ERROR THEORY
Please see our Risk Options section
This training can be offered as a one-day or two day course in seminar or standard training format. It is suitable for senior management, case conference chairs and all staff working with children who may need protection. It is not an easy course and is totally unsuitable for unqualified staff.
What
we are concerned with in Error Theory is the size of the ellipse in the
above diagram which outlines what is called "The
Fundamental Error" of incorrect action (false positive / false negative) in
any particular case.
The following (slightly amended) article on Error Theory appeared in the Greater London Post Qualification publication Contact magazine, No7, summer 2004. The Error List AS below has since been revised and extended as a working document
"There
are many studies regarding risk assessment in social work and also numerous
inquiry reports on fatalities that have had social
work input. However, as Dennis Howitt observed in his critical book (1992) Child
Abuse Errors: When Good Intentions Go Wrong,
social work has yet to develop an “Error Theory” to account for and assist
in predicting (wherever possible) what, where, when and why, errors or mistakes
of judgement might occur. So far as I am aware, all the social work books and
studies published to date on mistakes and errors fail to satisfactorily address
the issue of developing an “Error Theory”. This is not to say that the
authors of these
publications have not undertaken detailed analyses of mistakes and errors. This
has occurred and much pertinent advice disseminated along the way but mostly in
seeming ignorance of the absolutely enormous base of research literature on
decision and judgement
errors which exists outside of the social work arena. (1)
It
has long seemed passing strange to me that there has been hardly any attempt to
link the findings of social work studies of error
with the research on judgement and decision error outside of social work. There
may several reasons for this. One possible reason
has to do with alternative and conflicting perceptions on the nature of risk.
Another reason, not unconnected, has to do with the analysis and implications of
complexity within risk assessment tasks. The advent of the Framework for
Assessment within the childcare field
has enabled me to recently publish a formulation (2) of the astonishing
(potential) extent
of what this form of normative-based complex analysis technically demands
from field practitioners."
References:
(1) Social
Care Institute for Excellence (SCIE) publication (2005) Managing risk
and minimising mistakes in services to children and families
(2) Alan Cooper, Chapter 6: “Risk and the Framework for Assessment” in Calder MC and Hackett S (eds.) (2003) Assessment in Child care: Using and developing frameworks for practice. Dorset: Russell House Publishing.
The above was published in 2004
and my work on developing a Theory for Error within social care work has
continued and developed
since.
Also of relevance to how mistakes can occur within casework is my paper:
Alan Cooper: Game Theory and Partnership in Child Protection Practice published (2004) Vol 16, No 3, September BASW Practice Journal
 |
An amended Error section of Chapter 6 of the Calder and Hackett book referenced above is as below: |
Mistakes and Risk: Error Theory in Social Work
The literature on how people can make mistakes when confronting risk situations is vast. For some strange reason hardly any of this research is ever quoted in social work texts. And yet, as Stone (1992) and Reder et al (1993) noted, we keep making the same mistakes. So what does go wrong? Is it failures of individual competence or something much wider that is at issue? I would speculate that it is the latter. BA (degree) and much of post qualification training do not touch the deeper aspects of risk. I often find that qualified social workers cannot even give a basic definition of risk, let alone think about how probability and value judgements connect to risk and assessment. Social work educators neglect the deeper understanding of risk to the great peril of social work. There is much more to the assessment of risk than risk factor checklists which seem to constitute the sum total of what most students and social workers ever remember being taught both at Degree level and above.
And then there is the fact of non-existent risk management policies and the absence from many departments of effective workload management schemes in conjunction with staff shortages. Most important of all: I also believe that the sheer complexity of risk analysis has never been fully appreciated in social work. Howitt (1992), in his critical book on the child protection process, noted that social work has yet to develop an “Error Theory” in terms of the meta-level analysis of mistakes. Certainly, there have been a great abundance of inquiries. Despite the publication of several studies (DoH 1982, 1991, Reder et al 1993, Hagell 1998, Reder and Duncan, 1999) examining the lessons from these inquiries there has hardly been any attempt to link the findings of social work research with the enormous literature on judgement mistakes and risk errors outside of social work.
In my study of risk literature (both social work and outside of social work) I would submit that most mistakes uncovered by inquiries fall into the categories identified and listed below, although this listing is not presented as exhaustive. Furthermore, in principle, I believe that this list is as applicable to risk management in adult social work as it is to child protection work.
ORGANISATIONAL
ERRORS
v Procedural Error – policies and procedures not followed.
v Checklist Fixation Bias, including over reliance and fallacy of the base rate error.
v No Risk Management Policy: no guarantees for support, training or workload management within the context of working to policy and procedures. No senior management group that regularly monitor and record organisational risks.
v Inter-agency failures of communication: not sharing / comparing information.
v Intra-agency failures of communication: Adults and C&F not sharing / comparing information.
v Inter-agency Tensions / Role Confusion: can cause assessment paralysis.
v Intra-agency Tensions / Role Confusion: can cause assessment paralysis.
v Assessment Paralysis: groups cannot agree on problem or intervention strategy.
v Resource Tensions: the right intervention is not available or is deficient.
v Media Fear and paranoid risk adversity: lowers risk threshold – more CPR registrations and children in care.
v Management withdrawal / isolation: distance from staff and refusal to take case accountability – causes poor morale / staff tension – raises organisational risk.
v Management offers poor quality supervision: management inadequacy of knowledge and skills, uncaring attitude toward staff - raises organisational risk.
v Management / Staff Tension: poor morale, workplace stress and conflict.
v Workload stress and “burn out.” (associated with inadequate staffing and poor or non-existent workload management)
v Inappropriate allocation (inexperienced and inadequately trained staff)
v Inadequate knowledge and training on theory and research informing analysis and assessment
v Inadequate training on “what works” in terms of interventions and evaluation.
v Inadequate training on risk and risk management.
JUDGEMENT
ERRORS
v Non-Existence Error: (not seeing / talking to the child)
v Outcome (Goal) Error: non-specificity or confusion over framing outcomes
v Templating Error: mindset fixed on mistaken definite descriptions concerning a person’s behaviour, intentionality, circumstances or culture.
v Attribution Error: false ascription of intentionality
v Saliency Error: selecting from comparatively few factors – related to availability error
v Availability Error: false belief that information is complete and judgement is consequently narrowly based.
v Contrition Bias: believing the best in the face of evidence
v Conjunction Bias: taking a conjunction of separate circumstances as representative of greater likelihood of something being the case.
v Diminishment Bias: reducing the significance of important information
v Confirmation Bias: falsely believing that hypotheses should be confirmed and finding “confirmation” at every turn.
v Ratcheting Error: the upshot of confirmation error – the case is “magnified” or “diminished” out of all proportion and the reality cannot easily be recovered.
v Logical Reasoning Errors: e.g. Affirming the Consequent.
All of the above apply as much to Managers as to front line staff
There is insufficient space here to examine these errors in detail. Many of the items listed under “organisational errors” are almost self-explanatory. However, whilst most of these organisational factors work negatively to heighten the probability of things going dramatically wrong for the service user and the organisation, media fear tends to work against service users and their families because the organisation adopts a strongly “risk averse” philosophy and children come into the care system unnecessarily or unjustly. Indiscriminate use of checklists (checklist fixation) is associated with misplaced faith in their efficacy (Chris Beckett, 2001). This is what Pride (1987) and Howitt (1992) have railed against in their polemics against the child protection process.
The lack of organisational risk management policies that guarantee staff adequate training, and of applied and effective workload management systems together with organisational conditional support when things go right and when they go wrong is simply inadmissible. The non-existence of senior management groups which meet monthly (at least) to review levels of “organisational risk” (which should include reports from lower tier managers on levels of actual or probable judgement errors within case work practice) is something I find extraordinary. In a world where there is a universal expectation to “evidence” need, how else can there come about solid information for Councillors and Government on the need for additional human and training resources?
The tendency of management to avoid taking full and unequivocal accountability at all times, including when something goes wrong is hardly conducive to building staff morale. On a more positive note, the strong emphasis placed on effective “working together” by the DfES (2006) (**) will hopefully impact more and more progressively on all agencies involved in the assessment process. Sadly, it would seem, the DfES has forgotten intervention and its critical role in risk management. One would have thought that when such an august body as the Royal Society notes the importance of risk management and offers a definition:
“the making of decisions concerning risks and their subsequent implementation, … flows from risk estimation and risk evaluation,”
The reference to “implementation” would
automatically be taken to mean that it is continually necessary to assess and
evaluate risk during intervention. On this understanding, intervention cannot be
divorced from the assessment of risk at any time and should be part of the
initial assessment relative to the set of possible hypothesized outcomes. This
set is inclusive of the assessment of risk without any specific outcome-intended
intervention and with various possible hypothesised interventions.
Judgement errors are almost certain to occur
to most people at some time and in some place. They are therefore of enormous
importance to managers and practitioners. Nevertheless, to judge from my own
experience, few managers and practitioners have heard of them. Outcomes are
often framed in language that is not permissive of testing. Several fatal abuse
inquiries have shown that belief system “templating” has led directly to
signs and symptoms being ignored or “diminished” in significance. The child
“not existing” and therefore not seen is well documented in several
inquiries. Errors of false attribution (falsely believing something of someone);
errors of saliency (selecting too few factors for examination), errors of
availability (incomplete information falsely treated as complete when deciding
action), errors of contrition (wanting to believe the best in the face of
evidence) all seem to occur with regularity. Conjunction errors occur when
events which are non-connected are invested with causal connectedness. The error
of diminishment entails the “reduction” of the significance of information:
for example, the fatality of Lauren Wright (2001, Norfolk SSD) where the Team
Manager was reported to have unfortunately delayed an indicated urgently needed
visit because one had been planned for a few days time. The error of
confirmation bias is the tendency to look for “proof” of hypotheses instead
of refutation, which in turn can lead to “ratcheting” where the favoured
hypothesis is elevated to a state of unchallengeable sanctity. Lastly, logical
reasoning fallacies such as affirming the consequent are connected to checklist
fixation errors. For example, the “Anal Dilation” reasoning error of
Cleveland exemplifies this fallacy: Anal sex abuse causes anal reflex dilation
in some cases (statistical finding) - this is a legitimate conclusion from
observation. Anal reflex dilation (singular case) indicates sex abuse – this inference is not legitimate as the case may be a
false positive and the dilation caused by something else.
Most of these judgement errors are intimately
related to each other and psychologically driven. They derive from ignorance,
anxiety, fear, racism or uncertainty about working with people of other
cultures, but all occur within the context of the ethos and pressures of the
organisation.
In conclusion, my perception of risk errors and mistakes is that the organisation (including those providing Social Work Degree and PQ training programmes) should always bear the responsibility in terms of providing adequate training, adequate supervision and adequate environment.
(**) This has greater force with the advent of Every Child Matters and the Children Act 2004, Sections 10 and 11 of that Act and the publication of Working Together 2006.
Postscript:
A Theory of Error for Social Care Practice
- What exactly is a Theory of Error for social care?
- Does one exist already?
- If not, can one be formulated?
Theories tend to be (at their best) both explanatory and predictive. For example, the Theory of Evolution is a very powerful explanatory theory but is less good at prediction because the time scales upon which it operates are way outside human experience. The Special Theory of Relativity is powerfully predictive of relative motion but is poor in terms of explaining why things are as they are unless extended into the General Theory of Relativity.
There is no Theory of Error for Social Care as I pointed out in my essay (fragment above) of Risk and the Framework for Assessment. But the sociologist Dennis Howitt had noted this in 1992. But on the other hand, there is no shortage of Inquiry Reports and Serious Case Reviews all of which purport to discover “what went wrong and how”. For the most part, these reports fulfil this vital function and we have learned a great deal over the years on what goes wrong and how mistakes happen. But all these reports and most of the collective studies (excellent though they are) on child death cases seem to draw out only limited pointers for practice rather than attempt to construct a framework (theory) as to why mistakes and errors happen and why, despite everything that is done to shore up and improve practice, they keep happening.
My own efforts since 2002 in trying to get to grips with constructing a Theory of Error for social care practice have mostly fallen on completely deaf ears: there is virtually zero interest in this matter from social care providers and social care educators in the U.K. I find this rather strange but it is of a piece with the almost zero interest shown by social work in the vast research and evidence base on judgement and error theory in the fields of psychology and economics. Here and there a lone soul or two makes reference to these research findings in the social work literature and then the whole ship goes happily on its merry way while the mistakes keep occurring. It is all very odd and is almost enough to drive desperate souls like me into conspiracy theory! But no one ever does get it wrong until they get it wrong – do they? And who needs a theory then?
The approach I have adopted toward formulating a Theory of Error starts from various presumptions and then examining as a First Stage the many studies and reviews already available on child deaths. I have not extended this work to mistakes in mental health or adult care but I believe that the Second Stage of this work applies equally across the board to all aspects of social care.
First Stage: First principles and underlying (boot strap) theoretical structures
Second Stage: comprises translating identified mistakes in the social work literature into the categories of human judgement and organisational error identified in the Human Error Theory literature wherever possible or adding categories where required.
Third Stage: analysing causative paths associated with Judgement and Organisational Errors. This stage is a sort of half way house as it involves (usually) an inter-relation between human judgement error and organisational error.
Fifth Stage: enumerating layers of organisational hierarchy each layer having its own potential for error.
Seventh Stage: goes with the Sixth stage re discerning causative linkages between the Stages.
Eighth Stage: completed theory
So far my work has produced results for Stages 1 through to Stage 7 but more work needs to be completed and I have not yet produced Stage 8 in a form which is in any sense complete. However, I still believe that this work has implications for social care and especially for Child Death Panels and Serious Case Reviews.
See our Pages on Risk Dimensions, Probability, Judgement, Outcome formulation and Decision-making
