>> CLICK HERE TO ORDER 100% ORIGINAL PAPERS FROM AustralianExpertWriters.com <<

Oxfordshire Social Services

Br. J. Social Wk. (1990) 20, 525-546Allocating Blame in Social WorkGERALDINE MACDONALDGeraldine Macdonald was previously a social worker with Oxfordshire Social Services andis now a Lecturer in Applied Social Studies at Royal Holloway and Bedford New College,University of London. She is co-author (with B. L. Hudson) of Behavioural Social Work:an Introduction (1986).SUMMARYThis article takes issue with those who assume that the responsibility for bad outcomes insocial work, such as child deaths, is appropriately laid at the feet of individual workers. Itexamines the philosophical origins of such arguments, some recent applications withinsocial work literature and their appropriateness to the realities of social work practice. Theauthor argues that a morality of social work must recognize the social and organizationalcontext in which it occurs.The competence of social workers has been called into question by themedia, the Chairpersons of Inquiries, other professions and people ofstanding within social work. Central to many of the indictments made isthe assumption that the ultimate measure of competence is that ofparticular outcomes, and that the responsibility for such outcomes canbe traced back to individuals, or groups of individuals, with whom it isrightly placed. Accordingly, a bad outcome is presumed to indicate abad decision or negligent action.The significance of particular outcomes permeates many of theReports of Panels of Inquiry into the deaths of children who havesuffered deliberate harm at the hands of their parents. Even whenwriters expressly distance themselves from a misuse of hindsight, theyrapidly move to an appeal to outcomes in their adjudication of humanaction, as in this passage from the Report of the Panel of Inquiry intothe death of Jasmine Beckford:We are entitled to judge a person’s actions by reference to what was and should,reasonably, have been in his or her mind at the relevant time. We are notentitled to blame him or her for not knowing, or foreseeing what a reasonableCorrespondence to Geraldine Macdonald, Applied Social Studies, Royal Holloway andBedford New College, Egham Hill, Surrey TW20 OEX.0045-3102/90 $3.00 © 1990 British Association of Social WorkersDownloaded from https://academic.oup.com/bjsw/article-abstract/20/6/525/1657327 by rmit3 user on 25 June 2019526 GERALDINE MACDONALDperson would neither have known or foreseen. In assessing whether a reasonable person would have known or foreseen an event, we are entitled to haveregard to what actually happened, though, of course, the fact that an eventoccurred does not mean that a reasonable person would necessarily have knownthat it would occur or would have foreseen its occurrence. But the fact that it didoccur (and was not an Act of God but the result of human action or inaction)gives rise to a presumption—either that there was knowledge that it wouldoccur, or that foresight would have indicated its likely occurrence. A Child inTrust, 1985, p. 32 (emphasis added).Rarely has attention been given to the possibility that certain courses ofaction might be ‘right’ or ‘wrong’ despite rather than because of theiroutcome.This article argues that for professional activities such as social work,the allocation of responsibility or blame for decisions on the basis ofindividual bad outcomes is misguided. What is needed is a moralvocabulary appropriate to the task which recognizes the corporatenature of many social work responsibilities and their execution, and theknowledge-base required. In order to develop this position I propose toanalyse briefly four major forms of the proposition that the tightness orwrongness of actions is, and should be, determined by outcomes.A GENERAL CONSENSUS?Consider the following:A four year old child, known to Social Services, dies at the hand of her stepfather. A post-mortem reveals signs of old fractures and multiple bruising. She isunderweight.How do we react as workers? how does the public react? the media?With an impressive degree of consensus, it would seem, although thetone of the responses differs. The social worker and his colleagues madethe wrong decision. How could there be any doubt?Indeed, on first reflection that judgement seems a reasonable one.One of the tasks of social workers is to protect children and this childdied. Whether the decision was to return the child following a period incare, not to intervene after a phone call from a concerned neighbour, orthat weekly visits from the social worker and health visitor would act asa preventive and therapeutic measure, is immaterial: that the child diedis generally deemed to be evidence that a wrong decision was made.Such a judgement is the application of a common everyday response;from gardening to matrimony, we readily read bad outcomes asevidence of bad decisions.Downloaded from https://academic.oup.com/bjsw/article-abstract/20/6/525/1657327 by rmit3 user on 25 June 2019ALLOCATING BLAME IN SOCIAL WORK 527Further reflection might incline you to the view, with which I concur,that poor outcomes are neither necessary nor sufficient indicators ofwrong decisions. Bad outcomes may signal the need to review pertinentdecisions, but the criteria for assessment should not be the outcomes,but the appropriateness of decisions given the information available atthe time, together with such factors as the cost of getting additionalinformation; the quality of the information, whether it was the best thatcould be obtained at the time, and so on. This is not to say that outcomes do not matter; it is from aggregate outcomes that we (should)derive the criteria according to which our actions are appropriatelyevaluated.1But acceptance of this view is not straightforward. There exists a bodyof recent philosophical argument (serious argument by serious philosophers) which, to put it at its simplest, could be read as claiming thatthe initial response of the man-in-the-street captures an inescapabletruth about certain human actions. If so, and if aptly applied to socialwork, then we are seriously at risk—not from judgements of the tabloidpress, but from our self-appraisals.This risk has been most forcefully expressed by Hollis and Howe in a1987 paper entitled ‘Moral risks in social work’ (and more dramaticallyintheir 1986 popularized version ‘Child Death: Why Social WorkersAre Responsible’). In their opening sections they introduce work on‘moral luck’, and the unphilosophical reader would be forgiven forforming the impression that therein lay a (philosophically) convincingargument that certain actions, not dissimilar to social work, could beassessed only by their consequences. So I also begin with ‘moral luck’;partly because Hollis and Howe reject it for the wrong reasons, andpartly because to see exactly why that argument—whatever its peculiarmerits—is irrelevant to social work helps clarify the nature of socialwork action.MORAL LUCKIn a philosophically influential paper (1976) Williams has argued thatthere are some categories of action whose outcomes are so unpredictable and indeterminable that their justification—that is, judgementsabout their lightness or wrongness—can only stem from the particularoutcomes they produce. Williams constructs a hypothetical example of a. . . creative artist who runs away from definite and pressing human claims on‘This issue is distinct from the relevance of particular outcomes, and is one which I addresslater.Downloaded from https://academic.oup.com/bjsw/article-abstract/20/6/525/1657327 by rmit3 user on 25 June 2019528 GERALDINE MACDONALDhim in order to live a life in which, as he supposes, he can pursue his art(Williams, 1976, p. 22).He calls this hypothetical artist, Gauguin. Williams’s Gauguin is not aman who simply does not care about the claims of others, but one who:is concerned about these claims and what is involved in their being neglected(we may suppose this to be grim), and that he nevertheless, in the face of that,opts for the other life . . . which (for the sake of argument, he sees as) . . . a lifewhich will enable him really to be a painter (Williams, 1976, p. 23).For someone in Gauguin’s situation, he can only know that he has madethe right decision when (and if) he becomes a great painter; his successor failure holds the key to the evaluation of his action, which can,therefore, only be ‘retrospective’. Williams argues that until the outcome is known it is not possible to evaluate the decision: one cannotformulate a moral rule that will justify the decision to pursue art andignore the prior claims of others. Such a rule would read:one is morally justified in deciding to neglect other claims if one is a greatcreative artist (Williams, 1976, p. 24).And riders such as ‘If one believes oneself to be a great artist. . . .’ onlyserve to underline the absurdity of such an attempt at prior or contemporaneous justification.Williams distinguishes between two types of failure. If Gauguindrowns on his way to Tahiti, then his enterprise fails, but would notprovoke the thought that after all he was wrong and unjustified. Not so,if his paintings were failures.This distinction shows that while Gauguin’s justification is in some ways a matterof luck, it is not equally a matter of all kinds of luck (Williams, 1976, p. 25).Some luck, in a decision of Gauguin’s kind, is extrinsic to his project(outside his control), some intrinsic (is he, in fact, a gifted painter); bothare necessary for success, but only the latter relates to unjustification.Williams’s argument therefore has two components: first, someactions can only be justified by their outcomes and second, such actionscan only be unjustified (deemed wrong) by the intrinsic failure of theactor concerned to secure his intended outcome. Now many, if not all,decisions taken by social workers are taken on the basis of hoped for,even anticipated, but uncertain outcomes: child protection, fosteringand adoption, social care planning, to name but a few. But does thisplace them in the category of actions which can only be judged by theiroutcomes?In the context of Williams’s analysis the answers are simple. Williamsis concerned with decisions whose outcomes determine the very perception of the person concerned:Downloaded from https://academic.oup.com/bjsw/article-abstract/20/6/525/1657327 by rmit3 user on 25 June 2019ALLOCATING BLAME IN SOCIAL WORK 529In these cases, the project in the interests of which the decision is made is onewith which the agent is identified in such a way that if it succeeds, his stand-pointof assessment will be from a life which then derives an important part of itssignificance for him from that very fact; if he fails, it can, necessarily, have nosuch significance in his life. If he succeeds, it cannot be that while welcoming theoutcome he more basically regrets the decision (Williams, 1976, p. 35).This identity of actor and evaluator is the logical pivot of Williams’sargument. But for social workers and the decisions they routinely make,the analogy is a mirror-reflection. They are not making decisions thesuccesses of which affect their personal lives in any very direct or dramatic way, not so the ‘failures’. Further, Williams’s use of the termjustification is essentially about se//-justification. This is clearly inapplicable to an activity which is essentially interventionist and to agentswho, by definition, are not there to justify themselves, but to secureoptimum outcomes for others. Whatever the force of Williams’s arguments in their targeted sphere, they do not address the nature of socialwork action.OUTCOME AND CULPABILITYNagel (1979) bases his argument that outcomes determine culpabilitywithout recourse to moves such as the ‘standpoint of assessment’. Hisargument has, at first sight, an almost de facto, empirical basis:Actual results influence capability or esteem in a large class of unquestionablyethical cases ranging from negligence through political choice (Nagel, 1979, p.30).That these are genuine moral judgements . . . is evident from the fact that onecan say in advance how the moral verdict will depend on the results. If onenegligently leaves the bath running with the baby in it, one will realize, as onebounds up the stairs toward the bathroom, that if the baby has drowned one hasdone something awful, whereas if it has not one has merely been careless(Nagel, 1979, p. 30).Nagel’s point is that the behaviour—leaving the baby unattended—isthe same, but the evaluation will depend on the outcome, and we cansay which sorts of outcomes will result in which sorts of evaluation. Ifyou think no baby should be left unattended in this way, consider thecase of the child travelling unrestrained in a car; or the child left unattended while mother hangs the washing on the line. Most of us engage inactions with possible, albeit unlikely outcomes which, if we knew inadvance they would occur, would cause us to act differently. Nagel’s useof moral luck is anchored in the ways in which identical actions can haveDownloaded from https://academic.oup.com/bjsw/article-abstract/20/6/525/1657327 by rmit3 user on 25 June 2019530 GERALDINE MACDONALDdissimilar outcomes due to events outside of the control of the personconcerned:The driver [who accidentally runs over a child], if he is entirely without fault,will feel terrible about his role in the event, but will not have to reproachhimself. Therefore this example of agent-regret is not yet a case of moral badluck. However, if the driver was guilty of even a minor degree of negligence—failing to have his brakes checked recently, for example—then if that negligencecontributes to the death of this child, he will not merely feel terrible. He willblame himself for the death. And what makes this a case of moral bad luck isthat he would have to blame himself only slightly for the negligence itself if nosituation arose which required him to brake suddenly and violently to avoidhitting a child. Yet the negligence is the same in both cases and the driver has nocontrol over whether a child will run into his path (Nagel, 1979, pp. 28-9).Nagel’s intuitive position receives further support when one considersthat although some might wish to argue that all acts with potentiallynegative or harmful outcomes are equally culpable, the number of occasions on which society deems it appropriate to respond accordingly, isgenerally small. For example, the driver who has not recently checkedher brakes but drives without incident is typically not dealt with asseverely as a similar driver who, unable to stop in time, kills a child.There is an obvious pragmatic reason, there would be more courtappearances than the courts could handle, the problem of detectionaside; but Nagel might also argue that such a situation would transgressour generally accepted understanding of moralityHowever jewel-like the good will may be in its own right, there is a morallysignificant difference between rescuing someone from a burning building anddropping him from a twelfth-storey window while trying to rescue him. Similarlythere is a morally significant difference between reckless driving andmanslaughter (Nagel, 1979, p. 25).Nagel’s argument then rests on two claims:a) that we share an intuition that the acts he depicts (identical except intheir outcome) are morally distinct, andb) that maintaining such a distinction is socially viable—it is a good wayto arrange human affairs.We shall consider the second claim later, with regard to the context ofsocial work. Concerning the first, one might simply choose to disagree;but this falls short of a counter-argument.Alternatively, one might query the grounds on which Nagel’s intuition is shared. Might it not be that the person persuaded by Nagel’s‘baby in the bath’ example, is so persuaded because he feels the parentof the drowned child has in fact been more careless (left it longer, wentfurther out of earshot)? If so, then far from sharing Nagel’s intuitionDownloaded from https://academic.oup.com/bjsw/article-abstract/20/6/525/1657327 by rmit3 user on 25 June 2019ALLOCATING BLAME IN SOCIAL WORK 53I{same action, different outcomes), the assentient is invoking the outcome as a surrogate for a (in principle, directly observable) difference inthe action {different action, different outcome). It is possibly moreappropriate to regard Nagel’s ‘intuitions’ as modifiable psychologicalresponses, than as a defensible basis for moral assessment.Before turning to Nagel’s second claim, which is where the seriousimplications for social work are most manifest, let us consider a movewhich may short-circuit the discussion. Suppose that Nagel is dismissedon the grounds that his argument really only holds sway in circumstances when people have behaved negligently, and you do not holdyourself to be in this category. Is it still appropriate to hold you responsible for bad outcomes when you have dotted every i and crossed everyt? Hollis and Howe (1987) think it is.PREDICTABILITY; OCCUPATIONAL CHOICE ANDDUTYHollis and Howe address themselves specifically to social work. Theyrightly reject the suggestion that social work decisions should be judgedaccording to concepts of moral luck, as defined by Williams or Nagel,although their reasons for rejection are not persuasive. However, theystill wish to argue that individual social work decisions should be evaluated according to their individual outcomes. Their argument rests onfour main moves: first, deaths are preventable and predictable, andaccordingly represent culpable failures; second, social workers choosetheir profession and in becoming social workers accept the responsibilityfor outcomes, some of which will be bad; third, social workers have aduty to prevent death and protect children, and when they fail, theresponsibility rests with the individual; and finally, underpinning allthese, is the claim that the social worker is the effective decisionmaker—responsibility shared is not responsibility shed.PREDICTABILITYIn rejecting moral luck, Hollis and Howe go to the opposite extreme,suggesting that, far from being actions with unpredictable or accidentaloutcomes, social work decisions concern predictable outcomes. Withregard to the most difficult area of child protection, they state:Think of her [the social worker] as deciding in which of two categories the childbelongs. Category A comprises children so much at risk that they should all beremoved to a place of safety; category B comprises those who will be safe, if leftat home. The child’s death proves that it belonged in category A. Must she notDownloaded from https://academic.oup.com/bjsw/article-abstract/20/6/525/1657327 by rmit3 user on 25 June 2019532 GERALDINE MACDONALDhave been incompetent in assigning it to category B? (Hollis and Howe, 1987, p.130).In other words, children are either at risk, or they are not. But this is amisdescription of the predictive task. Category A (‘all should beremoved’) must refer to cases where the estimated probability of harmis so high as to justify removal, despite the associated costs of removal.Category B, on their definition, consists of children with a zero probability of abuse (for they say ‘the child’s death proves that it belonged incategory A [i.e. not in category B]’). These cannot, on any plausibleview of the world, be held to be exhaustive categories. We must, minimally, introduce a Category C where the probability of abuse is nonzero but still so low as not to justify costly intervention. To assert ‘yes,there is a possibility that this child might be abused, but the probabilityis so low, and the costs of removal (costs to the child, that is) so high,that it is inappropriate to institute care proceedings’ is scarcely selfevidently absurd. If so, then, even on Hollis and Howe’s analysis, deathindicates, not ‘membership of category A’ but ‘membership of A or C.Such an outcome is tragic; but the death itself cannot be even presumptive evidence of incompetence (which Hollis and Howe use interchangeably with responsibility); it may have resulted from a competentallocation to Category C.Hollis and Howe mask the complexity of decision-making in this areaby focusing on fatal outcomes. Many children give cause for concernbecause of generally poor parenting which might or might not includethe risk of physical abuse. Sometimes social workers have serious concerns about the welfare and safety of children, but have little evidenceon which to intervene: predictability and demonstrability are notsynonymous. And where the child’s physical safety is not the primeconcern (that is, where the child’s life is not deemed to be at risk) thedecisions are not just about protection from less than satisfactoryparents, but about when parenting becomes so inadequate or damagingthat the hurt and trauma of removal becomes justified. This is the realityof the Category C cases, and it is here that mistakes are, perhaps, mostlikely to occur, although not necessarily mistakes that culminate in theloss of life. But it is only for B, not C, that a negative outcome is proofof an inappropriate allocation.OCCUPATIONAL CHOICEHollis and Howe (1987) acknowledge the ‘uncertain grounds for deciding’ (p. 125) in the majority of the 100,000 referrals received by Britishwelfare agencies. However, recognizing that inadequate levels of knowledge, predictive finesse and skill in the area of child protection conDownloaded from https://academic.oup.com/bjsw/article-abstract/20/6/525/1657327 by rmit3 user on 25 June 2019ALLOCATING BLAME IN SOCIAL WORK 533stitute one source of inevitable errors, they argue that the social workeris as much responsible for bad outcomes beyond her control as thebomber pilot whose brief was not to kill civilians but who does sobecause his equipment was faulty:Our point is simply that to accept the social worker’s role and to intervene on itsauthority is to accept moral responsibility for what happens (Hollis and Howe,1987, p. 131).She became a social worker by her own choice and, even if her predicamentdawned on her only in midstream, she chose to continue. She has accepted theconstraints and is making judgements which would reconcile the conflicts, only ifthey displayed a professional accuracy which, in the present state of knowledge,they cannot possess (Hollis and Howe, 1987, p. 132).Aside from doubting whether any social worker formally makes such anundertaking (there is no social work equivalent to the HippocraticOath) it does not follow that acceptance of a difficult job is synonymouswith accepting moral responsibility for outcomes beyond one’s control,any more than a car mechanic could be deemed morally responsible forservicing a car and failing to notice a fault which he either had no reasonto be looking for, or which was, at the time of service, undetectable.Further, it requires an unduly simplified view of social work action—tothis I shall return.DUTYHollis and Howe’s second move rests on their analysis of the duty ofsocial workers; they present a conflict facing social workers from the‘conflicting imperatives of welfare and justice’ (p. 131). Whilst socialworkers are expected to intervene and protect children, and areauthorized to do so, they are not permitted to infringe the rights ofparents and families.The traditions might harmonise if she could be sure of intervening in all and onlythose cases where the child would otherwise be prevented from developing intoan autonomous adult. But we have given ample reason not to expect a perfectfit. So she will fall foul of one inheritance or the other from time to time (Hollisand Howe, 1987, p. 129).These conflicting demands form a second source of inevitable error(Pine, 1987, offers a similar analysis in the American context). Does thisinevitability relieve the social worker of her moral responsibility? Hollisand Howe remain adamant: it does not. This conclusion rests on theirassumption that it is the social worker’s duty to protect children, a dutywhich sometimes comes into conflict with other duties such as takinginto account the rights of parents, but which, in terms of her statutoryDownloaded from https://academic.oup.com/bjsw/article-abstract/20/6/525/1657327 by rmit3 user on 25 June 2019534 GERALDINE MACDONALDbrief, has primacy (she does not have the statutory brief to promote therights of parents). It is this duty that she has ‘taken on’ and for whichshe has accepted responsibility. They again make use of the ‘dirtyhands’ analogy, which they here describe as the invalidity of dutiesderived from one source (justice) being allowed to override another(welfare) and excusing the agent concerned.But this is a muddled and unhelpful use of the term duty. Whilst itpossibly makes sense to see X as having a duty to drive carefully or totell the truth, both actions being within X’s control, the duty ‘not to bemisunderstood’ or ‘not to have an accident’ poses a peculiar obligationwhich no individual could guarantee to deliver. Certainly we can have aduty to act so as to minimize accidents and incomprehensions, andcolloquially may collapse this to lacking accidents and misunderstandings. But the coalescence can only be colloquial; strictly the duty is aduty to endeavour, since the outcome is contingent upon the actions ofothers.Further, the ‘dirty hands’ analogy is formally erroneous. The reasonthe prison officer or the soldier cannot appeal to orders to absolvehimself from responsibility lies in the judgement that no moral agentwould accept or approve the aims of a corrupt organization. Statutorysocial work agencies, in contrast, exist because there is a general consensus2 that their role is desirable and/or necessary. Their allotted tasks(the ‘orders’) are, similarly, not comparable.THE CONTEXT OF SOCIAL WORK ACTIONIf the world is so arranged that errors are inevitable, it is, minimally,illogical to hold individuals morally responsible when there is no causallink between what they do or fail to do and bad outcomes. Hollis andHowe compound their error by dismissing the idea that the multi-disciplinary nature of decision-making in social work absolves the socialworker of any responsibility, a claim considered in more detail below.Their dismissal of the organizational context of social work and socialwork decision making is especially striking when one considers othercontributions made to the literature by Howe (1986a, 1986b), and hisfamiliarity with social work practice in local authority settings. Considerthe emphasis of this quotation:Thus, the social worker’s understanding of individual cases and the responsesadmittedly exist a variety of perspectives from which the enterprise of statutorysocial work is seen as, in varying ways, suspect. But these do not form the underpinning ofHollis and Howe’s argument; and, if accepted, they pose wider dilemmas for socialworkers than the assumption of responsibility for particular outcomes.Downloaded from https://academic.oup.com/bjsw/article-abstract/20/6/525/1657327 by rmit3 user on 25 June 2019ALLOCATING BLAME IN SOCIAL WORK 535she makes are outcomes of ‘wider social processes’; structural developmentsand the role of the state (Howe, 1986a, p. 138).Part of Howe’s concern elsewhere (19866) is to determine the relationship between what social workers do and the structural arrangements tohelp them do it; and part of his answer is thatSocial work has no essential nature. Its existence, its characteristics and style area product of time, place and the balance of power between those occupationsinterested in tackling certain types of behaviour and social concerns (Howe,19866, p. 160).This sensitive awareness of ‘time, place, and the balance of power’ fitsill with a simplistic view of the social worker as the sole decider. Pinker’sdissenting note to the Barclay report captures well the constraints onaction:It is more practical to define the role and tasks of social work in terms of therange of activities which social workers undertake in response to certain institutional imperatives. The first set of imperatives derives from the range andvariety of the needs that confront social workers. The second is contained in thelegislation that defines their mandatory Local Authority obligations and alsoindicates the range of their permissive undertakings. The third relates to theemployers . . . to whom social workers are accountable, who are themselvesaccountable in the discharge of their mandatory and permissive undertakings(Pinker, 1982, p. 237).Pinker’s insight rests uneasily with the argument of Hollis and Howe;but it is quoted, with approval, by Howe (19866, p. 161) in a volume inwhich he seeks ‘to understand the practice of social work by examiningthe ways in which fieldworkers and their work are organised in SocialServices Departments’ (p. 1).Disattention to context and to other sources of complexity also matter. Hollis and Howe fail to differentiate between categories of ‘highrisk’, some of which are more susceptible to successful intervention thanothers. Similarly, it is not surprising that Inquiries such as those chairedby Louis Blom-Cooper, also concerned with fatalities, feel able to cometo ‘simple’ conclusions:Dr Kempe, the author of the concept, the battered-child syndrome, once wrote‘If a child is not safe at home, he cannot be protected by case work.’Indeed, we have come to the firm conclusion that once a high risk situation isidentified by social services and is endorsed by a judicial process in the form of aCare Order, no amount of input from Social Services (even if fully supported bythe other agencies) can provide adequate protection for the child. (Louis BlomDownloaded from https://academic.oup.com/bjsw/article-abstract/20/6/525/1657327 by rmit3 user on 25 June 2019536 GERALDINE MACDONALDCooper Q.C. Commission Of Inquiry into the Circumstances surrounding thedeath of Jasmine Beckford. p. 208).If no form of casework were available other than that commonly onoffer at the time of Kempe’s work and if care orders were only soughtwhen there was unambiguous and demonstrable concern about thesafety of a child, then such a conclusion might be justified. Given theactual panorama of child protection cases, there is minimally, room fordisagreement.Finally, in using outcomes as evidence of wrong decisions Hollis andHowe make an error similar to that of Professor Greenland, expertwitness to the Inquiry into the death of Jasmine Beckford:Professor Greenland told us … that it seemed prudent to classify all nonaccidental injuries to young children as ‘high risk’ cases, since 80% of all children unlawfully killed by their parents had been previously abused (Beckford,p. 288).It does not follow that because eighty per cent of ‘unlawfully killed’children have been previously abused by their parents that even themajority of non-accidentally injured children are at risk of being killedby their parents. If an expert witness in the field of child protection andeminent lawyers responsible for determining the reasons for bad outcomes can make this fundamental error, and if social workers are taughtsuch faulty decision-making aids, then difficult decisions are going to berendered—erroneously—more straightforward than they in fact are.RESPONSIBILITYMany societies have arranged their affairs so that someone can be heldlegally responsible for an outcome for which he is demonstrably not toblame (Honor6, 1988, for example, argues for such an arrangement).But that people have done so is not, pace Nagel, proof that outcomesshould determine the moral culpability of the actor. Such laws typicallyreflect an attempt to organize society in ways that safeguard the interestsof citizens in particular situations and their social basis is, in part,attested by the ample debate concerning when it is in fact reasonable orjust to hold a person legally responsible for unforeseen and unforeseeable outcomes; consider, for example, debates about the age at whichchildren can be held responsible for their actions.But if outcomes are rejected as a major, if not the sole, source ofadjudication of moral culpability, the alternative is to consider theDownloaded from https://academic.oup.com/bjsw/article-abstract/20/6/525/1657327 by rmit3 user on 25 June 2019ALLOCATING BLAME IN SOCIAL WORK 537action itself and the contingencies surrounding it. The organizationalcontext of social work is a major source of such contingencies.SOCIAL WORKIn order to fulfil their statutory obligations, welfare agencies, includinglocal authority social services departments, employ individuals to act ontheir behalf. Some of these, such as fieldworkers, appear to have a greatdeal of autonomy, and often know their clients more intimately than dothe other professionals involved. Hollis and Howe take this as groundsfor asserting that the individual social worker has access to the ‘bestinformation, including the best-informed judgement of relevance andlikely consequences’ (1987, p. 132), and therefore, even when responsibility is shared, it is not shed. The ‘best information’ claim is questionable (what about expert knowledge?). More important, their neatalliteration (shared/shed) disguises a false dichotomy. They do not consider the possibility that (leaving aside those cases when individualsdeliberately mislead their managers, for example, the social worker whowrites reports about visits that never occurred) discussions about thesharing or allocation of responsibility amongst individuals may be misguided and that social work, because of its very nature, needs to moveaway from an individual morality.In order to begin to unpack the concept of an organizational ethic, Ishall consider two areas of practice which, I hope, will also highlight therelationship between it and individual responsibility, namely the caseconference and supervision.CASE CONFERENCESThe majority of decisions concerning children thought to be in need ofprotection are based on the deliberations of a group of people from avariety of professions: the case conference. The case conference itselfdoes not make decisions (other than whether or not a child’s nameshould be placed on the Protection Register); only recommendations,which are rarely contested—which is not to say that participants alwayscarry out agreed plans of monitoring or intervention.Such decisions are taken in this manner because it is recognized that amore accurate picture of a situation is gained by bringing togetherinformation not available to any one party. Further, the perceptual andpersonal biases of individual workers are exposed to contradictory viewswhich may lead to increased objectivity in a task that is otherwisevulnerable to subjective responses. However, from what we know of thepsychology of group behaviour, we have reason to be sceptical as toDownloaded from https://academic.oup.com/bjsw/article-abstract/20/6/525/1657327 by rmit3 user on 25 June 2019538 GERALDINE MACDONALDwhether the case conference is the rational and ‘objective’ body itendeavours to be.Consider the substantive social work discussions about the organization of case conferences. The recommendation that someone other thanthe line manager of the social worker concerned should chair the conference is, in part, aimed at minimizing one source of collusion or bias,and maximizing the chance that the chairperson will be objective. Askilled chairperson should be able to structure the conference to ensurethat the status of the contributions from participants is clear (for example, are they reporting a one-off incident that is atypical, or a regularoccurrence?) and that judgements reflect current knowledge concerningchild abuse and neglect. Only when there are strong reasons for thinkingthe case in question is idiosyncratic should courses of action be pursuedwhich are contraindicated by, for example, current research concerningrisk factors. This offers the participants, and the chairperson, someprotection against the very strong pulls of individual relationships,whether between social worker and client or social worker and otherconference member. All of this suggests that we need to evaluate bothindividual actors and our arrangements for aggregate action. The mapping of individual action, within the case conference, to the outcome ofthe conference is not straightforward. A fortiori the allocation ofresponsibility for a decision cannot be reduced to the actions of oneindividual.SUPERVISIONSupervision, despite its apparent simplicity in providing a more routineway of monitoring the development of cases and the appropriateness ofdecisions, is a similarly complex activity. Consider ‘the facts’, but froman ethical perspective. Any analysis of supervision shows that, for it tobe done well, does not simply require an act of will on the part of oneindividual. For supervision to achieve apt monitoring requires co-operation of actors, requires that they have received appropriate training,requires that the context support and require this activity, and more. Sothe appropriate moral imperatives are those which require all of these;not imperatives centring around responsibility lor particular outcomes.Inquiry after Inquiry has shown that supervision rarely fulfils its role,for reasons which include the social worker’s need for support in astressful area of work; overwork (no time for regular supervision or fordoing more than providing updates on a few cases); and overidentification with the client or the worker’s presentation of the client (cf. DHSS,1986). Good supervision would provide a regular source of ‘professionalchallenge to fixed perceptions’ (Sheldon, 1987), routinely asking theDownloaded from https://academic.oup.com/bjsw/article-abstract/20/6/525/1657327 by rmit3 user on 25 June 2019ALLOCATING BLAME IN SOCIAL WORK 539sorts of question which gently challenge the worker to provide evidencefor her views and proposals, and eliciting information that she might be(unwittingly perhaps) disattending to. It might be argued that such anapproach would add to the stress already placed on workers, but emotional support, though very important (and the author has had occasionto value it greatly) is vacuous if the work being supported is misplaced.The stress of a professional challenge comes only from its rarity; oftenit is provided only when something has gone awry. Routinized as part ofthe ‘ethos’ of a team, it provides a source of stimulation, underlining theworker’s role as an agent of planned change, in an arena in which it iseasy to miss things which, in retrospect, look obvious. Even socialworkers find it scarcely credible that in some cases of child deaths, theworker concerned failed to see the child for some considerable time.But in the midst of the pressures of work, the plausibility of our clientswhom as a general rule we rather like, and our tendency to selectivelyperceive data which support the decisions we advocated, or the progresswe hope is taking place, it is all too easy to understand, and in theabsence of active steps to prevent it, probably inevitable. ‘Bureaucratic’procedures such as the case conference and supervision, and includingsome other administrative requirements such as goal-oriented recordkeeping should be ways of enabling and constraining individuals tomake optimal decisions on behalf of their clients in the white-heatreality of day to day demands. The moral imperative is to make decisions correctly and to act accordingly.CORRECT DECISIONSTaking decisions ‘correctly’ is not an issue of surface propriety. All toooften the organizational response to tragic outcomes has been to heapadministrative procedure upon administrative procedure with insufficient attention to content. This is defensive social work (Harris, 1987).Correctly taken decisions are those for which appropriate information issought from diverse sources, appropriately weighed against availableknowledge, and whose outcomes are fed back into that knowledge base(both in terms of future organizational practice and, where pertinent, aresearch base) to inform future practice. This requires, minimally, amore rigorous and empirically-based practice than social work typicallyespouses. It has implications for the content of administrative requirements (task-oriented recording to replace unstructured running records;opinions to be differentiated from fact; and so on): for the laissez-fairenature of social work practice (do what feels right for you): and, byimplication, for the content of social work training, which is all too oftencluttered by unsubstantiated ideas of dubious utility.Downloaded from https://academic.oup.com/bjsw/article-abstract/20/6/525/1657327 by rmit3 user on 25 June 2019540 GERALDINE MACDONALDINDIVIDUAL PREVENTABLE ERRORBut what of those occasions when individuals simply fail to execute theirresponsibilities with the result that their actions or inactions result intragic outcomes? Where does responsibility lie for this individual andpreventable error? In part, obviously, with the errant individual.But consider a ferry which sinks because the person whose job it wasto close the bow doors was asleep as the ship left port (we might call thisthe Herald of Free Enterprise problem). In this situation, as in thecircumstances surrounding the death of Kimberley Carlisle (Report1988), there were low cost managerial or organizational modificationswhich, had they been made, would have rendered the outcome robustagainst these actor failures. For the Herald of Free Enterprise, an electronic device which indicated in the bridge whether or not the doors hadbeen closed. The recommendations of the Inquiry into Kimberley’sdeath contain a number of such changes ranging from the inadvisabilityof senior social workers carrying child protection cases, to the need forthe Health Authority to ensure that there is a ready supply of appropriate percentile charts and that they are used (for all children up to the ageof two and those older children whose development gives rise to concern).3 Sometimes these procedures are generated by the examinationof bad outcomes and would not necessarily have been self-evidentbeforehand. Others (including some which feature all too often inInquiries) represent failures on the part of organizations to set up andmaintain such helps and constraints. A morality of action requires thatorganizations attend to the structure of decision making and the contingencies of individual action in such a way that should an individualfail, the failure is identifiable prior to a bad outcome.INEVITABLE ERROR?So far, we have dealt with two situations: the first in which an individualconducts himself according to the requirements of good decision-making and still a bad outcome arises. The second is one in which anindividual errs in some serious way within an organization which may ormay not be equipped to identify that failure.However, there is considerable evidence that the nature of occupations such as social work is such that their very performance depends oncorner-cutting behaviour which, with the benefit of hindsight, is all tooeasily implicated in inquiries into tragic cases. Such an analysis has3It also contains a number of other recommendations pertaining more to the seaworthiness of the ship, such as slowing down the turnover of staff and providing better trainingand adequate resources. When these are contributory factors then the issue of responsibility and moral blame is, I think, shifted to a wider arena.Downloaded from https://academic.oup.com/bjsw/article-abstract/20/6/525/1657327 by rmit3 user on 25 June 2019ALLOCATING BLAME IN SOCIAL WORK 54Iimmediate force for anyone who has ever worked as a social worker, buthas been given more substance by the work of Lipsky (1980). Lipsky’sargument is not that, courtesy of Mrs Thatcher, social workers areunder-resourced and therefore cannot do their jobs properly. Rather,the overcrowding of their day is an inevitable aspect of all those occupations involved in providing ‘free’ service to the general public, andincludes the police, teachers and health workers, whom he collectivelyrefers to as ‘street-level bureaucrats’:A distinct characteristic of the work setting of street-level bureaucrats is that thedemand for services tends to increase to meet the supply . . . Appreciation of thedemand-supply dilemma in street-level bureaucracies does suggest that theproblem of the quality of service delivery is not likely to yield easily to anyimaginable resource increments (Lipsky, 1980, pp. 33, 38).On this analysis, lack of time for reflection is an ineluctable characteristic of the social worker, whom he describes as ‘trapped in a cycle ofmediocrity’ (p. 32).. . . street level bureaucrats work with a relatively high degree of uncertaintybecause of the complexity of the subject matter (people) and the frequency orrapidity with which decisions have to be made. Not only is reliable informationcostly and difficult to obtain but for street-level bureaucrats high case loads,episodic encounters, and the constant press of decisions force them to actwithout even being able to consider whether an investment in searching formore information would be profitable (Lipsky, 1980, p. 29).And again:Street-level bureaucrats characteristically have very large case-loads relative totheir responsibilities. The actual numbers are less important than the fact thatthey typically cannot fulfil their mandated responsibilities with such case loads(Lipsky, 1980, p. 29).We are back at the most persuasive of Nagel’s intuitions (‘the driver . . .guilty of even a minor degree of negligence . . . will blame himself forthe death’) (Nagel, p. 28). If Lipsky is correct then the range of actionsexhibiting ‘a minor degree of negligence’ is now much wider than wehad earlier taken it to be. This means that, on Nagel’s analysis, all socialworkers—good as well as bad—would be at risk of serious blame fromchance outcome. Such a means of ascribing responsibility becomesmeaningless; it ceases to be a good way of arranging human affairs.However, given Lipsky’s analysis, we must next ask whether an argument advocating the considered nature of decision-making and actiondoes not impose on social work an unreasonable, if not pernicious, setof moral prescriptions?Downloaded from https://academic.oup.com/bjsw/article-abstract/20/6/525/1657327 by rmit3 user on 25 June 2019542 GERALDINE MACDONALDACCOUNTABILITYIt is because of the pressures inherent in social work that the individualistic morality currently espoused by the BASW Code of Ethics is somisplaced. A particularly unfortunate injunction is:They will not reject their client or lose concern for their suffering, even if . . .obliged to acknowledge an inability to help them ( B A S W Code of Ethics, 1986,10:5.).Such moral imperatives place limitless and undefined demands on socialworkers, and enshrine irreconcilable demands. The moral constraintsthat social workers require are not admonitions to ‘care’ but the constraints afforded by, for instance, structured supervision, goal settingand case closure. If corners must be cut and the possibilities for actionare timelimited (that is, any emergency attended to will mean disattending to other situations) it becomes the responsibility of organizations toface these pressures squarely and ‘come clean’ about the boundaries ofacceptable deferments or non-performances. If organizations are notprepared to indicate which areas of work should have priority overothers, and do not make it possible for workers to remedy omissions(whether maintaining records or receiving training appropriate to a casethey hold) then a morality based on outcomes exercises a strong appealand leaves workers in a very vulnerable position. This is not an argument to further bureaucratize social work, despite Lipsky’s somewhatunfortunate phrase; social work is inevitably a very private activityrelying, to a great extent, on the knowledge and expertise of individualworkers. The unique combination of care and control that characterizessocial work demands that it remains so. However, the social worker’sremit is a public one, as is the accountability. The argument presentedhere is simply that the organizations in which they work could, andshould, facilitate the optimum execution of that remit and relieve someof the conflicting pressures that exact such a heavy toll on workers.To follow Nagel in ascribing responsibility on the basis of outcomessimpliciter leads to bad social work. It generates inequities in treatment(she is pilloried, he, acting similarly, is promoted); it leads to stress andburnout amongst individual social workers; and their response mayrationally be (pejoratively) ‘defensive’ social work. Moreover, by stressing the individual action the stance diverts attention from the consideration of social work as an organized, interdependent activity.Hollis and Howe’s caricature of the social worker as an actor in a GreekTragedy encourages the individualization of social work action; toaccept it is to encourage the pageant of the bleeding heart, to legitimatethe agonized breast beating, the appalling sense of irreconcilable forces.Downloaded from https://academic.oup.com/bjsw/article-abstract/20/6/525/1657327 by rmit3 user on 25 June 2019ALLOCATING BLAME IN SOCIAL WORK 543Whereas the stock dilemma used to be to decide on welfare grounds which oftwo risky courses would be best for the child, it is now entangled with the risk oflegal and political action to protect parents’ rights (Hollis and Howe, 1987, p.129).It contains no viable prescription for action.In contrast, if we focus on the action, we can concentrate on maximizing service delivery and optimizing the organization and conduct ofsocial work. When actions are evaluated on an outcome basis, there isimmense scope for ‘sloppy’ practice (why do so many people risk driving above the legal speed limit, if not because most of the time they ‘getaway with it’?). An action-centred approach would both facilitate goodpractice and curb poor social work.The analysis presented here restores the ‘stock dilemma’ to centrestage, and adds context. If the best possible decision is made, then evenif a bad outcome occurs, it is morally correct (though at this stage in theargument the word ‘morally’ adds little). And if poor decisions aremade, or misguided courses of action followed with no bad outcome?On this analysis, one would hold individuals responsible for their actionsrather than the outcomes of their actions and, where appropriate, wouldpenalize them accordingly. Similarly, a morality of action would punishdrunk drivers as severely as drunk drivers who had killed or severelyinjured a pedestrian. Again, this is not an argument for the transformation of social work into a profession whose internal discipline isreminiscent of the armed forces. But strictly speaking, the moral culpability of the worker who fails to demand to see a child about whom he orshe has reason to be concerned but who is in fact safe, equals that of theworker who behaved similarly and who later learns the child died thatnight. We should be no more critical of the latter than of the formerand, in an organization which fails to facilitate or constrain theindividual in such a difficult task as social work, perhaps we should beless critical of either and more critical of the context in which heoperates.DO OUTCOMES MATTER?Of course, aggregate outcomes matter—they are our judgement on oureffectiveness and influence the knowledge base against which individualcases should be assessed. It is, I would argue, precisely the style of socialwork entailed by an apt morality of action which would routinely generate such information. An interventionist discipline is subject to apeculiar requirement to restrict itself to demonstrably effective interventions or to evaluate work in areas where no such data exist (for aDownloaded from https://academic.oup.com/bjsw/article-abstract/20/6/525/1657327 by rmit3 user on 25 June 2019544 GERALDINE MACDONALDfuller discussion of this see Hudson and Macdonald, 1986, Ch. 1). In thiscontext it becomes obvious that a particular negative outcome does notprovide evidence of a misinformed course of action; we are dealing withaggregate probabilities, and there will inevitably be negative outcomesfrom optimal decisions.CONCLUSIONThe argument has been against the use of individual outcomes aspresumptive evidence of individual blame. I have argued, instead, for aconsidered and concurrent (as opposed to retrospective) evaluation ofactions, which attends to the relationship between individual andorganizational responsibility for good practice, based on a considerationof the empirical context in which social work takes place.Sound decisions cannot be made without reference to aggregate outcomes, which includes information collected routinely by organizationson their effectiveness and efficiency (for example, Freeman and Montgomery, 1988) as well as other research of relevance to any particulartask. Some bad outcomes arise from the fact that social workers arevulnerable to more ‘extrinsic luck’ than might be apparent, given theidiosyncratic nature of social work training.If you find the formal argument unconvincing, consider this tale.As our opening quotation shows, Blom-Cooper is aware of thedangers of retrospective evaluation. Here is the same writer (with colleagues) considering the events which led to the death of KimberleyCarlisle. The report contains much awareness of the external constraintson action:To say that the room where the Carlisles talked to Mr Ruddock was incommodious would be to overstate its suitability. . . . Nothing could be less suitablefor watching and listening to interaction between members of the family. . . . Werecommend that urgent consideration be given . . . to providing . . . proper facilities for interviewing. . . . Partly because of these wholly inadequate conditions,Mr Ruddock misread the signs (A Child in Mind, 1987, pp. 108-10).However, on the next page the writers report Mr Ruddock’s observation of the family group as they walked back to the car—presumably inprinciple a good interaction, not suffering from the disadvantagesimposed by the cramped interview room:‘I walked with the family to the door of the building, and watched as theywalked across the road to where their old car was parked. I still have a clearmental picture of the way in which they all walked across the road and got intothe car, parents holding children by the hand, children leaping around in the carDownloaded from https://academic.oup.com/bjsw/article-abstract/20/6/525/1657327 by rmit3 user on 25 June 2019ALLOCATING BLAME IN SOCIAL WORK 545as they got in, laughing, shouting and playing happily with each other. It wasalmost an archetype for a happy family scene.’Far from being reassured, Mr Ruddock should have been alive to the risk ofbeing manipulated. Plainly he had been deceived . . . (A Child in Mind, 1987,pp. 111-12).The question—not posed by the Inquiry—is: what factors would enablea social worker to discriminate when such a ‘happy family scene’ is not a‘happy family scene’ at all, but a deception? With the benefit of hindsight, we know that the family was not ‘a happy family’: but the factremains that Mr Ruddock’s goal when interviewing the family was toform an assessment of family relationships and of Kimberley’s wellbeing. Imagine the same account ending: ‘And so I immediately soughta place of safety order.’Even experienced assessors fall into the trap of retrospective judgement. If the actors (including the organization) were at fault their faultshould be locatable—and should have been located—independently ofthe death of Kimberley Carlisle. Moral rules which disregard the possibilities and constraints of human behaviour are dangerous—both toworkers and the people they are endeavouring to help.REFERENCESBASW (1975) ‘A Code of Ethics for Social Work’, in Watson, D. (ed.) A Code of Ethicsfor Social Work: the Second Stage, London, Routledge and Kegan Paul.DHSS (Social Services Inspectorate) (1986) Inspection of the Supervision of Social Workers in the Assessment and Monitoring of Cases in Child Abuse When Children, Subjectto a Court Order, have been returned home.Freeman, I. and Montgomery, S. (eds.) (1988) Child Care: Monitoring Practice, ResearchHighlights in Social Work 17, London, Jessica Kingsley.Harris, N. (1987) ‘Defensive social work’, British Journal of Social Work, 17, pp. 61-9.Hollis, M. and Howe, D. (1986) ‘Child Death: Why Social Workers Are Responsible1,Community Care, June 5, pp. 20-1.Hollis, M. and Howe, D. (1987) ‘Moral risks in social work’, Journal of Applied Philosophy, 4, pp. 123-33.Honore A. M. (1988) ‘Responsibility and Luck’, Law Quarterly Review, October 1988,pp. 530-53.Howe, D. (1986a) ‘Welfare Law and the Welfare Principle in Social Work Practice’,Journal of Social Welfare Law, May, pp. 130-43.Howe, D. (19866) Social Workers and their practice in welfare bureaucracies, London,Gower.Hudson, B. L. and Macdonald, G. M. (1986) Behavioural Social Work: An Introduction,Basingstoke, Macmillan.Lipsky, M. (1980) Street-level Bureaucracy: dilemmas of the individual in public services,London, Russell Sage.Nagel, T. (1979) ‘Moral luck’, in Mortal Questions, London, Cambridge University Press,p. 24-38.Pine, B. A. (1987) ‘Strategies for more ethical decision making in child welfare practice’,Child Welfare, 66, pp. 315-26.Downloaded from https://academic.oup.com/bjsw/article-abstract/20/6/525/1657327 by rmit3 user on 25 June 2019546 GERALDINE MACDONALDPinker, R. (1982) ‘Appendix B. An Alternative View’ in Barclay Committee Report,Social Workers, Their Roles and Tasks, London, Bedford Square Press.Sheldon, B. (1987) ‘The Psychology of Incompetence’, in Drewry, G., Martin, B., andSheldon, B. (eds.) After Beckford, London, Social Policy Institute.Williams, B. (1976) ‘Moral Luck’ Proceedings of the Aristotelian Society, supplementaryvolume L, pp. 115-35. Reprinted in Williams, B. (1981) Moral Luck and otherEssays, Cambridge, Cambridge University Press.INQUIRY REPORTSBeckford (1985) A Child in Trust: The Report of the Panel of Inquiry into the circumstancessurrounding the death of Jasmine Beckford (London Borough of Brent).Carlisle (1987) A Child in Mind: The Report of the Commission of Inquiry into thecircumstances surrounding the death of Kimberley Carlisle (London Borough ofGreenwich).Downloaded from https://academic.oup.com/bjsw/article-abstract/20/6/525/1657327 by rmit3 user on 25 June 2019
Australian Expert Writers