Child Abuse and Neglect: a socio-legal Study of Mandatory Reporting in Australia

cluster analysis: for example, identifying groups of records about reports by specific reporter groups, about specific types of abuse and neglect;

anomaly detection: for example, identifying shifts in numbers and or outcomes of reports of a specific type of abuse, or by a specific reporter group, in any given year or as a sustained trend over an extended period;

association rule learning: for example, identifying relationships between a particular reporter group and their reports of a specific type of abuse, and between the reports made by a particular report group and the outcomes of those reports; and

summarisation: generation of full data reports, as well as executive summaries and policy briefs.

Interpretation of data

The patterns and trends identified in the data mining process then must be interpreted by situating them in the context of relevant legal and other contextual factors. Due to the scope of the project (for example, noting the limitation produced by the fourth qualitative stage not being funded), it was not possible for the research team to identify and evaluate the impact of all factors impacting reporting practice generally, separated by reporter group, or by type of abuse or neglect.

However, major legal factors were able to be considered, namely the presence and nature of mandatory reporting duties, and the timing of the introduction of new legislative mandatory reporting duties. As well, we were also able to consider major known contextual factors which would also likely influence trends in numbers and outcomes of reports. Some of these are factors identified in previous analyses of the approaches adopted by respective State and Territory systems such as agency intake methods, nomenclature, and the processing of intakes (Holzer & Bromfield, 2008). For example, the characterisation of an intake as a ‘notification’ is affected by factors including whether the caller or the agency defines the intake. In the ACT and Tasmania, intakes are caller-defined, producing higher numbers of ‘notifications’, and Victoria has a similar but more rigorous approach; in contrast, in other jurisdictions the intake is agency-defined: that is, the agency determines whether the intake is a ‘notification’ regarding suspected child abuse or neglect, or is a less serious ‘child concern’ report regarding a child’s overall welfare). Agency decisions about whether to investigate a notification or not are influenced by multiple factors including: differences in the availability of alternative diversionary services, and different levels of resources available to support investigations. An agency finding that an investigated notification is ‘substantiated’ is influenced by multiple factors including:

• 
  differences in the availability of alternative diversionary services (where a notification is referred to such a service, this will not be counted as a substantiated notification);
  
• 
  different thresholds for reaching a finding of ‘substantiated’;
  
• 
  availability of evidence of harm in any given case even where there is sufficient evidence of abuse or neglect;
  
• 
  availability of evidence of abuse or neglect in any given case even where there is sufficient evidence of harm;
  
• 
  whether the jurisdiction focuses on evaluating substantiation of existing harm, or risk of harm occurring in the future.
  

Other important contextual factors affecting numbers of notifications and responses to them would include locally-situated events. These could include events and developments such as:

• 
  the presence of child protection inquiries or reports;
  
• 
  high levels of media attention on particular cases or departmental processes;
  
• 
  introduction of a new industry-based reporting policy such as a standard operating procedure for police;
  
• 
  changes to child protection agencies (such as an influx of funding or staff) and their approaches;
  
• 
  heightened reporter education via awareness-building by professional education, media campaigns or other strategies; and heightened sensitivity of the general public to a particular form of child abuse or neglect created by policy, media or social discourse.
  

To gain a more nuanced appreciation of the kinds of contextual factors that may have influenced the data in this study, and to supplement their own knowledge of contextual factors, the research team invited and held meetings with senior staff in State and Territory departments as critical friends to elicit their views about what might underpin key findings from the data. Their accounts were informed by their knowledge of local contexts and agency developments. These were noted and carefully considered by the research team in their interpretations of the data.

Contributions to knowledge

Despite these multiple complexities, the descriptive statistics and data mining facilitate understandings of important elements of the context, especially within jurisdictions annually and over the decade. This is important because it provides more nuanced insights into the reporting practices of different reporter groups for different types of abuse and neglect, rather than simply aggregating all reports of all kinds of abuse and neglect by all reporter groups, and drawing undifferentiated conclusions about the nature and outcomes of ‘reporting of child abuse and neglect’.

This approach has been urged in research by the lead investigator, based on the notion of a differentiation thesis (Mathews, 2014; Mathews, 2012). It proceeds from the basis that heterogeneity exists across both reporter groups, and types of abuse and neglect. It proposes that because of the differential nature of various reporter groups, and the differentiation between types of child abuse and neglect, there are likely to be significant differences amongst various reporter groups’ reporting practices for different types of child abuse and neglect. Gaining a detailed understanding of specific components of different reporter groups’ reporting practices regarding different types of abuse and neglect provides more sophisticated insights into:

• 
  the nature of reporting trends;
  
• 
  where reporting appears to be more and less effective; and
  
• 
  where efforts to improve practice may be best directed.
  

Systemic differences between jurisdictions limit the extent to which cross-jurisdictional comparisons can be made. However, some comparisons may be possible where sufficient similarities exist, and/or where clear and significant trends are identified in the data. This is especially so where, by all accounts, there appears to be only one significant differentiating variable, such as the presence or absence of a mandatory reporting duty.

Effectiveness of reporting practices

This part of the project could also assist in informing government departments in their assessments of the effectiveness of reporting practices. While there are challenges in conceptualising how ‘effectiveness’ of reporting may be evaluated, it seems reasonable to proceed on the basis that identification of trends in numbers of reports and in outcomes of reports, placed in the context of legislative mandatory reporting duties and changes in them, could provide some useful information contributing to any assessment of the effectiveness of reporting practice.

Accordingly, the research project provides detailed data in each jurisdiction specific to the reporting practices of key reporter groups, about specific types of abuse and neglect, over time. In addition, to provide meaningful findings for individual State and Territory governments (but not for strict cross-jurisdictional comparison) about the proportional contribution to reporting made collectively by their jurisdiction’s major occupational reporter groups, we conducted some analyses by combining the reporting practice of the ‘major mandated reporter groups’ in the State/Territory. In doing this, we adopted a similar approach for all jurisdictions, presenting the combined practice of four key groups who collectively make the vast majority of all reports by mandated reporters: police, teachers, doctors and nurses. These groups were chosen because generally they were: (1) clearly defined occupational/professional reporter groups; (2) groups who make a significant contribution to reporting; (3) groups designated as mandated reporters under the legislation; (4) designated as mandated reporters for the whole decade; and (5) data was individually available for analysis regarding these groups. For some jurisdictions (Qld, WA), one or more of the four groups had not always been mandated by law, but they were selected because they nevertheless made a large contribution to reporting practice. For four jurisdictions (ACT, NT, SA, Tas), one variation was necessary to this approach, including two additional groups (social workers and childcare workers) because they met the five conditions stated above.