No evidence scheme placing social workers in schools reduces need for social care, says What Works

One of the largest ever randomised controlled trials (RCTs) undertaken in sector finds no reduction in section 47 enquiries, leading centre to say no more money should be spent on project placing practitioners in schools

Image of social worker in school talking to teenager (credit: Valerii Honcharuk / Adobe Stock)
(credit: Valerii Honcharuk / Adobe Stock)

A scheme placing social workers in schools does not reduce children’s need for statutory social care services, one of the largest ever randomised controlled trials carried out in the sector has found.

Researchers found no statistically significant difference in outcomes for children between schools with a social worker and those without, What Works for Early Intervention and Children’s Social Care (WWEICSC) said today, following the trial which covered the 2020-21 and 2021-22 academic years.

WWEICSC – formerly What Works for Children’s Social Care – was commissioned by the Department for Education (DfE) to pilot SWIS in three local authority areas in 2018-19, and then to roll it out to 21 areas for the RCT.

School social work model ‘should not be continued’

“Whilst there was plenty of qualitive evidence suggesting that those involved – teachers, social workers and students – felt positive about social workers in schools, the RCT evidence did not find that it is having an impact on children’s outcomes,” said WWEICSC director of policy and practice Donna Molloy, in a blog post today.

“[Because] of the cost of SWIS and the lack of evidence, our conclusion is that the SWIS model tested in the trial should not be continued or further rolled out by the DfE or schools.”

The primary outcome the study was designed to test was the impact of placing social workers in schools on the rate of section 47 chid protection enquiries.

No impact on social care outcomes

It found that the rate of section 47s was 5.5% higher in the 136 schools studied that were part of the SWIS project, compared with the 132 control schools, after adjusting for baseline rates of section 47s, school size and the percentage of children on free schools meals. This difference was not statistically significant.

Likewise, the effects of SWIS on the other outcomes studied – rates of children’s social care referrals, section 17 children in need assessments, and children entering care, and the number of days spent in care per child – were similarly small and also not statistically significant.

The researchers also concluded – in a separate report – that there was no evidence that domestic abuse knowledge, attitudes and practice were better among designated safeguarding leads (DSLs) in SWIS schools than control schools. DSLs are responsible for child protection in their schools.

Overall, when the costs of SWIS, schools’ referrals to children’s social care and children’s social care interventions for pupils were taken together, the average was higher in intervention schools (£245,000 per 1,000 pupils per year) than controls (£210,500). As a result, researchers concluded that the scheme was not cost-effective.

Role ‘well-received by social workers, school staff and students’

The results came despite the role being generally well-received by the social workers themselves, school staff and students.

While social workers felt balancing the different aspects of their role – statutory work, preventive practice and working with school staff – was challenging – they valued the non-statutory elements of the role and the opportunities for informal interactions with pupils, staff and parents.

Most pupils interviewed who had had direct involvement with SWIS were positive, reporting that they trusted the social worker and that the practitioner understood them better than any other member of school staff.

And despite the domestic abuse report finding no evidence of impact on DSLs’ knowledge, attitudes and practice, safeguarding leads in SWIS schools said having a social work improved their confidence and ability in identifying and responding to domestic abuse.

‘Unmet need for support below statutory threshold’

Molloy said she recognised the findings would be “disappointing for many people – not least those involved in the intervention, who viewed it positively and felt that it led to good conversations and relationships”.

She stressed that the findings concerned the impact on need for statutory children’s social care services and interventions and was not a comment on the delivery of non-statutory family support within schools.

“Professionals and families clearly felt the need for additional support below the statutory threshold to address unmet need in schools and so a valuable next step could be to evaluate models of delivering school-based family support,” she added.

Following the report’s publication, the DfE confirmed it had accepted WWEICSC’s recommendation to cease funding SWIS.

“We know that looked after children often need additional support in school. However, we want to make sure that taxpayers’ money is used efficiently, and that support is delivered in the best way,” a spokesperson said.

“We are providing support in schools and colleges for children with a social worker through the virtual school heads programme, backed by over £57m investment.”

Virtual heads are council practitioners responsible for promoting educational outcomes for children with a social worker – whether looked after, in need or on a child protection plan – as well as those who have left care through adoption, special guardianship or a child arrangements order.

‘Disappointing results’

The Association of Directors of Children’s Services (ADCS) said it was “disappointing” that SWIS did not have “the kind of positive outcomes for students that we expected”, though said it awaited the further findings due next year on its impact on attendance.

The chair of its educational achievement policy committee, Heather Sandy, added: “As the evaluation notes, the project was well received by social workers, teachers and students and this is something local authorities have continually heard. Schools play a key role in safeguarding and ADCS is clear that the value of joint working between schools and local authority children’s services should not be undermined by these findings.”

She said school staff were “well placed to notice changes in [children’s] behaviour or concerns” and that the ADCS “[welcomed the government’s commitment to give schools a more prominent role in local safeguarding arrangements”.

This was set out in the DfE’s draft children’s social care strategy and will involve consulting on strengthening the role of education settings in safeguarding arrangements, as part of a spring 2023 update to Working Together to Safeguard Children.

It will follow this by consulting on whether and how to make education the fourth statutory safeguarding partner, alongside councils, police and health, this autumn.

Study strengths and limitations

The research team reported that the study had the following strengths:

  • The sample size (over 250 schools) was sufficiently big to have a strong chance of detecting a meaningful effect, in relation to section 47 enquiries, had there been one.
  • No schools dropped out over the trial period and there was no missing data in relation to the main outcomes studied.
  • The intervention and control schools were similar in terms of their initial children’s social care outcomes.
  • There was an extensive economic evaluation of the cost-effectiveness of SWIS.

But it said that the research also had these limitations:

  • Response rates to surveys of social workers (34% – 51%), school staff (43% – 60%) and students (11%) were relatively low, potentially resulting in selection bias.
  • There was also a relatively low completion rate for information on the time spent by social workers in schools, potentially leading to underestimates or overestimates of costs.
  • Planned observations by researchers of social workers in schools did not take place due to pandemic social distancing measures, leading to a loss of insights into the way practitioners worked.

Are you one of the social workers employed through SWIS and would like to share about your experience? Email or comment – anonymously, if preferred – below. 


4 Responses to No evidence scheme placing social workers in schools reduces need for social care, says What Works

  1. Anonymous April 2, 2023 at 10:44 am #

    Yes, referrals did not decreased why should they? Do we want to decrease the number of kids that receive support? DfE accepted recommendations to just save the Tory Government more money without seeing the results about attendance or hearing first hand from the children we have supported as SWIS The evidence is in the child’s voice, not penny pinching.

  2. Gaving April 3, 2023 at 8:37 am #

    The whole point of an RCT is that you account for all the variables except for the intervention via randomisation. This trial is a good example of why RCTs are not helpful when considering really complex ‘interventions’ such as placing social workers in schools. Literally, putting social workers in schools is not an intervention, it is not the same thing in every school, in every LA. It depends on the individual social workers, the schools, the local population, local needs, the availability of resources, the mindset of the school staff and the LA staff and how these interact etc. etc. The report of this trial even admits that the intervention varied from place to place. In which case, what exactly is it they were testing? And what exactly is it that “doesn’t work” ??

    If you are testing a new vaccine, then you can find out whether it works better than no vaccine or better than another vaccine or differently for men and women, for black people and white people, for children and adults, etc. because each dose of the vaccine is chemically and medically the same. Because the ‘input’ is the same, you can measure differences in outcomes and be reasonably confident they were caused by the input. If, however, you attempted to report an RCT of vaccine, and claimed that it didn’t work, while also admitting “of course, the actual contents of the vaccine were different in different places”, you’d never get published and would be laughed out of the lab.

    Running this expensive trial tells us nothing about whether putting Social Workers in Schools “works” because there is no such (simple) intervention to be tested.

  3. Geoff April 3, 2023 at 2:28 pm #

    The whole research premise is puzzling, with the view that having social workers in schools would prevent “inappropriate referrals”, and thus create value to the public purse.

    It very much ignores the underpinning lack of early help resources to meet need, and that this is the key aspect. This bit is particularly baffling given that it actually says that it wasn’t the role of SWIS to provide early help, just providing statutory services in a different office (the school):

    Social workers who are part of SWIS were physically located in schools and had an allocated caseload
    of children who have been referred to CSC and met the local authority’s threshold for a service
    (ranging from those who are children in need to those who are children looked after). Whilst the main
    focus of SWIS was statutory work, there may have been aspects of the role that could be described
    as preventative, such as advising staff and families when they have concerns that they would like to
    discuss, however SWIS did not include Early Intervention work which continued to be provided as a
    separate service.

  4. Anonymous April 5, 2023 at 10:31 pm #

    The social workers had no opportunity to do preventive work and the pilot has been a place to ‘dump’ cases that were not appropriate for the project (children attending alternative provision but ‘on roll’ at the designated school, one child on roll but with siblings in other schools in which the main issues lied etc) thus reducing case loads of other teams. Disappointing outcome for something that would have really helped if implemented correctly.