Alliance Evidence in Mandated Work: What Transfers to Child Protection

Power and choice in child protection

In voluntary therapy, a person chooses to attend, and the alliance is a bond between helper and person with agreement on goals and tasks. YFL's essay Building the Working Alliance Based on Confidence That Can Be Earned covers that construct and its application under mandate. The narrower question here is evidential: what has been measured, in child protection settings, about the quality of the relationship between worker and parent, and what happened to outcomes? Figures are given as the original authors report them.

What parents reported

Gary Dumbrill, a social work researcher at McMaster University in Ontario, Canada, interviewed 18 parents who had received child protection services and built a grounded theory from their accounts. Parents perceived workers as using power either over them, as control, or with them, as support. Their responses fell into three groups: fighting workers openly, playing the game by feigning co-operation, or working with them collaboratively. Parents who experienced power used over them tended to fight or play the game, and parents who experienced power used with them tended to work with the intervention (Dumbrill, 2006).

Two features matter for the question of transfer. First, "playing the game" is the pattern YFL's essay Disguised Compliance: The Confidence Measure's Hardest Case examines: cooperation performed without any change behind it. Second, Dumbrill found no evidence linking whether a case was voluntary or non-voluntary to whether parents experienced power as over or with them. The sample is small and from one province, so this is an observation, not a general result. Child protection systems differ between Canada and England in law and structure, so transfer to England is a further question; comparative accounts of the two systems exist (Swift, 2011, for Canada; Parton & Berridge, 2011, for England), but the chapters have not been read for this draft.

What workers did

Donald Forrester, a researcher in child and family social work who later became Professor of Child and Family Social Work at Cardiff University, led two studies of how social workers talk to parents.

In one, 24 taped interviews between social workers and an actor playing a parent were rated. On average, workers asked many closed questions, used few reflections and rarely identified positives, and most showed low empathy. The factor that most strongly influenced the simulated parent's responses was empathy: empathic workers produced less resistance and more disclosure, and this did not come at the cost of failing to raise concerns (Forrester et al., 2008a).

In the other, 40 social workers from seven London local authorities responded to nine vignettes. The authors described a consistently confrontational style, with little listening, and judged it likely to be a systemic issue (Forrester et al., 2008b).

Both studies used simulated or vignette conversations, so they show what workers do in test conditions, not what parents experience over time, and neither measured later outcomes for children.

A training trial. A randomised trial in one local authority tested a package of motivational interviewing training and supervision. Workers were randomly assigned to the package (28) or to a control group (33), and families were then randomised to trained (67) or untrained (98) workers. Family meetings were rated for motivational interviewing skill, and research interviews gathered Working Alliance Inventory data and, 20 weeks later, other outcome measures. The skill difference was statistically significant but not substantial (2.49 in the control group, 2.91 in the trained group, p = .049), and no other outcome measure differed significantly between groups. The authors concluded that the package did not raise skills enough to influence engagement or outcomes (Forrester et al., 2018). This tests training, not the relationship itself.

What happened in Hertfordshire

Family Safeguarding Hertfordshire (FSH) was a whole-system reform of children's services in Hertfordshire, England, supported by the Department for Education's Innovation Programme. It aimed to improve the quality of work with families through multidisciplinary teams, structured tools to support direct work, and enabling factors such as reduced caseloads (Forrester et al., 2017).

2017 evaluation. The comparative study was a before-and-after design without a control group, with a staggered roll-out, and all pre-FSH families came from a single office. Rated practice skills changed in one respect: "good authority" rose from a mean of 2.37 to 2.69 (p = 0.022), mainly through greater clarity about concerns, which rose from 2.10 to 2.64. "Care and engagement" rose from 2.23 to 2.38, which was not statistically significant. The proportion of families with a child entering care fell from 12% to 10%, also not significant, and days in care in the five months after allocation fell from 20.54 to 9.81 per family (p = 0.06). The evaluators listed limitations including few follow-up interviews, and key performance indicators with no pre-FSH data (Forrester et al., 2017).

2020 evaluation. An independent evaluation covered Hertfordshire and four other authorities. It compared the period after introduction with pre-existing practice, using time-adjusted analysis, with statistical-neighbour comparisons for looked-after children. It reported statistically significant reductions in new looked-after-children rates among under-12s in four of the five authorities, with population-adjusted changes of between 9% and 30% in the direction of fewer children, and significant reductions in child protection plans in the four authorities with data, from 7% to 46%. The evaluators noted that they could assess only two years of delivery in the new authorities and, for police call-outs, that there was no comparator data from authorities without Family Safeguarding and that call-outs might fall to some extent under any well-functioning model (Rodger et al., 2020).

What these findings can carry

The sequence is suggestive: a reform aimed at better practice was followed by clearer communication of concerns and lower use of care and protection plans. But several steps are missing. The relationship, as parents would experience it, was not the measured variable: practice skills were rated, and the one rated skill that rose significantly was good authority, mainly through clarity about concerns, not care and engagement. The reform also changed team composition, caseloads and tools together. And without a comparison group, trends elsewhere in the system cannot be ruled out.

The mandated setting adds one further difficulty. A parent's cooperation can be real or performed, as Dumbrill's "playing the game" shows, so a behavioural measure of alliance may overstate it. YFL's essay Disguised Compliance: The Confidence Measure's Hardest Case sets out the markers that distinguish the two. Whether parents' own reports of the relationship predict outcomes in mandated settings is not answered by the sources gathered for this essay.

The alliance idea has some support in child protection settings, but of a weaker kind than in therapy. Among the sources gathered for this essay, the observations are small, simulated or before-and-after, apart from one randomised trial of training that found little effect on engagement or outcomes, and none isolates the relationship. Other research may exist that was not reached here. YFL offers this as information for readers to weigh, not as instruction.

YFL's essay Does the Helping Relationship Heal? What the Evidence Shows sets out the therapy evidence that this essay sets against child protection. Cures Ahead of the Evidence: When Interventions Judged Beneficial Were Tested explains why a before-and-after result without a comparison group needs particular care, and Small Effects at National Scale: The Prevention Paradox and the Helping Relationship takes up what modest average effects mean across a whole system.

---

References

Dumbrill, G. C. (2006). Parental experience of child protection intervention: A qualitative study. Child Abuse & Neglect, 30(1), 27-37.

Forrester, D., Kershaw, S., Moss, H., & Hughes, L. (2008a). Communication skills in child protection: How do social workers talk to parents? Child & Family Social Work, 13(1), 41-51.

Forrester, D., McCambridge, J., Waissbein, C., & Rollnick, S. (2008b). How do child and family social workers talk to parents about child welfare concerns? Child Abuse Review, 17(1), 23-35.

Forrester, D., Lynch, A., Bostock, L., Newlands, F., Preston, B., & Cary, A. (2017). Family Safeguarding Hertfordshire: Evaluation report (Children's Social Care Innovation Programme Evaluation Report 55). Department for Education.

Forrester, D., Westlake, D., Killian, M., Antonopoulou, V., McCann, M., Thurnham, A., Thomas, R., Waits, C., Whittaker, C., & Hutchison, D. (2018). A randomized controlled trial of training in motivational interviewing for child protection. Children and Youth Services Review, 88, 180-190.

Parton, N., & Berridge, D. (2011). Child protection in England. In N. Gilbert, N. Parton, & M. Skivenes (Eds.), Child protection systems: International trends and orientations. Oxford University Press. [Pages to be confirmed.]

Rodger, J., Allan, T., & Elliott, S. (2020). Family Safeguarding: Evaluation report. York Consulting LLP for the Department for Education.

Swift, K. (2011). Canadian child welfare: Child protection and the status quo. In N. Gilbert, N. Parton, & M. Skivenes (Eds.), Child protection systems: International trends and orientations. Oxford University Press. [Pages to be confirmed.]