regarding the effectiveness of programs. The lower proportion of evaluation projects could be attributed to the recognised difficulties for service providers in conducting outcome and impact evaluations, such as a lack of resources (Tomision & Poole, 2000).

The high proportion of projects received via the 2005 and 2010 audit forms that were evaluations (36.2%), when compared to the proportion of evaluation projects overall (17.2%) (most of which were identified via published research) suggests that while evaluations are being undertaken, they are not being published. As noted by Higgins et al. (2005), the publication of evaluation results may not be a high priority for organisations undertaking evaluations, but this is an important way to share information about what works and prevent unnecessary replication of efforts that might have been shown not to work.

Population group gaps

Although research relating to Indigenous Australians is becoming more prominent, the full extent of the diversity of Australia's population is not reflected in research addressing the protection of Australia's children.

Children with a disability continue to be an under-researched population, despite the fact that they are at a higher risk of abuse and neglect than children without a disability (Sullivan & Knutson, 2000; Westat, 1993). Considering the cultural diversity of Australia's population, and the significant cultural issues within child protection services (Sawrikar, 2007; Kaur, 2007), research regarding CALD families and the protection of children also appears to be a significant gap. The significant challenges that adolescents face within child protection systems and out-of-home care have received some attention; however, overall, the proportion of projects focusing on that subgroup is relatively small (Robinson, 2010).

Methodological gaps

As far as Australian research that focuses upon a specific type of maltreatment is concerned (i.e., one or more of the five maltreatment types), sexual abuse is a much more common focus than any other specific maltreatment type. Leading researchers and advocates commenting on research within the fields of child welfare and child protection in Australia have consistently noted the abundance of small, qualitative research projects and the comparative shortage of larger scale, longitudinal, quantitative research projects (Cashmore & Ainsworth, 2004; Bromfield & Osborn, 2007; Cashmore, Higgins, Bromfield & Scott, 2006). The findings from this audit support this assertion, as there were:

• more qualitative than quantitative projects;
• few longitudinal studies; and
• very few randomised control trials.

As noted previously however, only a very small proportion of projects had the relevant information available to determine the type of methodology used; therefore, these findings should not be viewed as being representative of the entire field of research that relates to the protection of Australia's children. The fact that previous audits have made similar findings suggests that this finding is indicative of the overall body of research. As a result of the lack of large-scale, longitudinal, quantitative research projects, it is difficult to identify cause and effect relationships, track impacts on children and families of events over time, identify critical factors such as prevalence and incidence, and produce findings that can be applied to the general population. The lack of randomised control trials limits the extent to which the effectiveness of programs and interventions can be identified. There are, however, challenges associated with this type of research within the field of child welfare research.⁴⁹

In regard to longitudinal studies, Cashmore and Ainsworth (2004) called for a "national or cross-state longitudinal study of children and young people in the child protection and care systems" (p. 26). The South Australian longitudinal study, a data source used to undertake a number of quantitative studies identified in this audit, while not national in scale, points to a positive direction in Australian research. However, apart from projects using this particular dataset, longitudinal research is uncommon.⁵⁰ A notable exception is the NSW Pathways of Care study of children and young people in out-of-home care (see section on Randomised control trials and longitudinal studies).

More longitudinal research is required across the board; however, there is some indication that there is more longitudinal research being conducted within the field of out-of-home care, when compared to the fields of child maltreatment and child protection services research.

Data gaps

The content gaps identified above can only be overcome if reliable and accessible data are available. The accessibility of data was a key issue noted in the Towards a National Research Agenda forum (AIFS, 2009). Cashmore and Ainsworth (2004) also identified the need for data repositories that enable researchers to carry out secondary analyses of existing data. Access to case-level data (rather than aggregate data) is also a significant concern for researchers working in the area (Cashmore & Ainsworth, 2004). Developing unit record data relating to the statutory child protection system is one of the priorities noted in the document Implementing the First Three-Year Action Plan 2009-2010 (the first of the three-year action plans under the National Framework; COAG, 2009a). This will allow state/territory departments and other researchers to undertake more detailed analyses of the data.

It is difficult to determine from the audit findings whether certain content gaps relate to a lack of interest and/or expertise on the behalf of researchers or to a lack of reliable and accessible data. For example, the lack of research on CALD families could relate to the fact that child protection services do not record the cultural background of families unless they are from an Indigenous background (Sawrikar, 2007). This makes it difficult for researchers to address, for example, the prevalence of families from a CALD background coming into contact with child protection services.

Furthermore, the type of research that can be conducted is limited by the datasets available. Many of the research priorities identified during the Towards a National Research Agenda forum would benefit from longitudinal data. The lack of longitudinal research in Australia relates not only to the difficulties of undertaking longitudinal research, but also to the lack of reliable and accessible longitudinal administrative datasets. As noted in the forum, researchers need access to Australian longitudinal data that specifically focuses upon disadvantaged and vulnerable children and their potential for experiencing abuse and neglect.

4.2 Outcomes and progress since previous audits

Table 34 compares the findings from the Cashmore and Ainsworth (2004) audit, the Higgins et al. (2005) audit and the current audit.

Note: ^aAudit of Australian Out-of-Home Care Research (2004). ^bNational Audit of Australian Child Protection Research 1995-2004 (2005).

Some of the areas where there are notable differences are:

• the specific maltreatment types addressed;
• types of organisations conducting projects;
• degree programs;
• level of funding; and
• sources of funding.

It is difficult to make meaningful conclusions based upon the differences between the three audits, primarily because of their differing methodologies. Self-selection, inconsistent use of coding categories by audit form respondents and the aforementioned limitations of relying upon abstracts to glean information about projects may have resulted in many of the differences identified in Table 34.

As an illustration, clearly one of most significant differences relates to maltreatment types. The Higgins et al. (2005) audit identified much larger proportions of projects addressing each of the different specific maltreatment types (e.g., 82.2% of projects addressed sexual abuse, compared to 27% for the current audit; 84.4% addressed physical abuse, compared to 6%). Furthermore, whereas the current audit identified a comparatively larger proportion of projects relating to sexual abuse when compared to most other maltreatment types, the Higgins et al. (2005) audit did not.

These differences are likely to be the result of differing methodologies, rather than any significant outcome or progress over the last five years. In their analysis of maltreatment for the 2005 audit, Higgins et al. (2005) relied solely on the information received via the audit form. Information about projects received via the audit form provide more detail than information about projects identified through abstracts. This may have influenced the difference between the findings of the two audits - audit form respondents for the 2005 audit were simply able to provide more detailed information. By coding information on all projects with available information for the current audit (n = 489), we were able to include a much larger body of research in the analysis of maltreatment types.

Higgins et al. (2005) also identified that the information received via the audit form may not be accurate as "the validity of categorisations assigned by some respondents to the 2005 audit form may be questionable" (p. 26). Hence, this limitation of the Higgins et al. (2005) methodology may have affected their findings regarding the number of projects addressing specific types of maltreatment; hence the reason for the change in method for the current audit.

For these reasons, the best way of assessing progress since the time previous audits were conducted is to compare the different time periods within the current audit. The key changes over time include:

• a decrease in projects that focus upon sexual abuse;
• a slight increase in the proportion of projects that focus upon physical abuse, neglect and psychological maltreatment;
• an increase in the proportion of projects that address all types of maltreatment;
• an increase in the proportion of projects that address out-of-home care;
• a substantial increase in the proportion of projects undertaken by researchers from the discipline of social work and a gradual decrease in the proportion of projects undertaken by researchers from the discipline of psychology; and
• a substantial increase in the proportion of projects that address issues relating to Indigenous Australians.

4.3 Implications for research and researchers

The implications of the findings of the current audit for research and researchers relate both to knowledge gaps and infrastructure. In relation to knowledge gaps, in the past it may have been difficult for researchers working in the field of child welfare to identify exactly which topics were most relevant and of the highest priority to policy-makers and practitioners. By synthesising the needs of practitioners and policy-makers, the development of a national research agenda (one of the national priorities for 2009-12 implementation plan for the National Framework (COAG, 2009a) could enable researchers to identify exactly what type of evidence is needed in order to improve outcomes for Australia's children.

Another key implication for research and researchers in relation to knowledge gaps are the challenges associated with conducting research within the child welfare field. Challenges such as ethical dilemmas surrounding the suspicion or disclosure of abuse, and difficulties investigating causal relationships (Bromfield & Arney, 2008), make it difficult to employ the types of methodologies and research designs identified as lacking in Cashmore and Ainsworth (2004), Higgins et al. (2005) and the current audit. However, these methodologies are critical to answering many of the research questions identified above.

In regard to infrastructure, there are a number of key points to make. Firstly, Cashmore and Ainsworth (2004) noted that: "the future of research depends on the development of a new generation of researchers to replace the small number of established researchers" (p. 50). There appears to be a new generation of potential emerging researchers in this field, with almost one in five projects (18.5%) completed as part of a degree program.

However, in order to attract and retain a new generation of researchers after they have completed their degrees, there needs to be adequate funding and infrastructure available for universities and other institutes and organisations in order to continue to support research endeavours. There is evidence to suggest that much research is conducted on a small budget and, in some cases, with no external funding whatsoever. This affects not only the type of research that can be done (i.e., small-scale qualitative studies are usually much less expensive than large-scale longitudinal studies), but also the attraction and retention of new researchers.

4.4 Implications for policy and practice

Policy-makers and practitioners require research-based evidence in order to bring about improved outcomes for children, as Bromfield and Arney (2008) pointed out:

It is vital that interventions with vulnerable children and their families are accountable, that they actually work and, at the very least, that they do no further harm. (p. 11)

The implications of the content and methodological gaps identified above are that the evidence base that policy-makers and practitioners rely upon is limited, thereby affecting their ability to bring about improved outcomes for Australia's children.

Although international research can be relevant to policy-makers and practitioners, some of the unique characteristics of the Australian context (e.g., geographical size, remote regions, Indigenous cultures, other intersecting social and economic systems relating to health, income support and so on) mean that research relating specifically to the Australian context is required.

As the work of policy-makers and practitioners relies upon research, the dissemination of research findings is a critical issue. Whereas much research identified through the audit has been published, the publication of research findings in journal articles or the presentation of findings at academic conferences is not always the best way of disseminating research to policy-makers and practitioners. The Protecting Australia's Children Research and Evaluation Register, is an important step towards making research more accessible to these audiences (see section 5).

4.5 Limitations

The current audit relied upon literature searches as the main method for identifying projects that met the inclusion criteria, in recognition of the limitations of relying solely upon audit form data for the purposes of analysis. It was expected that gathering data via both a library search and the audit form would provide a more complete, comprehensive picture of research being undertaken when compared to an audit that relied solely upon individuals manually entering information in an online audit form.

One of the limitations of data sourced via a library search is that they do not provide the same level of detail as audit form responses (as only abstracts were reviewed). As Higgins et al. (2005) noted, reviewing entire documents is a time-consuming and expensive process and this was not possible with the resources available for the current audit. In some cases, the abstract did not provide enough information to determine whether those projects met the inclusion criteria for the audit, and thus they were not included in the audit. There is a possibility therefore that some projects that met the inclusion criteria were not included in the audit.

One of the drawbacks of relying upon two separate groups (i.e., project researchers and the audit team) to categorise projects is that there may be some inconsistencies between the two. As noted previously, a number of audit form respondents identified their project as addressing the topic of the "prevention of child abuse and neglect", yet when asked whether their project addressed the issue of child abuse and neglect, they responded "No". A limitation of the audit form was that it did not "pre-fill" responses to questions such as these.

Another limitation of the audit form was that it did not include a question that enabled an analysis of geographical gaps. Audit form respondents were asked to identify in which state or territory the researchers undertaking the project were based; however, it did not ask them about the geographical focus of the project (e.g., specific state/territory focus, rural and remote focus). A preliminary analysis of the title and abstracts of all 1,359 projects (by searching, for example, for the names of specific states/territories) identified 387 projects where a single specific state or territory was identified. These data were deemed invalid, however, as projects that did not include information about geographic focus in the title or abstract of their project may have been missed.

4.6 Implications for future audits of research

Research audits are intended to provide an indication of the extent, characteristics (e.g., subtopics, methodology, funding body) and quality of research pertaining to a particular topic. For research audits with a relatively broad scope, such as this one, there are a number of key issues to consider:

• An audit form that is made available to researchers to record the details of their research is unlikely to capture all research that has been undertaken about a topic if that topic is relatively broad in scope (e.g., protecting Australia's children). Even when an intensive process of promotion that targets key organisations is undertaken, some relevant research will not be captured, either because the relevant parties are not aware of the audit or do not have time to complete an audit form.
• Research audits that combine an audit form and a literature search are likely to provide a more comprehensive picture of the extent of research pertaining to a particular topic when compared to audits that rely solely upon an audit form.
• Audit forms are especially useful for capturing information about unpublished research (e.g., research currently being undertaken), as this research will not be captured via traditional literature search methods.
• The need to capture detailed information about research via an audit form needs to be balanced with the understanding that researchers' time is limited. The more complex and detailed an audit form is, the less willing researchers will be to complete it.
• Information provided via an audit form can appear inconsistent. For this reason, it would be extremely beneficial for audit forms to be designed to minimise inconsistencies (e.g., where a respondent notes their project is about the prevention of abuse and neglect, the question about whether their project addresses child abuse and neglect is automatically answered as "Yes").
• In order to identify geographical gaps, the audit form needs to explicitly ask if the research focuses on a specific region (e.g., state or territory), as the other material pertaining to the project (e.g., title, aim/purpose and findings) may not include this information.
• In order to identify the characteristics (e.g., methodology) and quality of research, it is necessary to have available relevant information about that research. Identifying this information by referring to a full publication (e.g., journal article, research report) is an extremely time-consuming process as it requires audit team members to review a vast amount of material. The more extensive the scope of the topic, the more time-consuming this process will be. Moreover, using this method of identifying relevant information requires access to full-text publications, and this capacity is limited among organisations that exist outside a university context. An audit is not a substitute for a systematic review of particular content areas.
• Identifying relevant information by referring to publication abstracts is a more efficient method of identifying the characteristics and quality of research. However, abstracts do not always contain all the relevant information (e.g., only 8.2% of projects identified for the current project explicitly stated whether they used a qualitative and/or quantitative methodology) and not all abstracts are publicly available, thus restricting the extent to which the characteristics and quality of a body of research can be determined.
• It is preferable to include all research in an audit, regardless of whether abstracts are available, as it provides an indication of the extent (i.e., amount) of research on a topic. However, it is also important to note that identifying relevant research without an abstract can be difficult if the title of the research does not indicate the topic of the research.

In conclusion, and perhaps most importantly, although a written report of a research audit provides aggregated information about research undertaken during a certain period of time, providing researchers and policy-makers with access to a searchable database of research identified through an audit is equally, if not more useful. The following concluding section of this report describes the resource that aims to provide this service, the Protecting Australia's Children Research and Evaluation Register.

5. Protecting Australia's Children Research and Evaluation Register

In 2005, Higgins et al. (2005) concluded their audit of research relating to child protection by calling for a database of research relating to child protection and child maltreatment that functions as a "living resource" rather than an "archival record" (p. 28). They argued that if such a database were maintained, "it would become a more accurate reflection of the scope of child protection research in Australia" (p. 28). They also highlighted the benefits of providing professionals working in policy and practice with greater accessibility to research, claiming that access to research "is the cornerstone of effective dissemination to inform policy and practice" (p. 28).

During the period that this audit was being conducted, the Australian Institute of Family Studies developed the infrastructure for a searchable online database of all the projects included in the audit (n = 1,359). The database includes detailed information about each of the projects, including the names of chief investigators, aim/purpose of the research, findings and implications of the research, methodological information (e.g., data sample, data size) and the contact details of researchers.

The database has the capacity for the records of projects currently in the database, including information about new projects, to be updated over time, so it serves the function that Higgins et al. (2005) noted as being especially significant; that is, functioning as a "living resource" rather than an "archival record".

There are three key benefits of the database:

• providing researchers in the field with information about research that has already been undertaken and research that is currently underway - this limits the potential for duplication of effort by researchers and enables them to identify gaps in the research using a relatively straightforward method;
• providing policy-makers and practitioners with access to information about research to inform the planning and delivery of services - one of the key barriers to utilising research in policy and practice is a lack of access to research, and providing access to the database fulfils this need; and
• undertaking regular audits of research in a more effective, efficient manner - the database allows researchers to provide information about their research at a time that is convenient to them (rather than inputting information via an audit form that is only available for a certain period of time) and provides those undertaking an audit with an easily accessible database of research, thereby reducing the need for comprehensive literature searches and audit forms.

In order to function as a "living resource", it is critical that the database is regularly promoted to those organisations and individuals who are likely to be undertaking research relating to the protection of Australia's children. A database that is widely used and easily accessible will provide the incentive for researchers to include their research, as it provides them with an opportunity to promote their own research within the field. The Protecting Australia's Children Research and Evaluation Register has the potential to make a significant indirect contribution to improving outcomes for children and families in Australia.

References

• Australian Bureau of Statistics. (2010). Australian demographic statistics, June 2010. Canberra: ABS. Retrieved from www.abs.gov.au/ausstats/[email protected]/mf/3101.0.
• Australian Institute of Health and Welfare. (2011). Child protection Australia 2009-2010. Canberra: AIHW.
• Australian Institute of Family Studies. (2009). Towards a National Research Agenda. Unpublished report.
• Berlyn, C., & Bromfield, L. M. (2010). Child protection and Aboriginal and Torres Strait Islander children (NCPC Resource Sheet). Melbourne: National Child Protection Clearinghouse. Retrieved from www.aifs.gov.au/nch/pubs/sheets/rs10/index.html.
• Board of Inquiry into the Child Protection System in the Northern Territory. (2010). Growing them strong, together: Promoting the safety and wellbeing of the Northern Territory's children. Report of the Board of Inquiry into the Child Protection System in the Northern Territory 2010. Darwin: Northern Territory Government.
• Bogossian, F. E. (2003). The mother's health study: A randomised controlled trial of a social support intervention on the health of mothers in the year after birth. Unpublished PhD thesis, University of Queensland.
• Bromfield, L., & Arney, F. (2008). Developing a road map for research: Identifying the priorities for a national child protection research agenda (Child Abuse Prevention Issues No. 28). Melbourne: National Child Protection Clearinghouse. Retrieved from www.aifs.gov.au/nch/pubs/issues/issues28/issues28.html.
• Bromfield, L., & Osborn, A. (2007). Getting the big picture: A synopsis and critique of Australian out-of-home care research. Melbourne: National Child Protection Clearinghouse.
• Bromfield, L., Higgins, D., Osborn, A., Panozzo, S., & Richardson, N. (2005). Out-of-home care in Australia: Messages from research. Melbourne: National Child Protection Clearinghouse.
• Bromfield, L. M., & Holzer, P. J. (2008). Australian Institute of Family Studies submission to the Special Commission of Inquiry into Child Protection Services in NSW (PDF 649 KB). Melbourne: Australian Institute of Family Studies. Retrieved from www.aifs.gov.au/nch/pubs/submissions/nswchildprotect/inquiry.pdf.
• Bromfield, L., & Horsfall, B. (2010). Child abuse statistics (NCPC Resource Sheet). Melbourne: National Child Protection Clearinghouse. Retrieved from www.aifs.gov.au/nch/pubs/sheets/rs1/rs1.html.
• Bromfield, L., Lamont, A., Parker, R., & Horsfall, B. (2010). Issues for the safety and wellbeing of children in families with multiple and complex problems (NCPC Issues No. 33). Melbourne: National Child Protection Clearinghouse. Retrieved from www.aifs.gov.au/nch/pubs/issues/issues33/index.html.
• Buckley, H., Corrigan, C., & Kerrins, L. (2010). Report of an audit of child protection research in Ireland 1990-2009 (PDF 881 KB). Dublin: Children Acts Advisory Board. Retrieved from www.tara.tcd.ie/jspui/bitstream/2262/41271/1/Full-Report-Audit-of-Child-Protection.pdf.
• Cashmore, J. A., & Ainsworth, F. (2004). Audit of Australian out-of-home care research. Sydney: Child and Family Welfare Association of Australia, Association of Child Welfare Agencies.
• Cashmore, J. A., Higgins, D. J., Bromfield, L. M., & Scott, D. A. (2006). Recent Australian child protection and out-of-home care research: What's been done and what needs to be done? Children Australia, 31(2), 4-11.
• Council of Australian Governments. (2009a). Implementing the first three-year action plan, 2009-2020: Protecting children is everyone's business. National Framework for Protecting Australia's Children. Canberra: COAG.
• Council of Australian Governments. (2009b). Protecting children is everyone's business: National Framework for Protecting Australia's Children. Canberra: COAG.
• Cunneen, C., & Libesman, T. (. (2000) Postcolonial trauma: The contemporary removal of Indigenous children and young people from their families in Australia. Australian Journal of Social Issues, 35(2), 99-115.
• Dawe, S., Harnett, P., & Frye, S. (2008). Improving outcomes for children living in families with parental substance misuse: What do we know and what should we do (Child Abuse Prevention Issues No. 29). Melbourne: National Child Protection Clearinghouse. Retrieved from www.aifs.gov.au/nch/pubs/issues/issues29/issues29.html.
• Edwards, B., Wise, S., Gray, M., Hayes, A., Katz, I., Mission, S., Patulny, R., & Muir, K. (2009). Stronger Families in Australia Study: The impact of Communities for Children. Canberra: National Evaluation Consortium (Social Policy Research Centre, at the University of New South Wales and the Australian Institute of Family Studies).
• Fraser, J. A. (2000). The role of home visiting as an early intervention strategy for prevention of child abuse and neglect. Unpublished PhD thesis, School of Applied Psychology.
• Fraser, J. A., Armstrong, K. L., Morris, J. P., & Dadds, M. R. (2000). Home visiting intervention for vulnerable families with newborns: Follow-up results of a randomized controlled trial. Child Abuse and Neglect, 24(11), 1399-1429.
• Higgins, D. (2004). Differentiating between child maltreatment experiences. Family Matters, 69, 50-55.
• Higgins, D. (2010). Community development approaches to safety and wellbeing of Indigenous children (PDF 336 KB). Canberra: Australian Institute of Family Studies. Retrieved from www.aihw.gov.au/closingthegap/documents/resource_sheets/ctgc-rs01.pdf.
• Higgins, D., Adams, R., Bromfield, L., Richardson, N., & Aldana, M. (2005). National audit of Australian child protection research 1995-2004. Melbourne: Australian Institute of Family Studies.
• Higgins, D. J., Bromfield L. M., Higgins, J. R., & Richardson, N. (2006). Protecting Indigenous children: Views of carers and young people on out-of-home care. Family Matters, 75, 42-49. Retrieved from www.aifs.gov.au/institute/pubs/fm2006/fm75.html.
• Higgins, D. J., & Katz, I. (2008). Enhancing service systems for protecting children. Family Matters, 80, 43-50.
• Holzer, P., Higgins, J., Bromfield,  L., & Higgins, D. (2006). The effectiveness of parent education and home visiting child maltreatment programs (Child Abuse Prevention Issues No. 24). Retrieved from www.aifs.gov.au/nch/pubs/issues/issues24/issues24.html.
• Horsfall, B., Bromfield, L., & McDonald, M. (2010). Are social marketing campaigns effective in preventing child abuse and neglect? (NCPC Issues No. 32). Retrieved from www.aifs.gov.au/nch/pubs/issues/issues32/index.html.
• Kaur, J. (2007). Working with families from culturally and linguistically diverse communities in Queensland: An Australian exploratory study. Children Australia, 32(4), 17-24.
• King, N. J., Tonge, B. J., Mullen, P., Myerson, N., Heyne, D., Rollings, S., et al. (2000). Treating sexually abused children with posttraumatic stress symptoms: A randomized clinical trial. Journal of the American Academy of Child and Adolescent Psychiatry, 39(11), 1347-1355.
• Lamont, A. (2009). Evaluating child abuse and neglect intervention programs (NCPC Resource Sheet). Retrieved from www.aifs.gov.au/nch/pubs/sheets/rs5/rs5.html.
• Lamont, A. (2010). Effects of child abuse and neglect for children and adolescents (NCPC Resource Sheet). Retrieved from www.aifs.gov.au/cfca/pubs/factsheets/a146141/index.html. (republished by Child Family Community Australia, 2014)
• Mayfield, B. (2009). A systematic map of the out-of-home care literature 1995-2008. Unpublished report.
• McHugh, M., & Valentine, K. (2010). Financial and non-financial support to formal and informal out of home carers: Final report. Report prepared for the Department of Families, Housing, Community Services and Indigenous Affairs. Sydney: Social Policy Research Centre, University of New South Wales.
• McHugh, M. (2009). A framework of practice for implementing a kinship care program: Final report. Sydney: Social Policy Research Centre.
• Muir, K., Katz, I., Purcal, C., Patulny, R., Flaxman, S., Abello, D. et al. (2009). National evaluation (2004-2008) of the Stronger Families and Communities Strategy 2004-2009. Canberra: Department of Families, Housing, Community Services and Indigenous Affairs.
• Robinson, E., & Miller, R. (2010). Adolescents and their families: Best interests case practice model specialist (Practice Resource Guide). Melbourne: Department of Human Services.
• Sawrikar, P. (2009). Culturally appropriate service provision for culturally and linguistically diverse (CALD) children and families in the New South Wales (NSW) child protection system (CPS). Sydney: Social Policy Research Centre, University of New South Wales
• Sullivan, P. M. & Knutson, J. F. (2000). Maltreatment and disabilities: A population-based epidemiological study. Child Abuse & Neglect, 24, 1257-1274.
• Tomison, A., & Poole, L. (2000). Preventing child abuse and neglect: Findings from an Australian audit of prevention programs. Melbourne: Australian Institute of Family Studies.
• Victorian Child Death Review Committee. (2008). Annual report of inquiries into the deaths of children known to Child Protection 2008. Melbourne: Office of the Child Safety Commissioner.
• Victorian Child Death Review Committee. (2009). Annual report of inquiries into the deaths of children known to Child Protection 2009. Melbourne: Office of the Child Safety Commissioner.
• Westat, Inc. (1993). A report on the maltreatment of children with disabilities. Washington, DC: National Center on Child Abuse and Neglect.

Appendix A: Literature search terms

Note: ^a These search terms are based upon the subject headings used in Australian Family and Society Abstracts.

Appendix B: Audit form

Thank you for participating in the Protecting Australia's Children: Research Audit 1995-2010.

Contact details will be collected for the purpose of clarifying any information you provide. You will be given the option of including your contact details in a publicly searchable database.

If you are inputting information about multiple projects please complete one questionnaire for each individual project.

If you have any questions about the audit you can phone (03) 9214 7888 and ask to speak to the Protecting Australia's Children: Research Audit co-ordinator or contact the project team

There are 38 questions in this survey

Project details

We would like to begin by asking you for some general background information about the project 
Mandatory questions are indicated by a red asterisk *

1 What is the name of the project? *

2 Which of the following topics does the project address? *

Please choose all that apply:

• Building safe, supportive communities for children and/or families
• Support for vulnerable and at-risk families
• Prevention of child abuse and neglect (including primary, secondary and tertiary prevention)
• Child abuse and neglect (including sexual abuse, physical abuse, psychological maltreatment, emotional abuse/neglect, physical neglect and witnessing family violence)
• Child protection services and/or systems and/or practice
• Other child protection issues
• Support for children who have been abused or neglected
• Out of home care (including leaving care)
• Other related category

3 Please describe this other related category

Only answer this question if the following conditions are met: 
° Answer was 'Other related category' at question '2 [TOPICS2]' (Which of the following topics does the project address?)

4 What year did the project commence?

5 Has the project been completed? *

Please choose only one of the following:

• Yes
• No

6 What year was the project completed?

Only answer this question if the following conditions are met: 
° Answer was 'Yes' at question '5 [COMPLETE]' (Has the project been completed?)

7 Who is/are the chief investigators on the project?

If there are more than 4 chief investigators please list the names of the investigators in alphabetical order and include the first 4 only.

8 What are the disciplinary areas of the chief investigators?

Please choose all that apply:

• Social work
• Social policy
• Psychology
• Education
• Sociology
• Law
• Other:

9 What type of organisations are involved in conducting the project? *

Please choose all that apply:

• University
• Government department
• Research Institute
• Non-government organisation/agency
• Other:

10 Is the project being conducted as part of the requirements of a degree program? (e.g., Professional Doctorate/PhD, Masters or Honours research)

Please choose only one of the following:

• Yes
• No

11 This project is being conducted for the requirements of a:

Only answer this question if the following conditions are met: 
° Answer was 'Yes' at question '10 [DEG]' (Is the project being conducted as part of the requirements of a degree program? (e.g., Professional Doctorate/PhD, Masters or Honours research))

Please choose only one of the following:

• Professional Doctorate/PhD
• Masters degree
• Honours course
• Other

Research Purpose and Focus

We would now like to ask you some questions about the purpose and focus of the project 
Mandatory questions are indicated by a red asterisk *

12 What is the aim or purpose of this project? (maximum 200 words) *

 Feel free to cut and paste this information from another document

13 Is the project research only, evaluation only or both research and evaluation? *

Please choose only one of the following:

• Research only
• Evaluation only
• Research and evaluation
• Don't know/unsure
• Other

Research is conducted with to the intent to generalize the findings from a sample to a larger population. Evaluation usually focuses upon collecting data for the purposes of determining the worth, merit or value of a specific program or client group activity.

14 Please indicate what type of program is being evaluated

Only answer this question if the following conditions are met: 
° Answer was 'Evaluation only' or 'Research and evaluation' at question '13 [TYPE]' (Is the project research only, evaluation only or both research and evaluation?)

Please choose all that apply:

• Prevention program for children
• Tertiary intervention program for child victims
• Child-focused programs within adult-oriented services
• Family support program
• Parent education program
• Community education program
• Statutory child protection service
• Specialist/innovative Child Protection service program
• Other:

15 What is the research question or  hypothesis? (if relevant) (maximum 150 words)

Feel free to cut and paste this information from another document

16 Does the project address the topic of child abuse or child neglect? *

Please choose only one of the following:

• Yes
• No
• Don't know/unsure

17 What forms of child abuse or neglect are covered in the project?

Only answer this question if the following conditions are met: 
° Answer was 'Yes' at question '16 [CAN]' (Does the project address the topic of child abuse or child neglect?)

Please choose all that apply:

• sexual abuse
• physical abuse
• psychological abuse/emotional abuse
• physical neglect
• witnessing family violence
• all maltreatment types
• Other:

18 Please describe the major findings from the project (if available) (maximum 250 words)

If the project has not reached the stage where findings can be identified please leave this question blank. 

Findings from individual projects will not be included in the audit report (unless they have already been published in a publicly available report, journal article etc). 

Feel free to cut and paste this information from another document about your project (e.g., project proposal, ethics application, journal article, report)

19 Please describe any implications for policy and/or practice that emerge as a result of this project (if relevant) (maximum 250 words)

If the project has not reached the stage where implications can be identified please leave this question blank. 

Implications for policy and/or practice will not be included in the audit report (unless they have already been published in a publicly available report, journal article etc). 

Feel free to cut and paste this information from another document about your project (e.g., project proposal, ethics application, journal article, report).

20 Will the project address issues relating to Indigenous Australians?

Please choose only one of the following:

• Yes
• No
• Don't know/unsure
• Not applicable

Methodology

We would now like to ask you some questions about the project methodology 
Mandatory questions are indicated by a red asterisk *

21 From which of the following groups are you collecting data for this project? *

Please choose all that apply:

• Parents
• Foster carers
• Kinship carers
• Children/Young people
• Professionals (e.g., child protection workers, social workers, family support workers, teachers)
• Not applicable
• Other:

22 What term best describes the parents who will constitute the data sample for the project?

Only answer this question if the following conditions are met: 
° Answer was at question '21 [SAMP1]' (From which of the following groups are you collecting data for this project?)

Please choose all that apply:

• Parents from the general community
• Parents at risk of abusing or neglecting their children
• Parents whose children are in out-of-home care
• Other:

23 Which of the following groups do the professionals in the data sample come from?

Only answer this question if the following conditions are met: 
° Answer was at question '21 [SAMP1]' (From which of the following groups are you collecting data for this project?)

Please choose all that apply:

• Child protection workers
• Welfare/family support workers/community development workers
• Health professionals
• Education professionals
• Legal professionals
• Other:

24 What is the size of the data sample?

If you are uncertain what the size of the data sample will be please provide an estimate.

25 Will the data sample include:

Please choose only one of the following:

• Indigenous Australians only
• Indigenous and non-Indigenous Australians
• Non-Indigenous Australians only
• Don't know/unsure
• Not applicable

The term 'Indigenous Australians' as it is used here incorporates Aboriginal and Torres Strait Islander people.

Location

We would now like to ask you some questions about the location of the research 
Mandatory questions are indicated by a red asterisk *

26 What is the primary state/territory where the research is based? *

Please choose only one of the following:

• ACT
• New South Wales
• Northern Territory
• Queensland
• South Australia
• Tasmania
• Victoria
• Western Australia

This question concerns where the project is being conducted not where the data sample comes from.

27 Is the project being conducted across different regions in one state/territory? (e.g., two different suburbs, one metropolitan and one regional area)

Please choose only one of the following:

• Yes
• No

Funding

We would now like to ask you some questions about funding for the project

28 Please identify what type of organisation is funding the project

Please choose all that apply:

• Commonwealth Research Granting Agency (ARC, NH&MRC)
• State/Local Government
• University
• Philanthropic trust
• Self funded
• Not applicable
• Other:

29 What is the total amount of funding that will be provided to undertake the research project?

 Please provide the amount in whole dollars. Please provide the amount in Australian dollars. If no funding is being provided please enter '0'

Publications

We would now like to ask you some questions about publications, conferences and other resources produced as a result of this project.

30 Please list any journal articles that have presented the findings of this project

Feel free to cut and paste directly from a reference list. 

If you are not cutting and pasting from a reference list we would greatly appreciate it if you could provide details of products in the APA style of referencing. See below for an example of the APA style of referencing for journal articles. 

Journal articles 

Marks, N. F., Lambert, J.D., & Choi, H. (2002). Transitions to caregiving, gender and psychological well-being: A prospective US national study. Journal of Marriage and Family, 64(3), 657-667.

31 Please list any conference papers or conference proceedings where the findings from this project have been presented.

Feel free to cut and paste directly from a reference list. 

If you are not cutting and pasting from a reference list we would greatly appreciate it if you could provide details of products in the APA style of referencing. See below for an example of the APA style of referencing for conference papers. 

Conference paper 

McDonald, P., & Kippen, R. (2000, March). Building child-friendly communities in Australia. Paper presented at the Annual Meeting of the Population Association of America, Los Angeles, California.

32 Please list any reports that have presented the findings of this project.

Feel free to cut and paste directly from a reference list. 

If you are not cutting and pasting from a reference list we would greatly appreciate it if you could provide details of products in the APA style of referencing. See below for an example of the APA style of referencing for reports. 

Report 

Jones, B., & Karol, S. (2004). Children and homelessness: Evaluation of the CARE program (Report No. 3). Canberra: The Australian Association of Caring Agencies.

33 Please list any unpublished dissertations of theses that have been produced as a result of this project.

Feel free to cut and paste directly from a reference list. 

If you are not cutting and pasting from a reference list we would greatly appreciate it if you could provide details of products in the APA style of referencing. See below for an example of the APA style of referencing for an unpublished dissertation or thesis. 

Unpublished dissertation or thesis 

Smith, J. F. (2005). A study of risk factors for child abuse and neglect. Unpublished doctoral dissertation, University of Wollongong, Wollongong, NSW.

34 Please list any books or book chapters that present the findings of this project.

Feel free to cut and paste directly from a reference list. 

If you are not cutting and pasting from a reference list we would greatly appreciate it if you could provide details of products in the APA style of referencing. See below for an example of the APA style of referencing for books and book chapters. 

Book 

Heady, B., Warren, D., & Harding G. (2006). Families, incomes and jobs: A statistical report of the HILDA survey. Melbourne: Melbourne Institute of Applied Economic and Social Research. 

Chapter in an edited book 

Bittman, M. (2004). Parenting and employment: What time use surveys show. In N. Folbre & M. Bittman (Eds.), Family time: The social organisation of care (pp. 69-89). London: Routledge.

35 Please list any other products that have been produced as a result of this project.

Contact Details

Mandatory questions are indicated by a red asterisk *

36 We may need to clarify some of the information you have provided in this questionnaire. Please provide an email address so we can contact you for this purpose. (The contact information you provide here will only be made available to the researchers working on the project) *

Contact for clarification:

• Email address

37 One of the outcomes of this audit will be the development of a publicly available database that will incorporate information about research and evaluation projects. Please provide public contact details in the text boxes below. *

Public contact details:

• Name (this could be the name of a project/team or organisation)
• Internet or email address

38 Please click SUBMIT to finish the survey

Submit your survey. 
Thank you for completing this survey.

List of tables

• Table 1: Topics included and excluded from the audit, by National Framework supporting outcomes
• Table 2: Key audit findings
• Table 3: Distribution of projects identified, by year completed, 1995-2010
• Table 4: Distribution of projects identified, by year of audit and source
• Table 5: Distribution of projects identified, by type of project (where identified)
• Table 6: Distribution of evaluation projects identified, by year completed
• Table 7: Distribution of evaluation projects identified, by types of programs evaluated (where identified)
• Table 8: Distribution of evaluation projects identified, by type of program evaluated and year completed, 1995-2010 (where identified, n = 181)
• Table 9: Distribution of projects identified, by supporting outcomes (where identified)
• Table 10: Distribution of projects addressing the topic of child abuse or neglect, by year completed, 1995-2010
• Table 11: Distribution of child abuse or neglect projects, by subtopics addressed (where identified)
• Table 12: Distribution of child abuse or neglect projects, by maltreatment type(s) addressed
• Table 13: Changes over time in distribution of child abuse or neglect projects, by maltreatment types and year completed (1995-2010)
• Table 14: Distribution of out-of-home care projects, by year completed, 1995-2010
• Table 15: Distribution of out-of-home care projects, by subtopics addressed (where identified)
• Table 16: Distribution of projects, by discipline area of chief investigators (where identified)
• Table 17: Change over time in distribution of projects, by discipline area of chief investigators (where identified, n = 315) and year completed, 1995-2010 (where identified, n = 315)
• Table 18: Distribution of projects, by type of organisation conducting them (where identified)
• Table 19: Change over time in distribution of projects, by type of organisation conducting them (where identified, n = 361)
• Table 20: Distribution of cross-organisational and non-cross-organisational projects, by type of organisation conducting them (where identified, n = 361)
• Table 21: Level of project funding (where identified, n = 80)
• Table 22: Comparison of project funding for all three audits
• Table 23: Distribution of projects, by type of organisation funding projects (where identified)
• Table 24: Change over time in distribution of projects, by type of organisation funding projects and year completed, 1995-2010 (where identified, n = 200)
• Table 25: Distribution of projects, by type of degree program (where identified)
• Table 26: Distribution of projects, by sample type (for groups from which data were collected and where identified)
• Table 27: Distribution of projects, by type of professionals in data samples (where identified)
• Table 28: Distribution of projects, by status of parents in data samples (where identified)
• Table 29: Distribution of projects, by sample size (where identified)
• Table 30: Change over time in distribution of projects, by sample size (where identified, n = 288)
• Table 31: Distribution of projects by publication types
• Table 32: Changes over time in distribution of projects addressing issues relating to Indigenous Australians, 1995-2010
• Table 33: Distribution of projects, by population groups
• Table 34: Outcomes and progress since previous audits