A systematic review of ethnic representation in UK research involving children and young people living in care

Abstract

A substantial and growing number of children in care in the UK are of minoritized ethnicities. Increasing evidence of inequalities faced by the UK’s children in care from different ethnic groups highlights the need for research that represents the ethnic diversity of children in care. This systematic review aimed to determine the frequency and manner of ethnicity reporting amongst studies of children in care in the UK from 2012 to 2024. Seven databases across multiple disciplines were searched, which led to the identification and inclusion of 196 papers. Over half the included studies (51.5 percent) did not report sample ethnicity data. There was notable variation in the style of ethnicity reporting: binary (e.g. White and non-White) and multiple categories (e.g. Asian, Black, Mixed, Other, White) were the most common approaches. Where ethnicity was reported, we found similar numbers of studies with samples under- and over-representing White or non-White children in care. We explore explanations for low rates of ethnicity reporting amongst studies including children in care in the UK and the limited research investigating differences in outcomes of children in care of different ethnicities. Recommendations are made for future research practice, including reflections on the role of data harmonization.

Language note: In this article, we use the term minoritized ethnic group (Gunaratnam 2003) to acknowledge that people are actively minoritized by systems of power rather than existing innately as a minority ethnic group. While we recognize the increasing use of the term global majority (Pollock, McCaughan, and Scholar 2024) we have chosen not to use this term to emphasize the particular experience of minoritized ethnic communities in the specific UK context, given the timeframe of our literature review.

Introduction

In England, in 2023, of the approximately 80,000 children living in care, more than 23,000 children (29 percent) were identified as belonging to a non-White British ethnic group (Department for Education 2023). There has been a steady increase in the proportion of children in care of minoritized ethnicities, rising from 23 percent reported in 2007 to 29 percent in 2023 (Department for Education 2007–2024), which in part reflects the growth of minority ethnic communities in the UK’s wider population. However, this increase is also related to the increase since 2015 in unaccompanied asylum-seeking children (UASC) who are in care. In 2023, UASC made up 9 percent of the population of children in care, with changing profiles of ethnicities amongst these children (Department for Education 2024).

Given these demographic shifts, and in light of the UK Equality Act 2010, there is growing recognition of the importance of promoting culturally sensitive and responsive social care practices (Ahmed et al. 2022). Research guiding social care practice is increasingly expected to reflect population diversity and include the perspectives of those from minoritized ethnic backgrounds. However, there is evidence that minoritized ethnic populations are often under-represented in health and social care research (Brown et al. 2014; Smart and Harrison 2017; Ekezie et al. 2024; Stemp et al. 2025). This systematic review explores ethnic representation in UK social care research that concerns children living in care.

Members of minoritized ethnic groups, including children, can face inequalities regarding experiences and outcomes in social care contexts. For instance, empirical research has demonstrated complex interactions between ethnicity, sex, poverty, and rates of children entering care in the UK (Owen and Statham 2009; Bywaters et al. 2019). While early work suggested an over-representation of Mixed and Black ethnicity children and an under-representation of Asian children in care (Owen and Statham 2009), this picture becomes complicated when taking into account socioeconomic disadvantage (Bywaters et al. 2019). Epidemiological work has shown that while Black and Asian children are more frequently living in the most deprived neighbourhoods, in those areas, White and Black Caribbean children are more frequently on a child protection plan or are in care as compared to, for example, Black African and Asian children (Bywaters et al. 2019). Once children enter care, there are notable differences in placement types, with Black and Asian children less likely to be adopted but simultaneously more likely to experience the most restrictive care orders, including secure accommodation or deprivation of liberty orders (Selwyn and Wijedesa 2011; Edney, Alrouh, and Abouelenin 2023). Research has shown that ethnicity may also associated with numerous outcomes for children in care and care-leavers. Black children living in care more frequently experience multiple care episodes while Black and Asian children in care more frequently experience youth justice involvement (Ahmed et al. 2022; Hunter, Francis, and Fitzpatrick 2023). However, patterns are complex, with children from minoritized ethnic backgrounds experiencing stable reunification with families of origin after leaving care (Goldacre et al. 2022). White and Mixed ethnicity children have the lowest rates of being in employment, education or training after leaving care (Ahmed et al. 2022). Given empirical evidence showing inequalities between care-experienced children of different ethnicities within a group that already faces significant inequity, there is a need for representative research that captures the needs of children from other ethnic groups (Chand 2008). While this review focuses on the UK context, it is important to note these inequalities have been identified across numerous Western social care systems (Cénat et al. 2021), meaning developing understandings of ethnic disparities amongst children in care a topic of international concern.

The emergent evidence regarding ethnic inequalities and the importance of conducting representative research reflecting the needs of children living in care demonstrates the importance of reporting ethnicity. The benefits of reporting ethnicity for improving transparency regarding research representativeness, generalizability, and interpretation and facilitating data pooling have been raised in the context of health (Buttery et al. 2022), and education research (Gaias et al. 2020). Indeed, ethnicity reporting has facilitated the identification of many examples of associations between ethnicity and health outcomes that have informed practice, policy and further research (Flanagin, Frey, and Christiansen 2021). Some argue that accurate recording of race and ethnicity is essential to improve the outcomes of marginalized groups, including those of minoritized ethnicities (Shanawani et al. 2006). However, there has been limited work applying these principles of ethnicity recording and reporting to research in social care contexts, or specifically with children living in care.

While reporting the ethnicity of children in care participating in research is important, inconsistencies in ethnicity reporting limit our ability to compare and understand differences in the experience and outcome of children from different ethnic groups (Lam et al. 2023; Murali et al. 2023). Data harmonization refers to the practice of making data comparable and coherent with findings from related data (Stillwell 2021). Ethnicity data harmonization is mandated and outlined by research governing bodies, including the Government Statistical Service (GSS) in the UK. The GSS harmonized categories classify ethnicity into five aggregated groups: White, Black, Asian, Mixed, and Other (Race Disparity Unit 2023). However, while five-level harmonized ethnicity reporting categories offer an improvement upon non-reporting or binary classifications (e.g. White and Non-White) this approach has been criticized due to its inability to reflect heterogeneity within these groups (e.g. as described in Bywaters et al. 2019).

We conducted a systematic review of the literature on children living in care in the UK, aiming to determine (1) how commonly ethnicity was reported, (2) how ethnicity was recorded, (3) how samples compared with national statistics on ethnicity, and (4) examine differences in outcome by ethnicity of children in care.

Methodology

Design

A systematic review following Preferred Reporting Items for Systematic Review (PRISMA) guidelines, (Page et al. 2021), pre-registered on PROSPERO (CRD42023423500).

Inclusion and exclusion criteria

The following inclusion criteria were used to select articles for review:

• Peer reviewed empirical studies in English, published after 1 January 2012. Qualitative and quantitative studies were included. Systematic reviews, commentaries, grey literature and research published in books were excluded.

• The study sample was children in local authority care in the UK at the time they were recruited. Therefore, young people currently living with adoptive families were excluded. For longitudinal data, the sample had to have been in care at the baseline assessment—this led to the exclusion of papers specifically exploring precipitating factors relating to entry into care (e.g. Bywaters et al. 2019).

• Participants were below eighteen years old at the baseline data collection. Where the included group of young people had mixed ages, the mean or median, depending on which was reported, was required to be below eighteen years and the maximum young person age to be twenty-five years at baseline.

• International studies were only included if there was a subsample of UK children, and data was reported for that subsample.

Search strategy

We conducted systematic searches of seven databases, for papers published between 1 January 2012 and 30 April 2024. The databases were Applied Social Sciences Index and Abstracts (ASSIA via ProQuest), Social Science Database (via ProQuest), Scopus (via Elsevier), Educational Resources Information Center (ERIC via ProQuest), CINAHL, PubMed and PsychINFO. Search terms fell into the conceptual categories of the UK context (e.g. ‘United Kingdom’, ‘Northern Ireland’), children and young people (e.g. ‘youth’), and care status (e.g. ‘state care’ ‘looked after’).

In addition, the reference lists of relevant systematic reviews were hand-searched by the joint first authors to identify pertinent other papers missed by the database search. The complete search strategy can be found in Supplementary material.

Screening and data extraction

Screening and data extraction were conducted on the Cadima systematic review platform (Kohl et al. 2018). After completing database searches and removing duplicates, abstracts were screened by six research team members, with 50 percent of abstracts double-screened. The same six authors screened articles that progressed to the full-text stage, with 50 percent of abstracts double-screened. Any discrepancies were resolved in consensus meetings. For each eligible paper, data were extracted based on publication date, geographic location, study type, and study methodology (e.g. qualitative). For study type, we categorized studies into primary data (self-report, proxy-report, or both), secondary research (local service data, national administrative data, cohort data), and mixed-primary and secondary research studies. Where available, data were also extracted for participant age, gender/sex, placement type, and asylum-seeking status.

As this review focuses on ethnicity, we extracted data on whether or not ethnicity was reported and how it was reported. The corresponding author was contacted to request the information if a paper did not report ethnicity.

Data analysis

To answer research question one and two, the proportion of published studies reporting and recording ethnicity was answered with descriptive information. To answer the third research question, only papers published on children in care in England were included (rather than the four nations). This was decided because the vast majority of studies were from England, England is a more ethnically diverse, and it allowed for comparisons against the Department for Education Data on Looked After Children in England. Because of differences in how ethnicity was reported (discussed later), we were only able to explore question 3 using categories of White and Non-White. We considered a study sample representative if their proportion of non-White participants was within 5 percent of the proportion of non-White children living in care in England that year, according to national data (Department for Education 2007-2024). For this question, we excluded national administrative data studies as these include the whole population of children in care and, therefore, are inherently representative.

While a meta-analysis had been planned to synthesize differences in outcomes by ethnicity, there was an insufficient number of papers reporting sub-group analyses for a meta-analysis. Therefore, the findings of studies that included sub-group analyses by ethnicity were qualitatively synthesized.

Results

Search results

A total of 2,557 items were identified from database searches, which was reduced to 1,635 after the removal of duplicates. Following title and abstract screening, 1,295 papers were excluded. Of the 340 items reviewed at the full-text screening stage, 192 eligible papers were included, and 4 papers were identified through a search of reference lists (see Fig. 1 for PRISMA Diagram; Page et al. 2021). All included papers can be found in Supplementary Material.

Study descriptives

One-hundred-and-ninety-six articles were included, just under half (46.4 percent) exclusively included data regarding ‘children in care’ as defined for this review (see Methods section), and just over half (53.6 percent) contained both ‘children in care’ and other samples. The sample sizes of children in care varied from single case studies (n = 1) to extensive secondary analyses of national administrative data (n = 60,000) with a strong skew towards smaller samples. The median number of eligible participants (children in care) was 49.

The included studies are summarized in Table 1. The majority of included papers were quantitative studies (n = 93). Primary research studies made up the highest proportion of papers (n = 127), followed by secondary research studies (n = 55), and mixed-primary and secondary studies (n = 15). Most studies included samples from England (n = 126), followed by UK-wide samples or those across multiple nations.

Ethnicity reporting

Of the 197 included papers, 51.5% (n = 101) did not report the ethnicities of the children in care (see Table 2). Ethnicity data were particularly infrequently reported in primary research studies (n = 65 of 123; 52.8 percent) and mixed primary-secondary studies (n = 8 of 13; 61.5 percent). Ethnicity was not reported in just under half of the secondary data studies (n = 28 of 60; 46.7 percent).

Of the seventy studies which compared children in care with samples from other groups of children, thirty-eight did not report the ethnicities of children. However, six of these thirty-eight studies did report the ethnicity of the total child sample without reporting a breakdown for children in care specifically.

Among the ninety-five studies that reported ethnicity data, thirty-nine (41.1 percent) used binary categories. All of these studies used ‘White’ or ‘non-White’ as their categories, although how ‘non-White’ was defined or labelled differed, including: Black and Minority Ethnic (BAME); Black; or ‘Non-Caucasian’. Furthermore, it was common for studies to report ethnicity by comparing ‘White British or Irish’ to ‘Non-White British or Irish’, therefore presumably including White children from non-British backgrounds (e.g. White European) alongside children with non-White ethnic backgrounds (e.g. Joseph et al. 2014; O’Higgins 2019).

Thirty-seven studies reported ethnicity using more detailed categories (i.e. at least three categories). However, there was no common approach to this. The most common multiple category approach, used by 11.6 percent of studies reporting ethnicity, was the use of five categories (White, Black, Asian, Mixed, Other), in line with GSS harmonization guidance discussed in our introduction (e.g. Pearson et al. 2020) and the Department for Education’s reporting on children in care. Several studies split these categories further, distinguishing between Indian and Pakistani Asian young people or between Black Caribbean and Black African young people (e.g. Vallejos et al. 2016). A small number of studies made more nuanced and specific distinctions such as reporting numbers of Irish Traveller, Lebanese and White Serbian participants or reporting detailed ethnicities of each participant (e.g. Peet and Teh 2020). Two studies focused on children of similar or the same ethnicities, specifically Black children living in care (Mantovani and Thomas 2014a, 2014b).

Although not conventionally regarded as an ethnicity category, a small number of studies used unaccompanied asylum-seeking status (UASC) as an ethnic category (e.g. Stanley et al. 2016). Twenty-two (11.2 percent) studies were conducted exclusively with UASC and 13 (6.6 percent) included some UASC amongst a wider sample of children in care. However, many studies did not specify if UASC were included in their sample (155; 79.1 percent), with a very small number specifically noting no UASC were included (6; 3.1 percent). Of the studies including only UASC, six did not report children’s ethnicities, and sixteen defined ethnicity by children’s country of origin or language spoken (e.g. Armitage et al. 2022).

In the secondary data studies, there was no standardized or common approach to reporting ethnicity. Of the twenty-five national administrative data studies, ten did not report ethnicity, eleven used multiple categories and four used binary categories. Of the ten studies using cohort data and twenty using local administrative data, approximately half the studies did not report ethnicity (cohort n = 5, local n = 10), with similar proportions of the remaining using binary (cohort n = 2, local n = 4) and multiple (cohort n = 3, local n = 3) categories.

Ethnic representativeness

Of the 125 papers conducted exclusively in England, 76 consented in participants (e.g. primary research and cohort studies) and, therefore, had the potential for recruitment biases (e.g. the over or under recruitment of certain ethnicities). Of these studies, thirty-one reported ethnicity by a binary or by multiple categories enabling binary White and non-White categories to be created to compare to national statistics from 2012 to 2024 (Department for Education 2007-2024).

Ten (32.2 percent) of these thirty-one papers reported on samples that were considered generally representative of national statistics (falling within 5 percent of the national statistics for the year of publication for White and non-White children in care living in England). Eleven studies (35.5 percent) over-represented non-White participants while ten studies (32.3 percent) under-represented non-White participants.

Overall, there was no clear pattern of either under- or over-representation of White ethnicities as compared to non-White participants across this group of studies.

Outcomes by ethnicity

Of the ninety-six studies that reported participant ethnicity, 10.4 percent (n = 10) reported subgroup analyses by ethnicity. Given the few studies reporting ethnicity with shared quantitative child outcomes, it was not possible to synthesize these outcomes using meta-analysis. A descriptive review of study findings by ethnicity revealed limited insights into patterns of need between children in care with different ethnicities.

Some studies found significant differences in outcomes related to ethnicity. For example, White and Mixed ethnicity children were more likely to re-enter care compared to Black and Asian children (Goldacre et al. 2022; Mc Grath-Lone et al. 2017) and rates of new-born entry into care were higher for children of Other and Mixed ethnicities, followed by White and then Black children (Pearson et al. 2020). Children of a minoritized ethnicity had lower odds of being able to identify a trusted adult in their lives compared with White majority children (Suh and Selwyn 2023). Ethnicity was also found to impact the educational trajectories of children living in care (Sutcliffe, Gardiner, and Melhuish 2017). Furthermore, Eisen, Williams, and Cohen (2023) found some differences between the country of origin of UASC and rates of positive tuberculosis testing.

Regarding mental health outcomes, mixed findings were identified concerning the ethnicity of children living in care. Children classified by the study as BAME ethnicity were over-represented in the general population amongst children with high PTSD symptomology but not amongst children in care specifically (Hitchcock et al. 2021). Contrastingly, White children living in care were found to have higher reported peer problems and hyperactivity compared to non-White children (Hiller et al. 2023).

Further studies were unable to find significant differences in outcomes between children of different ethnicities. For example, there were no ethnic differences found in predicted fixed school exclusions (suspensions) using White and non-White categorization (Melkman 2022). Joseph and colleagues (2014) found no significant relationship between the child’s ethnicity and the formation of new secure attachments by adolescents in foster care.

Discussion

This systematic review of 196 included articles explored whether and how studies reported on the ethnicity of children living in care. We found notably low reporting of ethnicity, with around half (51.5 percent) of the studies included not reporting on the ethnicity of children in care, nor the ethnicity of young people if there was a wider sample of young people. There was considerable variation in the style of ethnicity reporting. Over 40 percent of studies that reported on ethnicity used binary White vs. non-White categorizations and 38.9 percent used multiple ethnicity categories, with a third of this group utilizing GSS-harmonized five-category approach of using Asian, Black, Mixed, Other, and White categories. Yet other studies took alternative approaches, reporting language or country of origin as a proxy for child ethnicity. Of thirty-one studies including children in care conducted in England where ethnicity was reported, a third closely reflected national ethnicity statistics for children in care. However, no systemic pattern of over- or under-representation of White or non-White children was found.

These low rates of ethnicity reporting and lack of harmonization in ethnicity reporting has detrimental implications for social work practice. Failure to report ethnicity limits transparency as to study sample representativeness, generalizability and the ability of research to answer questions about ‘what works for whom’. Furthermore, failure to report on ethnicity risks the uncritical adoption of guidelines, policies or practices that may not reflect the needs of all children living in care. Research that reports on ethnicity may facilitate the addressing of unconscious bias and racism in social work practice and policies—with social care studies examining ethnicity already raising important questions regarding inequalities (e.g. Selwyn and Wijedesa 2011; Bywaters et al. 2019; Ahmed et al. 2022; Edney, Alrouh, and Abouelenin 2023).

There are several reasons why ethnicity reporting may have been lower in some studies, particularly because of concerns about protecting the anonymity of participants. In small-scale studies restricted to a local area, reporting child ethnicity may risk identifying children in care, especially where research has been conducted in nations or areas with low ethnic diversity such as Northern Ireland and Wales (e.g. Larkin and Lefevre 2020). Researchers have also commented on the challenges of accurate identification of participant ethnicity, particularly amongst children in care where case records and children’s self-identification may not align, or data may not be available (Schmidt et al. 2015; Bezeczky and Wilkins 2022). In secondary research, the quality of ethnicity data can be compromised due to the lack of guidelines on ethnicity data collection, especially in the case of third-party reporting (e.g. ethnicity inferred based on physical appearance in hospitals; Lam et al. 2023). This may be a particularly salient challenge regarding reporting of ethnicity for those with non-visible minority status (Song 2020). A comparison across three health administrative data sources revealed that Mixed ethnic groups and ‘Other’ ethnic groups more frequently experienced greater inconsistencies with their ethnicity records than White British and South Asian groups (ONS 2023). While there may be important ethically informed reasons for declining to report child ethnicity for some research, this is unlikely to explain the generally low and inconsistent rates of reporting.

There was considerable variation in the way studies reported the ethnicity of children in care, which may reflect current disagreements regarding best practices. There are several challenges and debates surrounding the appropriate reporting of ethnicity of research participants. First, there are pragmatic statistical reasons that a study may have to collapse ethnicity data into ‘White’ and ‘Non-White’, to ensure analyses are appropriately powered. In these cases, even if not reported, it would still be beneficial for these data to be collected and acknowledged, to support the ability to harmonize across studies, collate larger datasets, and therefore answer important questions about the potential unique needs of children in care from different ethnic backgrounds. It is also important to emphasize the risk of binary White/Non-White ethnicity categorizations masking large inequalities, such as those between Asian and Black children entering care (Bywaters et al. 2019). Even in large administrative data where this should not be an issue, ethnicity was often not reported. Another issue is the ongoing debate around what best-practice should be for reporting ethnicity. While data-harmonized census categories (Race Disparity Unit 2023) are widely used, they risk masking key disparities within ethnic terms like ‘Asian’ (Bhopal 2004)—failing to detect differences between children in care with different outcomes (e.g. Bywaters et al. 2019). Ethnicity is acknowledged as a dynamic, multidimensional and relational concept that may mean different things in different social contexts and historical periods, which may not be reflected in a single census category (Bhopal and Donaldson 1998; Burton, Nandi, and Platt 2010).

This study’s findings did not raise notable patterns in over- or under-representation of White and non-White children in care in our included studies that reported ethnicity. While only a quarter of primary research studies conducted in England were found to be representative of national statistics in England at the time of publication, equal numbers of studies over-represented White and non-White children. Some of this over- or under-representation may also be a product of the recruitment area and the research questions the study was addressing. Different regions in England have considerably different ethnic diversity. Of note, representative samples are likely only to include small numbers of non-White children (given White children comprise >70 percent of the care system). This is particularly the case when there is a lack of harmonization in reporting. Thus, it is also important to acknowledge that a specific focus on children of minoritized backgrounds is essential to fully understand needs, such as the growing qualitative literature on the experiences of Black children in care (e.g. Mantovani and Thomas 2014a).

It is notable that several of the included studies made concerted efforts to improve the ethnic representativeness of study samples of children in care or reflected on their study’s limitations in terms of ethnic representativeness. For example, some primary research studies used purposive sampling techniques to ensure that children from a range of ethnic backgrounds participated (e.g. Diaz, Pert, and Thomas 2018; Larkins et al. 2015). Other researchers selected a broad geographic spread of recruitment sites or regions, especially those with diverse ethnicities of populations of children living in care (e.g. Green et al. 2014). Ring-fencing resources to support the inclusion of children from minoritized ethnic backgrounds, for example, using interpreters, was a strategy adopted by King and Said (2019) in their research with asylum seeking young people. Best practice guidelines exist supporting researchers to identify approaches to improve the inclusion of participants from minoritized ethnic backgrounds, such as those produced by the National Institute for Health Research (Farooqi et al. 2023). These guidelines include the aforementioned strategies, as well as holistic approaches to building equitable and inclusive research designs from conception to dissemination. Several studies that did not make use of specific approaches to improve ethnic representativeness did acknowledge a lack of ethnic diversity (often linked to geographical location) or reliance on binary ethnicity categorizations as limitations of their research (Mannay et al. 2022; Roberts et al. 2023). This has the benefit of highlighting constraints on study generalizability and representativeness, allowing practitioners, researchers and policymakers to be cautious in the interpretation of study findings. Future research ought to openly reflect on limitations related to ethnic diversity as well as adopting strategies to address such limitations from the point of study design, informed by best-practice guidelines. This will facilitate the availability of evidence to inform anti-oppressive social work practice and assist in the avoidance of cultural and ethnic biases.

We were unable to identify enough studies with shared outcome measures that investigated sub-group differences by ethnic group to allow us to perform a meta-analysis. This may reflect the lack of large, data-harmonized datasets or large-scale quantitative research with data regarding children living in care in the UK (Allnatt et al. 2022). However, developments in administrative data linkage in the UK, and investments in secondary analysis of datasets, such as the creation of Administrative Data Research UK (https://www.adruk.org/) offer promising opportunities for researchers to investigate the role of ethnicity across a range of outcomes for children living in care in the UK. In particular, in England, the EChild and Growing up in England linked datasets offer a promising opportunity for further investigation of ethnicity effects.

Strengths and limitations

This study is the first of its kind conducted in the UK investigating the ethnic representativeness of research with children living in care. It comes at a timely moment when the body of research exploring inequalities at the intersections of care experience and ethnicity is growing (Bywaters et al. 2019; Ahmed et al. 2022; Edney, Alrouh, and Abouelenin 2023). This systematic review provides insight into current practice in ethnicity reporting amongst studies with children in care, moving the dial on conversations as to how better consistency and data harmonization can be achieved in the future.

For this study, we have used a narrow definition of children living in care that does not include adopted young people, children in need or at risk living at home or informal kinship care, which is an important direction for future work. There is a growing and rigorous body of work where gains have been made in understanding ethnic disparities relating to children’s social care focusing on factors precipitating social care intervention and entry into care (e.g. Bywaters et al. 2019) which were not able to include due to review scope focusing on the outcomes and experiences of children already living in care. Furthermore, our inclusion criteria regarding age, geographic location and the requirement for articles to include empirical research mean our review does not capture the full breadth of research concerning care-experienced adults in the UK or children worldwide. We were unable to conduct a meta-analysis examining differences in outcomes by ethnicity due to the lack of availability of studies where sub-group analyses were conducted or with shared outcomes. Finally, while not a limitation, it should be acknowledged that there is likely to be overlap between children represented in multiple study samples, particularly in secondary data analysis studies.

Recommendations and conclusion

We found generally low levels of reporting of ethnicity, even in studies where ethical and statistical reasons were unlikely to be an issue. There were also notable inconsistencies in the reporting of ethnicities. Based on this review, there is a clear need for more consistent reporting of ethnicity across studies. While ethical and analytic reasons may prevent fine-grain reporting of ethnicity, wherever possible it would benefit the field to have a harmonized collection of this data, to allow downstream analyses of combined harmonized datasets (see FAIR data principles: Wilkinson et al. 2016). Currently, this would most usefully be all projects collecting ethnicity data following the GSS harmonized five-category system (Asian, Black, Mixed, Other, White) at a minimum in line with the findings of the Race Disparity Unit (2023), with further detail added where feasible and possibly to meaningfully explore. It should be noted that this is imperfect and received fair criticism (e.g. Bhopal 2004; Burton, Nandi, and Platt 2010). It is likely that these categories will evolve over time. Nevertheless, as it currently stands, the field being consistent with this five-level reporting would be a marked improvement to current practice and would support more sensitive informing of policy and practice in a way that can empirically draw on commonalities and differences between children from different ethnic backgrounds.

Acknowledgements

We thank the authors and participants of all included research studies. We would also like to acknowledge the support of Shayma Izzidien, Rachael Stemp, Professor Nick Midgley and the INCLUDE research team, who informed the approach taken in this review.

Supplementary data

Supplementary data is available at British Journal of Social Work online.

Conflicts of interest. None declared.

Funding

This study is an output of the ReThink Programme, funded by the UK Research and Innovation (UKRI) (Medical Research Council, MRC) awarded to RH and LH (Grant number: MR/W002132/2).

References

Author notes