Methodology: Who are the children entering care in England?

The findings reported in our article, Who are the children entering care in England?, are based on data from the Growing Up in England (GUiE) dataset. GUiE contains the 2011 Census record for every individual linked to data from a bespoke extract of the Department for Education's feasibility All Education Dataset for England (AEDE). The feasibility AEDE is a large, longitudinal, record-level dataset covering predominantly government-funded education.

GUiE is currently made of up two waves. The first wave contains information on the characteristics of individuals and their households from the 2011 Census, linked to the feasibility AEDE, which contains education data from the National Pupil Database (NPD) and Individualised Learner Records (ILR). The education data in the first wave cover the period from the academic year 2001 to 2002 to the academic year 2014 to 2015. The data also contain information on student characteristics from the school census and attainment information from the Young Person's Matched Administrative Dataset and, for those in further education, information on the types of qualification undertaken.

The second wave of GUiE covers the 2010 to 2011 academic year to the 2014 to 2015 academic year and joins additional vulnerability measures from the Department of Education (DfE) to the first wave to provide further de-identified individual-level information from the following datasets.

Children looked after (CLA) dataset: this contains data on children who receive accommodation from the local authority for a continuous period of more than 24 hours (for example, a child who receives a custodial sentence following conviction of a criminal offence and goes into a young offender institution); are subject to a care order (to put the child into the care of the local authority); or are subject to a placement order (to put the child up for adoption). In our article, we refer to these children as "children in care" or the "in care" group.

Children in need (CiN) dataset: this contains data on children who are unlikely to reach or maintain a satisfactory level of health or development; whose health or development will be significantly impaired without the provision of children's social care services; or who are disabled.

Exclusions dataset: this contains data on academic year of exclusion, reason for exclusion, type of exclusion (for example, suspension or permanent exclusion) and whether a child has been excluded on more than one occasion.

Absences dataset: this contains data on the types and frequencies of authorised and unauthorised absences, and persistent absenteeism.

English School Census (ESC) dataset: this contains information on the special education needs, health and disabilities of pupils in state-funded schools in England. However, this data table was not available to us at the time of conducting our analysis, so no findings are reported from the ESC.

Our analysis uses 2011 Census data from Wave-1 as well as Wave-2 data from the CLA, Exclusions and Absences tables.

There are several data limitations that are relevant for our analysis of children in care. Firstly, GUiE does not capture the whole population of children in care in England because to be included in GUiE, individuals must appear in both the feasibility AEDE and the 2011 Census. Also, while pupils in local authority-maintained schools are automatically included in the data, some attainment data for those who are home-schooled or attend independent schools are only captured in the NPD if they are taking regulated qualifications. In addition to this, because of the way in which Wave-2 datasets were linked to Wave-1 (matching IDs in Wave-2 to the AEDE), if a child was present in Wave-2, but was not present in the AEDE, their records would not be linked to the 2011 Census.

Secondly, information on the individual and household characteristics in GUiE are a snapshot from the 2011 Census, rather than longitudinal. This means that there will be a lag between the individual and household circumstances in the 2011 Census and the time a child is first seen in care in the data.

Thirdly, GUiE does not contain information on the parents or main carers of the children in dataset. As such, analysis of the individual characteristics of parents or main carers of children is not possible. Furthermore, the variables contained in each data table in GUiE Wave-2 are a subset of the complete set of variables available to DfE. Finally, GUiE is currently seven years old and so the characteristics reported in our article may not necessarily reflect the characteristics of children in care in 2022.

Despite the limitations described, GUiE is the first quantitative dataset that enables analysis of a broader set of personal characteristics than what is currently published by DfE and provides information on the household circumstances of vulnerable children. DfE currently publish the Children looked after in England including adoptions data and the Outcomes for children in need, including children looked after by local authorities in England statistics. Consequently, we believe that the findings from our article will be valuable to policymakers, third sector organisations and the public.

GUiE is available to researchers in the Secure Research Service (SRS). Further information about GUiE and its research potential can be found in the Growing Up in England user guide under the "assets" tab, via the Administrative Data Research UK web page or in our Growing Up in England (GUiE) releases.

The aim of our analysis was to examine baseline characteristics of children in care. Therefore, our main group of interest are children who were not looked after at the time of the 2011 Census who subsequently went into care. Specifically, children in our main group of interest had to meet two criteria.

The first period of being looked after in the data must occur after the 2011 Census: this is so that we can examine individual, family and household characteristics before the child goes into care by examining 2011 Census data. Note that this criterion implies a cut-off date of 31 March 2015 because of the period covered by Growing Up in England (GUiE) Wave-2. Furthermore, we cannot conclusively determine if the first period of care observed in the GUiE Children looked after (CLA) tables is the child's first period of care in their life as the dataset does not currently contain historical information for all the children who appear in the GUiE CLA dataset.

The basis for becoming looked after can be for any reason apart from respite care: that is, the child should not be in care because they are being looked after under an agreed series of short-term placements (for example, a short break for young carers). This is consistent with the Department for Education's Children looked after in England including adoptions and Outcomes for children in need, including children looked after by local authorities in England publications.

In our article, we refer to children meeting both criteria as the "in-care" group or "children in care". There were 29,700 individuals (rounded to the nearest 10) identified as a child in care according to the criteria specified in this article.

Unaccompanied Asylum-Seeking Children (UASC) are also included in the analysis, but we are not able to show any breakdowns by this characteristic because of low observation numbers.

The comparison group comprises children who did not go into care after the 2011 Census in the data. Note that children who were previously in care but are not in care at the time of or after the 2011 Census are also included in this group. There were 3,757,700 individuals in the "not in care" group.

Note that sample sizes may vary depending on the characteristics being examined in our article.

Determining the reference points for the "in care" and "not in care" groups for non-2011 Census characteristics

Reference periods need to be determined when comparing non-2011 Census characteristics between children in care to children not in care, such as absences and exclusions. As the aim of our analysis was to examine baseline characteristics of children in care, the reference point for children in the "in care" group for absences and exclusions was taken to be the latest information before the child goes into care. Specifically, for children in care, this was identified as the last completed academic year before the child is first seen in care for absences, and the reference point for exclusions was the last three completed terms before the child is first seen in care.

As an example, if a child goes into care in January 2013 (so in the spring term of the 2012 to 2013 academic year), the relevant absences data were drawn from the 2011 to 2012 academic year while the relevant exclusions data were drawn from the spring term 2011 to 2012, summer term 2011 to 2012 and autumn term 2012 to 2013. This approach was taken to ensure that absences or exclusion instances occurring after the child went into care were not included in our analysis. Our first preference was to use the last three completed terms for both absences and exclusions data as it is the timelier of the two options. However, absences data in Growing Up in England (GUiE) are not provided on a termly basis.

For the "not in care" group, the reference point was February 2013. The decision to use February 2013 as the reference point was based on the median number of months a child went into care after the 2011 Census. The median was 23 months. Consequently, for absences, data were drawn from the last full academic year before February 2013 (so from the 2011 to 2012 academic year). Exclusions data were drawn from the last three completed terms before February 2013 (so spring term 2011 to 2012, summer term 2011 to 2012 and autumn term 2012 to 2013).

Dealing with multiple schools in the same academic year

Where a child was identified as attending more than one school in the same academic year, the "main record flag" variable provided in the absence data was used to determine which observation to use when reporting statistics for absences. In the exclusions data, the "main record flag" variable was not available and nor were the variables used to derive it. Consequently, where a child was identified as being registered in more than one school in the same academic year, the observation relating to the current main (dual registration) was retained.

Dealing with inconsistencies and conflicting demographic information

Due to GUiE being a longitudinal dataset made up of multiple administrative datasets, there were several instances where demographic information such as sex and ethnicity were inconsistent across time or conflicting between sources. In these instances, the most recent value was used to either backfill missing information or reconcile inconsistencies within the same data source. For data conflicts in demographic variables across sources, we used 2011 Census values over those reported in the Children looked after (CLA) dataset because the CLA dataset is completed by the social worker rather than by a family or household member. No demographic information was provided in the GUiE Absences or Exclusions table.

Aggregation of placement type

The placement type variable analysed in our article is aggregated from the original placement type variable available in GUiE. The following levels were aggregated:

• "Fostered placement" is made up of "Foster placement with relative or friend" and "Placement with other foster carer"

• "Placed for adoption" is made up of "Placed for adoption with parental or guardian consent with current foster carer", "Placed for adoption with parental or guardian consent not with current foster carer", "Placed for adoption with placement order with current foster care" and "Placed for adoption with placement order not with current foster carer"

• "Independent or semi-independent living" is made up of "Independent living" and "Semi-independent living accommodation not subject to Children's homes regulations"

• "Children's home or secure children's home" is made up of "Secure children's home's" and "Children's home subject to Children's homes regulations"

• "Other care institution" is made up of "Residential care home", "NHS, Health Trust or other establishment providing medical or nursing care" and "Family centre or mother and baby unit"

• "Temporary arrangement" is made up of "All types of temporary move", "Temporary periods in hospital", "Temporary accommodation while normal foster carer is on holiday" and "Temporary accommodation of seven days or less, for any other reason"

Absences

A pupil's overall absence rate is the total number of overall absence sessions as a percentage of the total number of possible sessions available to that pupil in each school establishment. Overall absence is the sum of authorised and unauthorised absence and one session is equal to half a day. In our analysis, we refer to overall absence as "absence".

Care group

In our analysis, a child is in care if they met all of the following criteria:

• appeared in the Growing Up in England (GUiE) Children looked after (CLA) dataset

• their first period of care as seen in the data started after the 2011 Census (so on or after 28 March 2011)

• the reason for going into care was not because of an agreed series of short-term breaks, that is, the child is not looked after because they are in respite care

Children that did not meet all of these criteria were allocated to the "not in care" group.

Category of need

Category of need relates to the reason the child originally became looked after and should remain the same throughout their period of care. It records the main reason why a child is being provided with services and provides further insight as to why a particular child is being looked after.

Child in need

A child in need is defined under the Children Act 1989 as a child who is unlikely to reach or maintain a satisfactory level of health or development, or their health or development will be significantly impaired without the provision of children's social care services, or the child is disabled.

Educational deprivation

No person in the household has at least Level 2 education (for example, GCSE at grades A* to C), and no person aged 16 to 18 years is a full-time student.

Employment deprivation

Where any member of a household, who is not a full-time student, is either unemployed or long-term sick.

Episode of care

The period during which a child is looked after by the local authority is broken down into "episodes" of care. Each episode represents a period of being looked after under the same legal status and in the same placement. When either the legal status or the child's placement changes, a new episode must begin. Episodes must involve at least an overnight stay, which means that an episode cannot start and end on the same day.

Exclusions

Exclusions, where mentioned in our article, refer to both suspensions (formerly known as fixed-period exclusions) and permanent exclusions.

Health and disability deprivation

Any person in the household has general health that is "bad" or "very bad" or has a long-term health problem.

Household deprivation

The four dimensions of deprivation are employment deprivation, education deprivation, health and disability deprivation, and housing deprivation. A household in the 2011 Census is classified as being deprived in none, or one to four of these dimensions in any combination.

Housing deprivation

The household's accommodation is either overcrowded (with a bedroom occupancy rating of negative 1 or less), or is in a shared dwelling, or has no central heating.

Looked after child

Under the Children Act 1989, a child is looked after by a local authority if they fall into one of the following:

• is provided with accommodation for a continuous period of more than 24 hours (Children Act 1989, sections 20 and 21)

• is subject to a care order (Children Act 1989, Part IV)

• is subject to a placement order

In our analysis, we refer to these children as children in care or the in-care group.

Overcrowded household

A household is considered overcrowded by the bedroom standard if it has an occupancy rating of negative 1 or less (so the household has at least one bedroom too few for the number and composition of people living in the household).

Period of care

A period of care consists of one or more episodes, and their total duration must exceed 24 hours. A period of care can involve one or more placements and consists of one or more periods of legal status. The legal statuses of the episodes within any period must be either: 

• anything other than being looked after in a series of short term placements (so not in respite care, codes V3 or V4)
• V3 or V4 (but not both)

Permanent exclusion

A permanent exclusion refers to a pupil who is excluded and who will not come back to that school (unless the exclusion is overturned).

Suspension

Suspension refers to a pupil who is excluded from a school for a set period of time. A suspension can involve a part of the school day and does not have to be for a continuous period.

A pupil may be suspended for one or more fixed-periods up to a maximum of 45 school days in a single academic year. This total includes exclusions from previous schools covered by the exclusion legislation.

Suspensions were formally known as fixed-period exclusions.

We welcome feedback and the opportunity to learn from best practices.

This publication and the accompanying publication, Who are the children entering care in England?, is part of a programme of work exploring the characteristics and outcomes of vulnerable children. The programme is broken into three strands of research; strand one examines the baseline characteristics of children that entered care after the 2011 Census, strand two will explore the pathways and experiences of children in the care system, and strand three will analyse educational outcomes of children in care.

Future work includes analysing the factors determining the probability of a child going into care, to understand the relative contribution of each factor in isolation.