Lies, damned lies and child protection statistics

Lies, damned lies and statistics: so the famous saying goes. The problem is, in the counting of social phenomenon (as opposed to physical entities), the way we choose to count things always reflects underpinning social processes rather than objectively verifiable realities. So, the issue is not so much a matter of calling out ‘lies’, but one of discerning the social priorities and concepts driving the categorisation processes used to sort the things at hand.

The recent furore over the statistics regarding emotional abuse and neglect included in the Salvation Army’s State of the Nation report (and the subsequent media reporting) is a fine example of this predicament. The socially defined nature of child abuse makes measuring its incidence far from an exact science: one person’s ‘poor parenting’ is another person’s abuse, and if those two people are from different cultures, different periods in time, or different professional disciplines, their perceptions can differ immensely. Clarifying exactly how certain definitions will be applied to actual practice situations is often a matter of debate. This doesn’t necessarily signal intentional obfuscation, but the Salvation Army is correct in their view that there ought to be transparency and consistency in our decision-making criteria.

In addition to the fundamental conceptual problems with defining child abuse, there are organisational and macro-level factors that can influence how rates appear in child protection statistics. For example, international research shows that the presence of NGOs in particular geographical locations can influence whether cases are substantiated and progress to further legal intervention or not: NGOs can provide social workers with alternatives to statutory pathways for ‘cusp’ cases that sit on the edge of the statutory threshold. If statutory services can rely on an NGO to support and monitor the family, substantiation and legal interventions may not be pursued. In geographical areas where NGO provision is absent, statutory interventions may be the only option. Uneven provision of support services means that similar cases can be treated differently, with one case ‘substantiated’ (and counted in the statistics) while the other is not. That is why the purpose of any child protection agency, and its interactions with other organisations, must be known in order to understand the extent to which their stats can be considered a true picture of incidence, and to determine what other factors might be driving changing statistical outputs.

As the Salvation Army report points out, an interesting pattern in the last year or so is that, although notifications continue to rise, every other decision point, including substantiation, has dropped. Why is this? Is it due to changing practices around decision points, smarter or more efficient means of sifting out those that don’t require CYF intervention, or is it that the threshold for CYF services is being pushed higher? If so, is that because the presence of solid NGO services or Children’s Teams are providing a better alternative to CYF intervention, or is it because, in an environment of static resources, CYF is simply raising its threshold in line with what they are realistically able to manage?

Lets look at the numbers. Notifications come in two main kinds: general care and protection and police family violence. The first has dropped a little, from 88, 768 in 2014 to 83, 871 in 2015; while police family violence notifications continue to rise: from 57,889 in 2014 to 67,034 in 2015 (an increase of 15%). Overall, the combined rate of notifications in 2015 was 150,905 compared to 146,657 for the previous year (a 2.8% increase).
Every decision point after notification within the CYF system has dropped, some significantly, while the number of children in care remains stable. For example, notifications requiring further action, as a percentage of cases notified, has dropped from 61% to 54%. The real reason for these drops at each point can be found in the final stats relating to types of abuse substantiated. Within this substantiated group, the rates of physical and sexual abuse are constant, but findings of neglect and emotional abuse have both dropped significantly (by around 21% each).

Substantiated abuse findings by finding type
Type of abuse finding F2011   F2012   F2013   F2014   F2015
Emotional Abuse         12,711 12,454  12,777   10,406   8,318
Physical Abuse               3,253  3,330    3,343     3,305    3,235
Sexual Abuse                1,514    1,418   1,459      1,329    1,275
Neglect                          4,813    4,970    5,405      4,583   3,644
Total abuse findings  22,291  22,172  22,984   19,623  16,472

What does this mean? Is CYF simply ignoring cases of emotional abuse and neglect, or are they getting better at sifting out cases that don’t require legal intervention? Historically, several things have been happening. Firstly, notifications to CYF have increased astronomically in a 15 year period, roughly ten times now what they were then. This is a similar pattern to other Anglophone coutries, and has led in other jurisdictions to attempts to lower the rates of notification, and improve triaging cases to maintain a particular threshold. For example, the millions (yes millions) of notifications in the US have led some states to engage in large-scale education campaigns with all potential notifiers to stop them referring cases that do not require statutory intervention. Pushed up by mandatory reporting, within a litigious culture, notifications for any possible case of abuse resulted in a system that was overwhelmed with notifications that didn’t need legal intervention. Other states simply refuse to accept any notifications for domestic violence (these make up one third of New Zealand notification rates and are still increasing) or notifications where the only issue is parental substance abuse.

In New Zealand we had very similar issues with a notification rate that was threatening to overwhelm the system. This was one factor prompting the workload review in 2014, and it seems one of the outcomes of the review was a tightening of the criteria used to define emotional abuse and neglect. This change has filtered through in the numbers of substantiated cases, particularly with reference to family violence automatic notifications, where it seems CYF is exercising greater discretion about which cases require a legal reponse from them (bearing in mind that many different situations are classified as domestic violence and all are automatically notified to CYF). As is also common in other countries, the ongoing ambivalence in policy and practice towards domestic violence in child welfare contexts, which pushes up the notifications immensely, is evident (Hester, 2011). Many such cases are closed by CYF because their response tools – statutory intervention on children’s care – are not always well suited to the problem cause of inter-parental violence.

Having clear and consistent thresholds is important, after all, if a child in one place is deemed ‘abused’, and another in similar circumstances is not, this is an important justice issue for both parents (rights intruded on unnecessarily) and children (who may not get the protection they deserve). So, improving consistency is a good thing, and the CYF changes reflect attempts at achieving greater consistency in line with a clearer definition of their role. Another positive is that children for whom the solution that CYF offers (statutory intervention) is not appropriate, are being referred elsewhere. CYF is essentially a response organisation. Whether you think that is their correct role or not is a different question, but it is how their current role is configured. To ask them to engage with families that are high need but not chronically harmful is like asking life guards to teach children to swim: it’s not their job.

The bigger question is this: since we have defined the CYF role in this way over the last 20 years, are the other services charged with picking up the preventive role adequately resourced to meet the needs of those turned away by CYF? Here we have a problem, as family poverty, poor housing, and lack of access to adult mental health services can all have huge impacts on parenting, yet the provision of these essentials is often left wanting. We have a decent range of primary health and some lower level targetted family support services, but high level intensive family preservation services are simply not available. As CYF become clearer about their decision thresholds, and define their role more clearly as a legal intervention only service, who picks up the other services needed for prevention? Children’s teams are working hard but, without adequate resourcing of direct services, this seem destined to a very tough road indeed. We can only hope the CYF review will look beyond Child Youth and Family to attend to the broader child welfare system – from basic universal provisions through to specialised services.

Emily Keddell


Hester, M. (2011). The three planet model: Towards an understanding of contradictions in approaches to women and children’s safety in contexts of domestic violence. British Journal of Social Work, 41(5), 837-853. doi:10.1093/bjsw/bcr095

[fac_icon icon=”camera-retro”] Image Credit | Simon Cunningham

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: