Transparency data

Transparency Data: Characteristics of children affected by household imprisonment in Wales

Published 10 September 2026

Applies to Wales

1. Executive Summary

1.1 Overview

This report presents analysis carried out by the Better Outcomes through Linked Data (BOLD) Partnerships team, based in the Ministry of Justice (MoJ), using anonymised data held in the Secure Anonymised Information Linkage (SAIL) Databank.

In its 2024 Labour Party manifesto, the government committed to better identifying and supporting children affected by parental imprisonment. Children affected by parental imprisonment have historically been difficult to identify within routine data systems. There is limited information on how many children are affected, where they live, and their characteristics, which has constrained the evidence available to inform policy and service planning, as highlighted in the Welsh Government’s Actioning a Schools and Prisons Independent Research Evaluation (ASPIRE) review.

This report presents descriptive statistics on a cohort of children affected by household imprisonment in Wales. The cohort includes children aged 10 to 14 on the day of the 2011 Census Wales^{[footnote 1]}, living in households where an adult was linked to imprisonment between 1 January 2011 and 31 December 2014. The aim is to describe what is recorded about this cohort in the linked data. This includes demographic characteristics, household circumstances, and selected indicators of service contact or need.

This household-based approach is used as a proxy for parental imprisonment because the administrative data held in the SAIL Databank cannot reliably identify the relationship between a child and an imprisoned adult within the household. As a result, the cohort also includes children affected by the imprisonment of a caregiver or other adult within the household. This approach provides broader population coverage than studies based on small samples, data from individual services, or families known to support organisations. It may include children who do not visit the person in prison, recognising support needs may extend beyond what can be identified in a prison setting.

This is an early example of BOLD analysis using linked Welsh population and prison data. The approach highlights limitations of the data and linkage methods while providing a broader view of children’s household circumstances than prison records alone.

A transparency data publication released alongside this report provides updated estimates on the number of children with a parent in prison in England and Wales.

1.2 Headline findings

Findings from this analysis indicate children in the study cohort experienced higher levels of socioeconomic deprivation, family vulnerability, educational need, and contact with social care services and the Children and Family Court Advisory and Support Service (CAFCASS) Cymru than those in the comparison group. Overall, there was little difference in demographic characteristics (sex, age and ethnicity) between the two groups.

Key headline findings:

• larger and more complex households were more common among children in the study cohort, including single-adult family households (30% vs. 24%) and other complex family arrangements (12% vs. 7%)
• a higher proportion of children in the study cohort experienced two or more forms of deprivation (52% vs. 20%)
• short custodial sentences were most common, with half of imprisoned household members (50%) serving sentences of less than six months
• free school meal (FSM) eligibility was more than twice as common among children in the study cohort (71% vs. 30%), indicating higher levels of economic disadvantage
• special Educational Needs (SEN) provision was recorded for over two-thirds of children in the study cohort (68%), compared with 43% of children in the comparison group
• interaction with social care services or CAFCASS Cymru was nearly four times as common among children in the study cohort (29% vs. 8%)
• previous involvement with safeguarding services was more common among the study cohort, including a higher proportion of children who had been previously looked after or recorded on the child protection register, although the numbers affected were small
• family and household vulnerability factors were over five times more common among the study cohort (11% vs. 2%), including parental domestic abuse, substance misuse, mental ill health, physical ill health, and learning disabilities

These descriptive statistics suggest that the needs and circumstances recorded among children in the study cohort extend beyond the individual child to the wider household. This highlights the importance of coordinated early support for affected children and families.

1.3 Methodology   

The analysis used anonymised linked data accessed through the SAIL Databank. SAIL securely links anonymised health and administrative datasets primarily about the population of Wales. Results of the analysis are presented at cohort level. The analysis does not identify individual children or families.

The baseline cohort was based on 2011 Census Wales data and included children aged 10 to 14 on the day of the census (27 March 2011). Most outputs use Census information about the child, the household, and the adult linked to imprisonment. The analysis also used additional linked education, social care, and prison datasets to describe selected indicators of service contact or need.

Children were included in the study cohort if they were identified as living in a household in Wales where an adult was linked to imprisonment between 2011 and 2014 and also included in the Welsh census. A comparison group was established comprising all other eligible children in Wales aged 10 to 14 who were not identified as affected by the imprisonment of a household member during the study period.

This report uses descriptive statistics only to describe patterns observed. It does not establish causal relationships or test for statistically significant differences. Differences observed between children affected by household imprisonment and the comparison group may reflect household imprisonment, underlying differences between the groups, or a combination of both.

This work was produced using deidentified administrative data accessed through the SAIL Databank. Where data from organisations other than MoJ have been used, their inclusion does not imply endorsement by those organisations of the interpretation or analysis presented. This work uses research datasets which may not exactly reproduce Accredited Official Statistics aggregates.

 1.4 Limitations

The linkage demonstrates the value of SAIL data for criminal justice research by enabling prison records to be linked with wider population data. However, there are key limitations to this approach:

• the lack of a single unique identifier for both children and adults limits the completeness and accuracy of data linkage
• the study cohort was identified using prison custody records. Prison data are collected mainly for operational purposes and whilst they provide comprehensive coverage of people in custody some information may be incomplete or recorded inconsistently because the data are collected primarily for operational rather than research purposes
• some indicators reflect contact with specific services rather than the full extent of children’s experiences or needs
• the household-based approach cannot identify children whose imprisoned parent was not recorded as living in the same household at the 2011 Census, for example where parents were separated

This publication presents findings from Wales only. Future work with local authorities in England could build on this approach, as also referenced in the “Updated estimates of children with a parent in prison” publication. Following the manifesto commitment in 2024, the government is taking action with pilots set up by MoJ to trial the operational recognition and support of children with a parent in prison, working closely with local authorities.

2. Introduction

2.1 The Better Outcomes through Linked Data (BOLD) Partnerships

BOLD is a cross-government programme led by the MoJ. It was created to show how people with complex needs can be better supported by linking and improving the government data held on them in a safe and secure way.

The BOLD programme demonstrates the value of linking data to improve our evidence on what works in four main areas: supporting victims, reducing homelessness, tackling substance misuse, and reducing reoffending.

You can find more information on the BOLD programme at Ministry of Justice: Better Outcomes through Linked Data (BOLD) - GOV.UK.

This publication has been produced as part of the BOLD Prevention workstream and complements the Ministry of Justice’s “Updated estimates of children with a parent in prison” publication.

2.2 Background

Children affected by parental imprisonment experience numerous difficulties across health and social outcomes and so a cross-government response is critical in order to address this issue.

In 2024, the Labour Party manifesto committed to better supporting and identifying children affected by parental imprisonment. The MoJ and the Department for Education (DfE) are jointly responsible for taking this commitment forward - recognising children affected by parental imprisonment is led by the MoJ.

Responsibility for supporting children who may be vulnerable due to parental incarceration sits with DfE in England and the Welsh Government in Wales, as Education is a devolved responsibility. The MoJ and DfE are working together to pilot approaches to identify children with a parent in prison who live in pilot areas in England and connect them with focused support according to their individual needs.

Strengthening family ties is also a vital aspect of the work of the MoJ and His Majesty’s Prison and Probation Service (HMPPS) to support family cohesion, promote wellbeing and safety, and enable positive rehabilitation and desistance from crime, as long as it is in the child’s best interest to maintain a strong relationship with their parent or carer who is in prison.

The Farmer Review on “The Importance of Strengthening Prisoners’ Family Ties (2017)” pointed to significant gaps in official data on the impact of parental imprisonment, noting that there was not a process within the MoJ to centrally record the number of children affected. Prisoners may be unwilling to share information about their children to staff for a range of reasons, including concerns about involvement from statutory services.

The challenge of collecting data on this group means there is a lack of evidence on the number of children experiencing parental imprisonment and how this may affect their wellbeing and future outcomes. The lack of reliable data creates a challenge when designing policy. Building the evidence base and improving our collective understanding of the impact of the issue is a critical step towards being able to develop an effective cross-government approach and ensuring that the right support can be provided to the children who need it.

Since 2011 the Welsh Government has had a commitment to the United Nations Convention on the Rights of the Child (UNCRC). Children affected by imprisonment should be reflected as children with individual rights, strengths and needs and not primarily defined by the offending behaviour of an adult.

3. Datasets

3.1 Data linking 

All data for this analysis was accessed through the SAIL Databank, a trusted research environment that enables anonymised records from multiple administrative datasets to be linked securely at the individual and household level (Ford et al., 2009; Lyons et al., 2009; Jones et al., 2014).

Children affected by household imprisonment were identified through linkage between prison custody records and household-level information. For adults who entered prison after Census Day, 2011 Census household information was used to identify children living in the same household. For adults who were already in prison on Census Day, residential records from their last known address before imprisonment were used.

Once children affected by household imprisonment had been identified, their anonymised records were linked to education, social care and prison datasets available within SAIL. This enabled the analysis of demographic characteristics, household circumstances and selected indicators of service contact and need.

Not all individuals could be linked across all datasets (see Figure 2 of the Technical Guide for linkage rates). For example, 13,120 individuals out of the 18,880 individuals in the Data First Prisoner Custodial Journey Dataset (PRIS) were not successfully linked to households in the demographic datasets. This reflects differences in dataset coverage, availability of administrative records, and quality and availability of identifying information used for record linkage, which may affect the ability to establish links between data sources. Findings should therefore be interpreted as representing those individuals who could be linked across the relevant data sources.

Further details on cohort construction, linkage procedures, inclusion criteria and dataset coverage are provided in the Technical Guide.

3.2 Data sources

The analysis in this report is based on linked administrative data from the data sources shown in Table 1, all of which were accessed through the SAIL Databank.

Table 1: Data sources used in this publication

Data source	Description
ONS 2011 Census Wales (CENW)	A population-wide Census providing a snapshot of demographic, household and socioeconomic circumstances on 27 March 2011.
Welsh Demographic Service Dataset (WDSD)	Administrative dataset containing demographic and residential address information for individuals registered with a GP in Wales.
Data First Prisoner Custodial Journey Dataset (PRIS)[footnote 2]	Administrative dataset containing information from the Prison National Offender Management Information System (p-NOMIS), the MoJ’s prison operational case management system. It contains information on custodial episodes in prisons in England and Wales.
Children in Need Wales (CINW)	Information on the health, education and parental characteristics of children in need in Wales 2009–2016.
Children Receiving Care and Support Census (CRCS)	Information of children (under 18) who have a care and support plan within Wales, replacing the Children in Need Wales (CINW) census in 2016.
Looked After Children Wales (LACW)	Individual level demographic and episodic information of children who received care support 1999-2022.
Children and Family Court Advisory and Support Service Wales Family Justice Dataset (CAFW)	The Children and Family Court Advisory and Support Service (CAFCASS) provides information on children and families receiving support due to involvement in family court proceedings.
Education Wales (EDUW)	Welsh School and Pupil Data including attendance, outcomes, special needs provision and FSM eligibility.

For further information about the data sources used and their processing, please refer to the Technical Guide.

4. Methodology

The main analysis in this report focuses on children aged 10 to 14 at the time of the 2011 Census. The selection of this age group was informed by a cross-government stakeholder working group with representation from the DfE and was chosen to support an analytically consistent cohort definition across the linked datasets used in this study.

For this analysis, an adult prisoner was defined as an individual aged 18 years or over at the time of imprisonment between 1 January 2011 and 31 December 2014. This included custodial periods that began before 1 January 2011 but continued into the study period. Children aged 10 to 14 on Census Day who were subsequently identified as adult prisoners during the study period were excluded from the child cohort to prevent individuals appearing in both the child and prisoner cohorts. Further details on this definition are provided in Section 1.3 of the Technical Guide.

Although The Best Start for Life: a vision for the 1,001 critical days highlights early childhood as an important period for child development and later life outcomes, important developmental transitions occur throughout childhood. The selected age group captures a period of significant educational, social, and developmental change, including the transition from primary to secondary education.

Children in this age range are also commonly the focus of prevention and early intervention activity across education, health, and social care services, making it a particularly relevant stage at which to examine their circumstances and experiences.  

Additionally, this age range provides a strong foundation for future longitudinal analysis. By identifying children aged 10 to 14 at baseline, it is possible to examine characteristics and experiences recorded earlier in childhood whilst providing sufficient follow-up time to explore outcomes occurring later in adolescence and early adulthood in future publications, including outcomes relating to education, social care and imprisonment.

This age range is also consistent with standard age groupings used in official demographic statistics, supporting comparability with official demographic statistics and enabling findings to be interpreted within a wider population context.

For context, using linked data within the SAIL Databank, there were 2,905 children aged 0 to 17 in Wales with evidence of living in a household where an adult was linked to imprisonment between 1 January 2011 and 31 December 2014. These 2,905 children lived in 1,765 households.  

Of these children, 905 children were aged between 10 and 14 and identified as living in 730 households where an adult was linked to imprisonment during the 2011 to 2014 study period. This group is referred to as the study cohort throughout the report.

A comparison group was also created. This included 152,900 children aged 10 to 14 who were not identified through the household imprisonment linkage during the study period.

The comparison group should not be interpreted as children who were never affected by imprisonment. Some children may have been affected outside of the study period. Others may not have been captured in the available linked data.

Counts and percentages are subject to SAIL disclosure control and rounding procedures; see Technical Guide Section 1.9 for details.

5. Findings

5.1 Household and socioeconomic context

This section describes selected household and socioeconomic characteristics of children in the study cohort and the comparison group, based on 2011 Census data.

Of the 1,965 adults who were linked to imprisonment during the study period, at the time of the 2011 Census, 1,545 adults (79%) had not yet entered custody, while 230 (12%) were in prison. Information on custody status on Census Day was unavailable for a further 190 (10%) of adults.

As a result, Census-derived measures may reflect household and socioeconomic circumstances before imprisonment for some children and during a period of imprisonment for others. Where an adult was in prison on Census Day, Census information relates to the household in which the linked child was living at that time.

5.1.1 Household size

Children in the study cohort were more commonly found in households containing 6 or more people than children in the comparison group (see Table 2). The largest difference was observed for households containing 6 to 7 people (13 percentage points).

Household size includes all usual residents recorded in the household at the 2011 Census, including adults and children outside the study cohort age range. These figures reflect household composition at the time of the Census and not necessarily any subsequent changes in living arrangements. The number of dependent children is considered separately in Section 5.1.4.

Table 2: Household size of children in the study cohort and comparison group

Household size	Number in study cohort	Percentage in study cohort	Number in comparison group	Percentage in comparison group
2 people	25	3%	8,275	5%
3 people	110	12%	28,300	19%
4 to 5 people	500	55%	95,985	63%
6 to 7 people	215	24%	16,850	11%
8 or more people	60	7%	2,195	1%

5.1.2 Family size

Family size is based on the Census family definition: a married, civil partnered or cohabiting couple with or without children, or a lone parent with at least one child, living at the same address. This differs from household size, which includes all usual residents in a dwelling and may contain more than one family.

As Table 3 shows, the pattern of family size was broadly similar to that observed for household size. However, the study cohort showed a larger difference between household and family size. Compared with household size, the proportion of children in families of 6 to 7 people was 4 percentage points lower. In contrast, the proportion of children in families of 3 people was 2 percentage points higher.

Taken together, these findings suggest that children in the study cohort more commonly lived in households that included people beyond their immediate family than children in the comparison group. Family and household definitions can be complex. For example, a child and their mother (one family) may live with the child’s grandparents (another family).

Table 3: Family size of children in the study cohort and comparison group

Family size	Number in study cohort	Percentage in study cohort	Number in comparison group	Percentage in comparison group
1 to 2 people	55	6%	11,710	8%
3 people	120	14%	29,350	19%
4 to 5 people	495	55%	94,170	62%
6 to 7 people	180	20%	14,510	10%
8 or more people	50	6%	1,560	1%

5.1.3 Household composition

Household composition describes the structure of the household in which the child was living at the time of the 2011 Census.

A higher proportion of children in the study cohort lived in single-adult family households than in the comparison group (30% compared with 24%). Other complex family household types were more common in the study cohort (12% compared with 7%) as shown in Table 4. An example of a complex household might be a cohabiting couple consisting of a mother and her child, living with a partner and his parents.

Table 4: Household composition of children in the study cohort and comparison group

Household composition	Number in study cohort	Percentage in study cohort	Number in comparison group	Percentage in comparison group
One adult family household	270	30%	36,100	24%
Two-adult family household	525	58%	104,865	69%
Other complex family household type	110	12%	10,610	7%

5.1.4 Dependent children in the household

Under the 2011 Census definition, a dependent child is a person aged 0 to 15, or a person aged 16 to 18 who is in full-time education and lives in a family with a parent or grandparent. People aged 16 to 18 who live with a spouse, partner, or child are not classified as dependent children.

As shown in Table 5, a higher proportion of children in the study cohort lived in households with two or more other dependent children (48% compared with 33% in the comparison group). In contrast, households containing one other dependent child were more common in the comparison group (46% compared with 32%). Overall, children in the study cohort tended to live in households with a larger number of dependent children.

Table 5: Other dependent children in the household of the study cohort and comparison group

Dependent children	Number in study cohort	Percentage in study cohort	Number in comparison group	Percentage in comparison group
Only dependent child	185	20%	31,280	21%
One other dependent child	290	32%	70,025	46%
Two or more other dependent children	430	48%	50,295	33%

5.1.5 Household deprivation

Household deprivation is a derived Census measure based on information provided by household members. Households are classified according to whether they meet one or more deprivation criteria across four dimensions: employment, education, health and disability, and housing conditions. A household may be classified as deprived in none, one, two, three, or all four dimensions, depending on the number of dimensions in which it meets the specified criteria.

As Table 6 shows, a higher proportion of children in the study cohort lived in households that reported experiencing multiple forms of household deprivation than children in the comparison group. The most common category for the study cohort was households deprived in two dimensions (30%), while only 19% lived in households with no recorded deprivation.

In contrast, over half (51%) of children in the comparison group lived in households that were not deprived in any dimension. Overall, a greater proportion of children in the study cohort lived in households experiencing deprivation across multiple dimensions, indicating higher levels of socioeconomic disadvantage.

Table 6: Classification of household deprivation relating to children in the study cohort and comparison group

Deprivation	Number in study cohort	Percentage in study cohort	Number in comparison group	Percentage in comparison group
Not deprived in any dimension	175	19%	77,770	51%
Deprived in 1 dimension	255	28%	43,795	29%
Deprived in 2 dimensions	275	30%	21,235	14%
Deprived in 3 dimensions	165	18%	7,995	5%
Deprived in 4 dimensions	35	4%	805	1%

5.2 Stable demographic characteristics

This section describes selected demographic characteristics as reported in the 2011 Census. Information on children aged under 16 years was provided by an adult member of the household on their behalf.

5.2.1 Age distribution

The age profile of the study cohort was slightly older than that of the comparison group. Children aged 14 represented 24% of the study cohort compared with 21% of the comparison group, while children aged 10 represented 17% and 19%, respectively. The proportions of children aged 11, 12 and 13 were similar across the two groups, with each age accounting for around a fifth of children.

5.2.2 Sex

The sex distribution was similar across the study cohort and comparison group. Boys accounted for 51% of the study cohort and 52% of the comparison group, while girls accounted for 49% and 48% respectively.  

5.2.3 Ethnicity

The ethnic composition of the study cohort was similar to that of the comparison group. Children recorded as White accounted for 92% of the study cohort and 94% of the comparison group, while children recorded in all other ethnic groups accounted for 8% and 5%, respectively. Ethnicity was grouped into White and Other ethnic groups due to small numbers in some ethnic categories.

5.3 Individual circumstances

5.3.1 Disability

Disability status was derived from 2011 Census self-reported information on long-term health problems or disabilities.

Among children in the study cohort, 35 (4%) were recorded as having a long-term health problem or disability that limited their day-to-day activities a lot, 40 (4%) were recorded as being limited a little, and 830 (92%) were recorded as not being limited in their day-to-day activities.

Similar proportions were observed in the comparison group, with 2% recorded as limited a lot, 3% as limited a little, and 94% as not limited.

5.3.2 Carer status

Carer status reflects whether a child reported providing unpaid care at the time of the 2011 Census.

Around 20 children (3%) in the study cohort provided unpaid care for someone else, while 880 children (97%) did not provide unpaid care. In the comparison group, 2% of children provided unpaid care.

5.4 Safeguarding and household vulnerability

Measures of safeguarding involvement and selected indicators of household vulnerability are based on administrative records and may not capture all social care involvement.

Safeguarding records were identified from the available data sources up to age 18 or the end of dataset coverage, whichever occurred first. Available records begin in 2005 and do not capture the full childhood period for all children.

The CINW and CRCS datasets provide a snapshot of children who met the census criteria between 1 January and 31 March each year. CINW includes children with ongoing involvement from children’s social services. CRCS includes a narrower group of children whose assessed needs required an active care and support plan.

This analysis cannot determine whether safeguarding involvement occurred before, during or after a parent or caregiver’s imprisonment, or whether imprisonment contributed to the involvement.

5.4.1 Social care and CAFCASS Cymru involvement

A social care and CAFCASS Cymru involvement flag was derived using linked records from four datasets: CINW, CRCS, LACW and CAFW. Children were classified as having social care or CAFCASS Cymru involvement if they were identified in at least one of these datasets. The CAFW dataset does not capture all family justice involvement and only includes cases where CAFCASS Cymru was involved.

Around 265 children (29%) in the study cohort met this definition. The remaining 635 children (71%) were not identified in any of these datasets.

In the comparison group, 8% of children met this definition. The findings indicate that social care or CAFCASS Cymru involvement was nearly four times more common among children in the study cohort than in the comparison group.

It is important to note that engagement with social care services or CAFCASS Cymru does not necessarily indicate harm. Within the Welsh policy context, involvement may occur for a range of reasons, including care and support needs, safeguarding responses, family support, and court-related proceedings.

5.4.2 Looked after child (LAC) status

Previously looked after child (LAC) status^{[footnote 3]} was derived from the CINW dataset, which has coverage from 2005 to 2016. This measure identifies whether a child was recorded as having previously been looked after at the time of their most recent referral.

A small number of children in the study cohort were identified as having previously been looked after (15 children; 2%). Within the comparison group, 485 children were identified as having previously been looked after (0.3%).

5.4.3 Child Protection Register involvement

A child was counted if they were recorded as currently on the child protection register, had a recorded date of entry onto the register, or had a record indicating previous Child Protection Register involvement.

In the study cohort, 15 children (2%) had a record of Child Protection Register involvement, compared with 535 children in the comparison group (0.4%).

5.4.4 Recorded family and household vulnerability factors

A family and household vulnerability flag was derived from factors recorded in the child’s most recent CINW referral. Children were classified as having a recorded vulnerability factor if the referral included parental domestic abuse, parental substance or alcohol misuse, parental mental ill health, parental physical ill health, or parental learning disabilities.

In the study cohort, 100 children (11%) had at least one recorded family and household vulnerability factor. The remaining 805 children (89%) had no recorded vulnerability factors in the available data.

The proportion with at least one recorded family and household vulnerability factor was five times higher in the study cohort than in the comparison group (11% compared with 2%).

5.5 Characteristics of the imprisoned adult household member

This section describes selected characteristics of the adult household member identified in prison records during the study period, based on PRIS. Information on imprisoned household members was sourced from p-NOMIS, the operational case management system used in prisons in England and Wales. Due to data limitations, it is not possible to identify the specific relationship between the child and the household member in custody. Equivalent characteristics are not presented for the comparison group, as children in the comparison group were not identified as having experienced household imprisonment.

5.5.1 Sex of imprisoned household member

Of the imprisoned household members linked to children in the study cohort, 88% were recorded as male, while 12% were recorded as female. These figures describe imprisoned household members linked to children in the study cohort rather than the prison population as a whole. They should not be interpreted as the proportion of male and female prisoners in the wider prison population. Information on the wider prison population is available in the Ministry of Justice’s Offender Management Statistics publications. Further work would be required to determine the extent to which this finding reflects linkage methodology, characteristics of the linked population, or other factors.

5.5.2 Sentence length

Sentence length was derived using custodial start and end dates recorded in PRIS. For the custodial start date, we prioritised the first sentenced date unless the individual was recorded as being untried and remanded in custody, or no first sentenced date was available, in which case the first movement date was used instead. Further information on the derivation of custodial episode dates and limitations of this approach is provided in Section 1.4 of the Technical Guide.

Where a household member had multiple periods of imprisonment between 1 January 2011 and 31 December 2014, the earliest custodial episode within the study period was used. The effective release date was used to ascertain custodial end date.

These figures relate only to the custodial episode selected for analysis within the study period. They do not capture an individual’s full custodial history or the timing and duration of household imprisonment relative to other events in the child’s life.

As shown in Table 7, the most common sentence length category among imprisoned adult household members was less than 6 months (50%), followed by 1 to 2 years (13%). The least common sentence length category was 3 to 4 years (3%).

Table 7: Sentence length of imprisoned household members

Sentence length	Number in study cohort	Percentage in study cohort
Less than 6 months	990	50%
6 months to less than 1 year	240	13%
1 year to less than 2 years	260	13%
2 years to less than 3 years	125	6%
3 years to less than 4 years	55	3%
4 or more years	105	5%
Missing effective release date	185	9%

5.6 Education indicators

The education indicators presented in this section were derived from Education Wales datasets. Data were sourced from either the Pupil Level Annual School Census (PLASC) for children attending state-funded mainstream schools, or the Educated Other Than At School (EOTAS) Census for children educated outside mainstream school settings, including Pupil Referral Units (PRUs), hospital schools, and individual home tuition.

5.6.1 FSM eligibility

FSM eligibility was used as an additional indicator of socioeconomic disadvantage, based on recorded eligibility during the available education data coverage period (2004 to 2021).

In the study cohort, 640 children (71%) were recorded as having been eligible for free school meals, while 260 children (29%) had no recorded history of FSM eligibility.

The proportion of children with a recorded history of FSM eligibility was more than double in the study cohort than the comparison group (71% compared with 30%).

As FSM eligibility may change over time, this measure reflects whether eligibility was recorded at any point and does not indicate the timing or duration of eligibility.

5.6.2 SEN provision

SEN provision was derived using information recorded in the EDUW Pupil and EDUW EOTAS datasets. Children were classified according to whether they had a record of receiving SEN provision. Although Wales has since transitioned to the Additional Learning Needs (ALN) framework, SEN terminology is used throughout this report to reflect the source data and study period.

In the study cohort, 615 children (68%) were recorded as receiving SEN provision, while 290 (32%) were not.

The proportion receiving SEN provision was substantially higher in the study cohort than in the comparison group (68% compared with 43%).  

6. Discussion

6.1 Findings in the context of existing evidence

This report addresses a key gap identified by the Welsh Government-commissioned ASPIRE programme. The review highlighted limited systematic information on children affected by parental imprisonment and the challenges this creates for policy development, service planning and public awareness. By linking administrative records across multiple domains, this study provides new descriptive evidence on the characteristics and circumstances of children affected by household imprisonment in Wales.

The findings are broadly consistent with existing evidence. The ASPIRE review highlights that children affected by parental imprisonment may experience financial hardship, educational challenges, safeguarding concerns and exposure to other forms of adversity. In addition, this publication finds that children affected by household imprisonment more commonly live in larger households (with over a quarter living in households of six or more people), households with multiple indicators of deprivation, and more complex family structures. Around one-third lived in single-adult households. The majority were affected by the imprisonment of an adult serving a custodial sentence of less than six months. These findings provide further insight into the household circumstances of children affected by imprisonment in Wales and may help identify where support is best targeted.

Similar patterns were identified by the Thames Valley Violence Prevention Partnership (TVVPP), which found higher levels of educational support needs and greater representation in social care records among children recognised through Operation Paramount. Operation Paramount linked prisoner data from HMPPS with information held in local police systems between 2021 to 2024. Although the TVVPP study focused on a smaller cohort affected specifically by parental imprisonment, the findings are broadly consistent with those presented here.

Findings from the accompanying “Updated estimates of children with a parent in prison” publication are broadly consistent with those presented here. The estimates publication identified larger family sizes among prisoners who were parents (2.44 children per prisoner parent, compared with 1.77 children in the general population). This analysis similarly found that children in the study cohort more commonly lived in larger households and households with multiple dependent children. The estimates publication also highlighted the concentration of children with a parent in prison in more deprived areas, consistent with the higher levels of household deprivation observed among the study cohort in Wales. In addition, the estimates publication identified high levels of safeguarding-related vulnerability among children with a parent in prison, which is consistent with the broader pattern of vulnerability observed among children in this study.

The findings are intended to provide insight into the circumstances of children affected by household imprisonment and to inform further research, policy and support. They should not be used to make assumptions about the experiences, needs or likely life trajectories of individual children, nor should they be interpreted as evidence that imprisonment caused the observed differences.

As highlighted in Weaver and Nolan’s review of the evidence on families of prisoners, families affected by imprisonment may already experience social, economic or family-related challenges before imprisonment occurs. In many cases, imprisonment is not a single isolated event, and individuals may experience multiple episodes of imprisonment, release and re-imprisonment over time. As a result, it can be difficult to distinguish characteristics that pre-date imprisonment from those that arise during, between, or after periods of imprisonment. Imprisonment may compound existing disadvantage, create new challenges, or, in some circumstances, alleviate existing family stressors. This descriptive study does not seek to distinguish between these explanations.

6.2 Variation within the cohort

The findings support a key theme within both the ASPIRE review and wider literature: children affected by imprisonment are not a homogeneous group.

While some children in the study cohort were recorded as experiencing multiple forms of disadvantage, others showed relatively little evidence of recorded need across the indicators examined.

This variation in circumstances suggests that policy and practice responses should support proportionate, non-stigmatising engagement and access to support where needed, rather than assuming a single intervention pathway for all children affected by household imprisonment.

6.3 Data and methodological considerations

Household imprisonment is not routinely recorded within population, education or social care systems. Identification therefore relies on linking datasets collected for different operational purposes. Changes in household composition, dataset coverage, missing identifiers, recording practices and decisions about how household imprisonment is defined may all affect which children are identified and should be considered when interpreting findings. Further details of these methodological challenges including linkage between prison records and 2011 Census data, are provided in Section 1.3 of the Technical Guide.

Administrative data cannot capture the full lived experience of children affected by imprisonment. Qualitative research highlights important aspects of children’s experiences, including stigma, grief, trauma, family relationships, emotional wellbeing, coping strategies and the quality of support networks, which are not routinely recorded in administrative data. The findings therefore provide insight into recorded characteristics and service contact, rather than the full range of children’s experiences.

6.4 Future work

This publication contributes to ongoing work to improve understanding of children affected by household imprisonment using linked administrative data. It has, for the first time, used linked Welsh population-level data to derive a study cohort of children aged 10 to 14, supporting future longitudinal analysis of children affected by the imprisonment of an adult household member. The characteristics of this study cohort are consistent with characteristics described in the wider research literature (see Section 6.1).

Future work could examine additional characteristics and outcomes, continue to refine identification methods, investigate potential selection bias within the linked offender cohort, and extend the analysis using locally linked local authority data shares in England. This could help assess whether similar patterns are observed across different areas and data sources while accounting for regional variation. Although, the Education and Child Health Insights from Linked Data (ECHILD) Database is the closest equivalent linked data resource in England, it does not currently provide the household-level linkage needed to identify children affected by household imprisonment. As a result, local data linkages may offer the most feasible route for extending this approach in England.

Exploration of these areas could improve understanding of children’s circumstances, recorded service contact and longer-term outcomes. Further analysis is also needed to better understand how safeguarding-related information is recorded across the administrative datasets used in this study, and how data linkage and data sharing could be improved. This could help inform future policy decisions on how children affected by imprisonment are recognised and supported.

Alongside any changes to sentencing and prison release policy, the profile of children affected by imprisonment may also change. Using and repeating this analysis as necessary would provide valuable evidence on the characteristics and needs of affected children. It could also help ensure that the impacts of reforms on children and families are considered alongside impacts on offenders and the wider justice system.

7. Contact details and feedback

Any queries on this publication please contact CAPI@justice.gov.uk

8. Acknowledgements

We would like to acknowledge all the data providers who make anonymised data available for research. We also acknowledge the Ministry of Justice’s Data First programme , which enables secure access to linked administrative data for research and analysis. We also thank the SAIL Databank Quality Assurance team for their review, advice and support throughout the project.

The analyses, findings and interpretations presented in this publication are those of the authors and not necessarily those of the data providers, SAIL Databank, Swansea University or partner organisations.

9. References

Department of Health and Social Care. ‘The best start for life: a vision for the 1,001 critical days’. Department of Health and Social Care, 2021

Farmer M. ‘The Importance of Strengthening Prisoners’ Family Ties to Prevent Reoffending and Reduce Intergenerational Crime’. Ministry of Justice, 2017

Ford DV, Jones KH, Verplancke JP, Lyons RA, John G, Brown G and others. ‘The SAIL Databank: building a national architecture for e-health research and evaluation’. BMC Health Services Research 2009: volume 9, issue 1, article 157

Jones KH, Ford DV, Jones C, Dsilva R, Thompson S, Brooks CJ and others. ‘A case study of the Secure Anonymous Information Linkage (SAIL) Gateway: a privacy-protecting remote access system for health-related research and evaluation’. Journal of Biomedical Informatics 2014: volume 50, pages 196 to 204

Labour Party. ‘Change: Labour Party Manifesto 2024’. Labour Party, 2024.

Loucks N, Beresford S, Wright P, Raikes B, Rees A, Kenny F and Stevenson S. ‘Improving wellbeing and educational outcomes for children in Wales affected by parental imprisonment: literature, policy and practice reviews’. Welsh Government GSR report 25/2025, 2025

Lyons RA, Jones KH, John G, Brooks CJ, Verplancke JP, Ford DV, Brown G and Leake K. ‘The SAIL databank: linking multiple health and social care datasets’. BMC Medical Informatics and Decision Making 2009: volume 9, article 3

Olphin T, Massie R, Miller O and Blakemore R. ‘What Works Series: Operation Paramount - Statistical analysis of children in Oxfordshire affected by parental imprisonment’. Thames Valley Violence Prevention Partnership, Kidlington, 2025

Rodgers SE, Demmler JC, Dsilva R and Lyons RA. ‘Protecting health data privacy while using residence-based environment and demographic data’. Health and Place 2012: volume 18, issue 2, pages 209 to 217

Rodgers SE, Lyons RA, Dsilva R, Jones KH, Brooks CJ, Ford DV and others. ’Residential Anonymous Linking Fields (RALFs): a novel information infrastructure to study the interaction between the environment and individuals’ health’. Journal of Public Health 2009: volume 31, issue 4, pages 582 to 588

Scott S, Griffin N, Crowe L, Shildrick T, Young T, Loucks N and Minson S. ‘“There are a million and one other ways to punish someone than sending them away…” Exploring the impact of having a family member in prison on children and young people’s health and emotional wellbeing: a qualitative longitudinal study’. Children and Youth Services Review 2026: article 108904

Weaver B and Nolan D. ‘Families of Prisoners: A Review of the Evidence’. University of Strathclyde, 2015

1. Census Day is the reference date used for the population count. Census responses are intended to describe an individual’s circumstances on that day, irrespective of when the census form was completed. For the 2011 Census, Census Day was 27 March 2011. ↩

2. Ministry of Justice, released 01 January 2022, Health Data Research Gateway Metadata Catalogue, dataset, Ministry of Justice Data First Prisoner Custodial Journey - England and Wales ↩

3. The term “looked after child (LAC)” is used throughout this report because it reflects the terminology used in the source dataset. In the Welsh policy context, children may also be referred to as “children looked after” or “children who are care experienced”, depending on the context. ↩