Data quality
The quality of the data has improved since the first collection of individual social worker data in 2017. This is mainly due to improved guidance and data validation, and the work undertaken by local authorities to provide accurate data in their census return.
Quality assurance checks have been carried out at each stage of the data collection and production cycle of the statistical publication. Anomalous data were highlighted and verified by contacting the local authority and late returns were pursued to ensure the overall response was as complete and accurate as possible.
The data collection included validation checks, as covered in the data collection and cleaning section. All local authorities were asked to provide comments on the return relating to any validation queries and, where applicable, asked to confirm that year on year changes were valid.
The majority of local authorities derive their child and family social work workforce data from their management information systems. This data requires checking and managerial sign off before it is submitted. However, the DfE does not collect information on the specific data checks carried out by local authorities.
Some known data quality issues are listed below:
| Data quality | Issue | |
|---|
| Comparisons between years | Since 2013 the CSWW census has collected data for each full year ending 30 September. Data has been collected on an individual social worker level basis since 2017, whereas from 2013 to 2016 it was collected at an aggregate (local authority) level. Prior to 2013, the primary source for information on the child and family social work workforce was the National Minimum Data Set for Social Care (NMDS-SC), which was a voluntary return. The changes to the data collection mean the statistics from the different sources are not directly comparable. Data quality has continued to improve since the introduction of the individual level collection, with local authorities becoming more familiar with the return and their data collection systems becoming better placed to provide the data. Furthermore, the DfE has improved the guidance and validation checks for the data collection. Year on year comparisons therefore may not represent true changes in the number of child and family social workers (and associated measures) and may in part be a result of improved reporting. There are known discrepancies between years: for example, we would expect, at local authority level, that the number of child and family social workers in post for a reporting year would equal the number of social workers in post for the preceding reporting year plus the number of starters for the reporting year minus the number of leavers for the reporting year. However, this is not the case for a number of local authorities and may be due to improved reporting over time, or differences in how the DfE’s guidance on starters and leavers has been interpreted across different years. As a result of the above issues any comparisons between years should be treated with a degree of caution. |
| Starters | The number of starters during the year is derived from the date a social worker started in a child and family social worker role at a local authority. Some local authorities have reported that they have recorded start dates when an employee has moved child and family social worker posts within the same local authority or started at the local authority in a non-child and family social worker role. However, a start date should only be recorded where the person is a new child and family social worker: this includes child and family social workers who joined a vacant child and family social worker post at a new local authority and social workers who have previously worked in the same local authority, but in a non-child and family social worker role. It should not include a move or promotion from one child and family social work position to another child and family social work position within the same local authority or starts at the local authority in a non-child and family social worker role. As a result, the number of starters reported in this publication may be higher than the actual number, as moves or promotions from one child and family social work position to another child and family social work position within the same local authority and starts in a non-child and family social worker role may be included in the figures in some instances. |
| Duplicate records | Duplicates have been removed at a headcount level to account for cases where individuals have split roles; this avoids double counting. At FTE level, duplicates have not been removed so that the FTE for each role is captured. There are some instances where the sum of the FTE for a social worker undertaking split roles is greater than 1. In this case the FTE for some local authorities may be greater than the headcount figure. In the attrition analysis, where two or more records in the same year had the same Social Work England registration number, these were combined into a single record, with FTE and cases held added together. This was to allow for a one-to-one matching process when identifying movers and attrition into agency. However, the approach taken ensures that the FTE totals at a National level were not affected. |
| FTE | An FTE of zero was recorded for some child and family social workers in the census. These workers are included in the headcount statistics. However they are not accounted for in the FTE count, since their FTE value is zero. In some instances, local authorities have reported that these child and family social workers were working on a zero hour contract or casual basis and therefore did not have an FTE. The number of FTE child and family social workers at 30 September reported in this publication is therefore likely to undercount the actual total. In 2025, 0.6% of records for child and family social workers in post at 30 September had an FTE value of zero, the same as in 2024 and down from 1.0% in 2017. |
| FTE leavers | The FTE of leavers is based on their FTE at 30 September of the previous year. This was recorded as zero for some child and family social workers who left their post during the year. As a result these workers are included in the headcount of leavers, but are not accounted for in the FTE count of leavers, since their FTE leaver value is zero. The number of FTE leavers reported in this publication is therefore likely to undercount the actual total. In the year ending 30 September 2025, 5.5% of leavers had a value of zero recorded for their FTE at 30 September of the previous reporting year, compared to 6.1% in 2024 and 7.7% in 2017. Excluding leavers who also started during the same year (and so legitimately had an FTE of 0 at 30 September of the previous year), the proportion of leavers who had an FTE of zero at 30 September of the previous year was 1.0% in 2025 and 2024 and 2.6% in 2017. The missing information also affects the turnover rate at an FTE level, as these leavers are not accounted for in the calculation. |
| Attrition and attrition rates | To determine attrition FTE and attrition rates, leavers were joined with the local authority employed workforce and the agency workforce at two separate stages. Errors and missing information for Social Work England registration numbers in data collection may have caused a small degree of over-estimation of full attrition in the figures in years ending up to 30 September 2023, because this would have prevented leavers from being identified as movers or agency attrition. For data from year ending 30 September 2024 onwards, the effect that erroneous Social Work England registration numbers have had on the attrition analysis has been minimised through improved data collection and data cleaning methods regarding the registration numbers. This has resulted in more accurate identification of social workers and in turn more accurate estimates. Where local authorities’ data has been excluded (see ‘data quality for specific local authorities’) so their local authority employed, leavers and agency has been excluded, this will slightly overestimate full attrition, as movers into these local authorities will appear as having left the workforce entirely. Where local authorities have only submitted partial or no data for their leaver workers (see ‘Data quality for specific local authorities’), their local authority employed workers were also not included in attrition calculations. As a result, full attrition will be slightly overestimated. Attrition into agency will be less affected as agency workers are included in calculations. The attrition methods used will also slightly overestimate full attrition because the method relies on a fixed cut-off date of 30 September. Any leavers who start roles on or after 1 October would have been treated as full attrition, rather than movers or agency attrition. The impact of this is likely to have been small. This issue was explored in depth in the 2022 ad hoc publication, which found that over 70% of movers started their new role within a week of leaving their old role. Future methods development will continue to explore the impact of using a fixed cut-off date. The national level figures for the number of social workers in post, the number of social worker leavers, and the associated turnover rate in the attrition file i.e. data set 5 may have small differences with equivalent national level figures in other files. This is because the attrition analyses exclude those local authorities at national level with data quality issues e.g. Sunderland in 2024 and Gateshead in 2022 (see sub-section ‘Data quality for specific local authorities’ below for further information). |
| Full-time base salary | Base salary is only calculated for full-time child and family social workers in post at 30 September 2025, for workers which had a recorded base salary greater than zero (99.9% of all full-time workers). DfE assessed data quality to be insufficient to publish statistics on an FTE basis and for part-time workers. DfE will continue to work with local authorities to improve recording practices, with a view to potentially publish this information in future years. |
| Gaps and inconsistencies | Local authorities may not be able to provide a complete and consistent set of data for all statutory data variables. In this case data is shown as not available (denoted by ‘x’) within the data files. |
Data quality for specific local authorities
2025
Kensington and Chelsea and Westminster
Kensington and Chelsea and Westminster were unable to provide data following a cyberattack in November 2025, which had a significant impact on their information management systems. Because of this, in data sets 1 to 3, figures for these local authorities are not available (indicated by ‘x’), and for data sets 1 to 4 at national and regional level, figures from the previous year were used as best estimates to replace this year’s missing data for these local authorities to calculate totals for England, London, and Inner London. Figures for Kensington and Chelsea and Westminster have been omitted from data sets 5 and 6 (relating to attrition and base salary), and are stated as ‘Not known’ in data set 7 (relating to allowances and additional payments).
2024
Birmingham
After the 2024 collection had closed, Birmingham local authority informed the Department that there were data quality issues with the figures they reported in the collection. This affects their data on agency workers, caseload and sickness absence. To reflect these issues:
For the national and regional figures, 2024 data for Birmingham has been:
- Included in the cases held and average caseload figures, agency worker counts, and agency workers covering vacancies counts.
- Excluded from the sickness absence figures, sickness absence rates, agency worker rates, agency workers covering vacancies rates and vacancy agency cover rates (Birmingham has also been excluded from the denominator of these rates).
2024 data for Birmingham has been provided as ‘u’ in the underlying data for these measures to indicate low reliability.
In addition, the FTE agency workers figures reported for Birmingham in 2024 have been included in the figures for cases held, social and agency workers holding cases, and average caseload, at regional and national level.
Sunderland
During the 2025 collection, Sunderland local authority informed DfE that the data they had submitted for 2024 for FTE leavers was incorrect and a significant undercount of the true figure. To reflect this issue the reported 2024 FTE leavers data for Sunderland has been:
- Included in the national/regional figures for FTE leavers (apart from in data set 5 for attrition where the figures have been excluded).
- Excluded from the 2024 national/regional rates for turnover and the attrition analyses; Sunderland’s figures for FTE social workers in post have also been excluded from these calculations.
- Presented as ‘u’ in the underlying data at local authority level to indicate low reliability; the turnover rate has also been presented as ‘u’.
Bristol
During the 2025 collection, Bristol local authority informed DfE that the data they had submitted for the number of working days lost due to sickness absence for 2024 was incorrect and a significant undercount of the true figure. To reflect this issue the reported 2024 working days lost figure for Bristol has been:
- Excluded from the 2024 national/regional figures and excluded from the calculation of the national/regional sickness absence rates. The figure for social workers in post for Bristol has also been excluded from the calculation of the rates.
- Presented as ‘u’ in the underlying data at local authority level to indicate low reliability; the rate has also been presented as ‘u’.
Kensington and Chelsea
Analysis during the 2025 publication indicated that the reported numbers of FTE agency workers and FTE agency workers covering vacancies for 2024 were incorrect and significant undercounts of the true figures. To reflect this issue these figures have been:
- Included in the national/regional figures and rates for FTE agency workers and FTE agency workers covering vacancies.
- FTE agency workers presented as ‘u’ in the underlying data at local authority level to indicate low reliability.
- FTE agency rates, FTE agency workers covering vacancies rate and FTE vacancy agency cover rate presented as ‘x’ in the underlying data at local authority level to indicate not available.
Bolton
During the 2025 collection, Bolton local authority informed DfE that the data they had submitted as part of the 2024 census for cases held was incorrect and an undercount of the true number. Caution should therefore be taken when interpreting this figure.
Liverpool
During the 2025 collection, Liverpool local authority informed DfE that the data they had submitted as part of the 2024 census for agency workers was incorrect and an undercount of the true number. Caution should therefore be taken when interpreting this figure.
2023
Cambridgeshire
During the 2024 collection, Cambridgeshire local authority informed DfE that the data they had submitted for the number of working days lost due to sickness absence for 2023 was incorrect and a significant undercount of the true figure. To reflect this issue the reported 2023 working days lost figure for Cambridgeshire has been:
- Excluded from the 2023 national/regional figures and excluded from the calculation of the national/regional sickness absence rates.
- Presented as ‘u’ in the underlying data at local authority level to indicate low reliability; the rate has also been presented as ‘u’.
Sefton
After the 2023 collection had closed, Sefton local authority informed DfE that the data they had submitted for the number of working days lost due to sickness absence for 2023 was incorrect and a significant overcount of the true figure. To reflect this issue the reported 2023 working days lost figure for Sefton has been:
- Excluded from the 2023 national/regional figures and excluded from the calculation of the national/regional sickness absence rate; the figure for social workers in post for Sefton has also been excluded from the calculation of the rates.
- Presented as ‘u’ in the underlying data at local authority level to indicate low reliability; the rate has also been presented as ‘u’.
Sutton
After the 2023 collection had closed, Sutton local authority informed DfE that the data they had submitted for the number of FTE agency workers was incorrect and a significant undercount of the true figure. To reflect this issue, the reported FTE agency worker figure for Sutton for 2023 has been:
- Included in the 2023 national/regional figures for FTE agency workers.
- Excluded from the 2023 national/regional rates for FTE agency workers and agency workers covering vacancies; figures for Sutton on social workers in post and agency workers covering vacancies have also been excluded from these rates.
- Presented as ‘u’ in the underlying data at local authority level to indicate low reliability; the agency worker rate and agency worker covering vacancies rate has also been presented as ‘u’.
In addition, the FTE agency workers figures reported for Sutton in 2023 have been included in the figures for cases held, social and agency workers holding cases, and average caseload, at local authority, regional and national level.
During the 2024 collection, Sutton local authority informed DfE that the data they had submitted for 2023 on the number of vacancies and working days lost to sickness were incorrect and an overcount of the true figure. Caution should therefore be taken when interpreting these figures and the sickness absence rate.
Devon
During the 2024 collection, Devon local authority informed DfE that the data they had submitted as part of the 2023 census for vacancies was incorrect and an undercount of the true number. Caution should therefore be taken when interpreting this figure.
Lancashire
During the 2024 collection, Lancashire local authority informed DfE that the data they had submitted as part of the 2023 census for social workers in post was an undercount of the true number. Caution should therefore be taken when interpreting this figure.
2022
Gateshead
After the 2022 collection had closed, Gateshead local authority informed DfE that the data they had submitted for leavers was incorrect and a significant undercount of the true number. To reflect this issue the reported 2022 leavers data for Gateshead has been:
- Included in the 2022 national/regional figures for leavers.
- Excluded from the 2022 national/regional rates for turnover and the attrition analyses; the figures for social workers in post have also been excluded.
- Presented as ‘u’ in the underlying data at local authority level to indicate low reliability; turnover has also been presented as ‘u’.
Bury
After the 2023 collection had closed, Bury local authority informed DfE that the data they had submitted as part of the 2022 census for vacancies was incorrect and an overcount of the true number. Caution should therefore be taken when interpreting this figure.
Lancashire
After the 2023 collection had closed, Lancashire local authority informed DfE that the data they had submitted as part of the 2022 census for social workers in post and for starters and leavers were incorrect and were undercounts of the true numbers. Caution should therefore be taken when interpreting this data.
Data uses
The main internal users of these statistics are officials within the DfE, who use the data for policy setting, policy monitoring and policy evaluation purposes.
The main external users of these statistics are local authorities, who use the information to compare the statistics for their own local authority with other local authorities. Ofsted and the Children's Commissioner for England also use the data and statistics.