Dealing with uncertainty and
complexity in research and
To what extent is it possible and
feasible to formalise social work
• The key features of social work
• Reasons to turn to science and formalisation
– Good and bad
• The generalisability of research
• The contribution of research to decision making
• What else should we be doing?
What is social work?
• A human service aiming to resolve human
• Takes place within a human interaction – a
• Seeks to find solutions that fit the recipient’s
world – not imposing our values on less
Strategies for increasing
• Decision making tools
• All represent a shift to more formalisation of
practice and reduction in individual judgment
and decision making
REASONS TO TURN TO SCIENCE
GOOD AND BAD-
Good reasons for this
• Some confidence in the findings of scientific
research so share the learning
• Standardising practice, greater equity
• Reducing scope for personal bias, e.g. values,
Bad reasons for this
• Laziness – over-simplifying
• Ignorance – not understanding social work
• Cheapness – less need for
training/supervision, just check if have
followed the rules
• Fear – of taking responsibility for judgments-
‘it’s not my fault, I was just following the rules’
Objectivity: More layers of
meaning than a mille-feuille’
Daston & Gallison (2010)
* Ontological claim:
– An objective world of particulars independent of
• Character claim:
– Impartiality, detachment, disinterestedness, and a
willingness to submit to evidence
• Epistemological claim:
– beliefs, judgments, or products of thought about what is
really the case
Three distinct kinds of claim in ‘this
 It works somewhere.
This is the kind of claim we can clinch with objective methods
like RCTs; the kind of verdict about a programme that a good
post hoc evaluation can deliver.
 It works.
Means: ‘It works almost everywhere’, or at least ‘widely’, or
perhaps ‘everywhere, other things being equal’.
 It will work here.
This is what we want to know when we deliberate about
whether to adopt a policy.
Does EBP avoid value bias?
• Choice of question
– Maltreatment arises from multiple factors, some
structural, some individual
– Where to focus change effort?
– How to define ‘success’ – a value
• Choice of method
• Definition of terms – social construction
Does EBP remove the personal?
• Aspire to empower the recipient – respecting
their wishes and feelings
• Significance of worker variables – people
respond to specific individual
Does EBP achieve greater equity?
• But ….
– In practice, most interventions include scope for
adaptation to recipient
– Recipients respond differently so receive different
THE CONTRIBUTION OF RESEARCH
TO DECISION MAKING
• Decision tools in child protection build on
research evidence about frequency of
variables in pathways to harm
• But rely on cases known to child protection
• And on data recorded by them
How much confidence should you have in the
outcome of an actuarial prediction?
Need to know:
• Sensitivity – true positives
• Specificity – true negatives
• Base rate – prevalence in the population
articulating, checking, receiving additional
ideas, being challenged,
Improving feedback – what happened after
your judgement or decision?
• In a complex world, we need ‘requisite variety
• The unique worker does matter
• Formal methods all entail some intuitive
• We need to work on improving intuitive
Cartwright, N. & Hardie, J. (2012) Evidence based policy, A practical guide to
doing it better, Oxford, Oxford University Press.
Deaton, A. & Cartwright, N. (2106) Understanding and misunderstanding
randomized, controlled trials. National Bureau of Economic Research,NBER
Working Paper No. 22595
Munro E., Cartwright N., Montuschi E. & Hardie J. (forthcoming) Improving
child safety: judgment, expertise and using research, Durham University. Email
email@example.com for details of publication