Is the Care System to Blame for the Poor Educational Outcomes of Children Looked After? Evidence from a Systematic Review and National Database Analysis

Abstract

It is recognised internationally that children in out-of-home care (‘children in care’ or ‘children looked after’) generally have lower educational attainments than other pupils. This article provides two forms of evidence that challenge the view that care status in itself can explain this ‘attainment gap’. A systematic review of 28 studies was conducted to assess the evidence on whether being in care is detrimental to young people's educational outcomes. This is complemented by an analysis of administrative data from England, which compares the educational outcomes of children in care at age 16 to those of children in the general population and to other children supported by social services. Taken together, the findings suggest that while research demonstrates an important attainment gap between children in care and children in the general population, this difference is reduced and in many cases disappears when other important factors are taken into consideration to reduce selection bias. We find little evidence that being in care is detrimental to the educational outcomes of children looked after, but suggest that given the heterogeneity of the population, special attention should be paid to different groups of children and their particular needs while in care.

Educational outcomes of children in contact with social care in England: a systematic review

“…One consists of the series of eight statistical releases by the DfE [14], one is a research report of longitudinal data by the DfE [13], two were reports self-published by a research centre [12] and a children’s charity [16] and three were peer-reviewed journal articles [15, 17, 18]. The DfE statistical releases are annual cross-sectional population-level snapshots of Key Stage results and exclusion and absence rates.…”

“…All had clearly stated research questions and ‘recruited’ participants from clearly specified and defined populations with high response rates. In the cases of the DfE statistical releases [14], DfE research report [13], Sebba et al [12], O’Higgins [18] and Luke and O’Higgins [17], data were available on all or almost all children in the target population. In Henderson et al [15], which used the Next Steps study, the response rate at wave 1 was 74% [19].…”

“…None provided an explicit sample size justification as they used all data available. However, this does not detract from study quality, except perhaps in the case of Ellison and Hutchinson [16] where this is a higher risk of non-response bias, as the sample sizes were sufficiently large to support the analyses undertaken, and in the case of Sebba et al [12], Luke and O’Higgins [17] and Henderson et al [15], point estimates with standard errors are presented.…”

“…However, children were ‘followed up’ retrospectively and only children with complete data (including at Key Stage 2 exams (age 11) and at GCSE (age 16)) were included in the models ( n  = 570,470 (89%) children). Luke and O’Higgins [17] and O’Higgins [18] used essentially the same data. There is therefore a potential selection bias if CLA or CiN are less likely to sit their exams, which is likely given the higher rates of exclusions and SEN among the looked after and in need groups.…”

“…Although this classification does allow for analysis by different levels of exposure, it still does not represent the full range of naturally occurring sub-groups of care and need trajectories which may be related with educational outcomes. Luke and O’Higgins [17], who, as previously outlined, used the same data as Sebba et al [12], also used the same exposure categorisation except they further divided the long-term group into subgroups. The focus of O’Higgins [18] was asylum-seeking or refugee children in care for a year or more, compared to other CLA in care for a year or more, CiN and the general population.…”