Data and Methodology

Data Quality

Data Completeness

Some organisations who submit data for the Carers Census will not have the information available to complete all of the topics covered by the census. This is because some might only provide certain support services but not carry out support plans, whilst others may be commissioned to carry out support plans but not to provide support services. As such, the analysis presented in this publication is largely based on records where the information is available.

Some organisations have been unable to return information on all the carers they support due to differences in data recording. For instance, some organisations are only able to provide information on carers with a support plan in place, but will have also supported other carers whose information was not recorded in the same way. There were also some areas where a return was received from either social work or third sector but not from both, or from neither, leading to gaps in coverage across Scotland. As such, the figures reported here will be an undercount of the true number of carers being supported by local services.

The figures presented here will also not be reflective of the true number of Adult Carer Support Plans (ACSPs) and Young Carer Statements (YCSs) or the support which carers receive under the Carers (Scotland) Act 2016. This is partly due to differences in data recording, but also due to differences in the way support plans are implemented. Some areas will only use an Adult Carer Support Plan or Young Carer Statement as an escalation if the carer’s situation worsens. Others will have conversations with the carer about their role and needs and put support in place, but will not record the data as an ACSP or YCS. These issues should be considered when interpreting the results in this report.

Rounding and Comparability

All figures presented in this report are rounded and refer only to carers identified in the Carers Census rather than all unpaid carers living in Scotland. Therefore, statements such as ‘X% of carers were reported to have experienced impacts on their health’ refer only to carers reported on within the Carers Census rather than all unpaid carers.

The results in this report should not be directly compared to the 2018-19 publication, since the 2018-19 figures cover a 6-month period rather than a full year. It also important to note that comparisons between 2022-23 and 2023-24 will be based on the number or records available in each year, which may vary. For example, in 2022-23 there were 22,950 records with information available on support provided, compared with 27,510 records in 2023-24. As such, comparisons and changes over time should be interpreted with caution.

The data for the charts are available in the supporting documents.

We will continue to review the Carers Census for future publications, working with organisations who submit data to make further improvements to the process and to make the final data more useful for users.

De-duplication Process

As unpaid carers can sometimes be supported by more than one local service, it is possible for information on the same carer to be submitted by multiple organisations. To ensure that carers are not being double counted in the final results, the figures presented in this report refer only to records that have been de-duplicated. This means that only one record per unpaid carer is included in the final results. The process has been further developed and extended from 2022-23, to ensure that where possible, all unique records are retained.

First, instances where an organisation has returned more than one record for the same carer are examined. If information for the same carer is split over several records, these are combined into a single record that contains all the information for that carer.

Next, records where month and year of birth, gender or data zone were unknown or missing were removed. This is because all three of these identifiers are required in order to create an accurate enough de-duplication ID to allow us to determine if records submitted by different organisations refer to the same person or not. Of the records submitted in 2023-24, 7% were removed due to missing identifiers.

Initial deduplication used the data provider and Carer ID submitted, in conjunction with date of birth. If multiple records were submitted by an organisation that represented the same carer, the duplicates were resolved. If all information submitted across duplicate records was consistent, one was retained; otherwise if the level of completeness varied across the records, these were combined into a single record containing all information available for that carer.

If the Carer IDs were consistent but the date of birth was different, thus suggesting the records represented different carers, these Carer IDs were recoded to ensure that further processing that utilised the Carer ID treated such records as distinct.

De-duplication IDs were then created for each remaining record by combining the three identifiers: month and year of birth, data zone and gender. All records where the de-duplication ID appeared only once were included in the final analysis. As above, duplication of Carer IDs within these records was investigated, given that these IDs were often incremental numeric values. For example, multiple records with Carer ID 1 were submitted by different organisations, representing different carers. As a result, any duplicate Carer IDs identified during this stage were recoded to ensure uniqueness of this identifier.

In cases where the de-duplication ID was the same for multiple records, further analysis was undertaken to determine if those records referred to different carers or were duplicate records:

1. If multiple records submitted by a single organisation had the same de-duplication ID, but different Carer IDs (e.g. Carer 1 and Carer 2), it was assumed that these records referred to different carers. As such, these records were recoded to ensure the uniqueness of the Carer ID.
2. In cases where the same system was used by multiple providers (e.g. different Carer Centres run by the same organisation) and so used the same Carer IDs, a single record was taken for each carer. If providers each returned different parts of the data, these were combined into a single record.
3. The remaining records with the same de-duplication IDs were considered to be duplicates, and therefore were resolved into a single record per carer.

As a result of the de-duplication process outlined above, 86% of the records submitted in 2023-24 were included in the final data analysis – an increase from 2022-23 (79%). This is likely due to the further development and extension of the de-duplication process, resulting in more unique records being retained.

Table 1: Number of records included in analysis following de-duplication

Source: Carers Census, Scotland, 2023-24.

Analysis of records removed during de-duplication process

The de-duplication process removed 8,490 records (14% of all records submitted) from the dataset in 2023-24. Further analysis was carried out on these records in order to ascertain if certain groups of carers were impacted more than others.

Some areas were more impacted than others by the de-duplication process. However, this was not necessarily due to data quality issues such as missing identifiers. Areas where organisations work together to provide unpaid carers with support will be more likely to return information on the same people, which would lead to more records being removed during the de-duplication process.

For example, if a Carers Centre led on conversations with the carer to put a support plan in place while the Local Authority provided the support needed, then both organisations would return some information on the same carer. Therefore, to avoid double counting, the information would be combined into a single record to be included in the final analysis and so one record would be removed from the overall total.

In 2023-24, 10% of records for 18-64 year olds and 12% of records for 65+ year olds were removed. This is higher than the 0-18 year old age group, for whom 7% of records were removed during the de-duplication process. This means that the de-duplication process affected adult carer records (particularly older adult carers) more than young carer records.

Similar proportions of records for male (11%) and female (13%) carers were removed in 2023-24. However, there was slightly more variation across ethnic groups, with the proportion of records removed varying between 8% for Mixed or multiple ethnic groups, and 19% for African, Caribbean or Black ethnic group (not including records where ethnic group was missing or not known).

The proportion of records removed for each deprivation decile (as measured using the Scottish Index of Multiple Deprivation (SIMD)) varied, with a consistent trend seen of higher proportions of records being removed in the less deprived SIMD deciles. In 2023-24, 15% of records in the least deprived SIMD decile were removed, compared to 10% in the most deprived SIMD decile. This means that unpaid carers from less deprived areas were less likely to be included in the final results.

Finally, an increase in the proportion of records being removed was seen with increasing intensity of care, where 17% of those providing 50+ hours of care a week were removed, compared to 9% of those providing up to 19 hours per week. This suggests that unpaid carers providing more intense caring roles are more likely to be excluded from the final results than those providing less intense care.