Research and analysis

Local Civil Society Infrastructure (LCSI) R&D Programme - secondary data analysis

Published 11 July 2025

This research was supported by the R&D Science and Analysis Programme at the Department for Culture, Media & Sport. It was developed and produced according to the research team’s hypotheses and methods. Any primary research, subsequent findings or recommendations do not represent government views or policy.

Key Terms and Definitions

Local civil society infrastructure organisation (LCSI) – organisations registered as charities in England and Wales that meet at least one of the following criteria developed by 360Giving (Kane and Cohen, 2023):

• Their activities are classified as “Acts As An Umbrella Or Resource Body”

• Their beneficiaries include “Other Charities Or Voluntary Bodies”.

LCSI type – LCSI organisations are categorised according to the 360Giving typology (Kane and Cohen, 2023):

• General: local Councils for Voluntary Service, or equivalent organisations, that provide services and support to all charities in an area
• Volunteer centre: providing volunteer services in the local area. Where the volunteer centre is hosted by a CVS the “General” category was used instead.
• Specialist: providing support to a particular group of charities in a local area
• Service provider: providing specific services, such as a community accountancy service. Where the service provider is hosted by a CVS the “General” category was used instead.

Civil society organisation (CSO) – organisations registered as charities or Community Interest Companies (CICs) in England and Wales.

Local authority – local government administrative units in England as of April 2023. Local government undergoes restructuring, which requires adoption of a common structure for a consistent comparison across time. The focus of this report is on the 2023 boundaries. There are four types of local authority analysed in this research (UK Government, 2016):

• Unitary authorities: There are 62 unitary authorities and they provide all local government services in their areas. These are mainly in the cities, urban areas and larger towns.

• London boroughs: There are 32 London boroughs and they provide nearly all the services in their area. However, the Greater London Authority (GLA) provides London-wide government, including special responsibility for police, fire, strategic planning and transport.

• Metropolitan districts: There are 36 metropolitan district councils and are responsible for all services in their area.

• District councils: There are 164 district councils and they are responsible for local services such as rubbish collection, housing and planning applications; other services are the responsibility of county councils.

LCSI coverage – a local authority is determined as having a LCSI organisation present in it if it meets at least one of the following criteria:

• Registered here: an LCSI organisation has a registered / head office address in the local authority

• Registered elsewhere but operating here: an LCSI organisation has a registered / head office address in a different local authority but reports operating the local authority in its annual return to the Charity Commission for England and Wales.

LCSI sufficiency – after extensive discussion within the research team and consultation with DCMS, a formal definition of sufficiency is not adopted. Instead LCSI sufficiency is approximated using the following measures:

1. Typical (median) expenditure of LCSI organisations with a registered / head office address in the local authority, standardised by population size.

2. Total expenditure of all LCSI organisations with a registered / head office address in the local authority, standardised by population size.

These are important measures as they reflect the spending power of LCSI organisations within an area and relates this to the number of individuals who can potentially benefit from the activities of local civil society organisations. These measures refer to LCSI organisations registered in a local authority, rather than refer to all LCSI organisations that operate in a particular local authority.

Executive Summary

1,392 LCSI organisations were active in England at some point between 1991 and 2023. The number of active organisations peaked between 2003-2006 and has been on a steady decline since then. In the latest year of observation (2023) there are 530 LCSI organisations active across England.

Most LCSI organisations are classified as providing general support services, though there has been a substantial increase in the proportion of LCSI support provided by specialist organisations. LCSI organisations have a typical income of between £100k-£1m , and the vast majority have been in operation for over ten years.

The proportion of local authorities where LCSI organisations are registered has declined since 2001 and in district councils in particular. Declining coverage appears to be partly ameliorated by the presence of LCSI organisations that are registered in a different local authority.

In general, areas with larger numbers of civil society organisations tend to have more LCSI organisations present; and LCSI organisations are more likely to be present in more deprived local authorities. The same patterns are found for LCSI sufficiency, which is measured as the total or median LCSI expenditure per head of population in a local authority.

There is no evidence that higher levels of LCSI coverage or sufficiency are associated with higher rates of civil society organisation survival or foundation.

There is little change over time in the level of LCSI expenditure per head of population in a local authority. However, there is significant variation between local authorities in a given year: for example, in 2006 one local authority had total LCSI expenditure of almost £125 per head of population, while 25% of local authorities had £1.75 or less.

Characteristics of LCSI Organisations

The research identified 1,392 LCSI organisations that were active in England at some point between 1991 and 2023. The number of active organisations peaked between 2003-2006 and has been on a steady decline since then (figure 1). In the latest year of observation (2023) there are 530 LCSI organisations active across England. While this implies a high degree of organisational dissolution in this sector, there are a number of instances where the same organisation re-registers as a new legal form (and thus appears as both a dissolution and formation) or merges with other organisations (e.g., the Volunteer Centre merges with the local Council of/for Voluntary Service).

Figure 1: Number of LCSI organisations, by year.

Organisations identified as providing general support are the most common form of LCSI support in England. These include councils for/of voluntary service, voluntary action bodies, and organisations that provide networking opportunities to local civil society organisations. Over time there has been a substantial increase in the proportion of LCSI support provided by specialist organisations: much of this support is aimed at young people and minority groups, with other organisations focusing on providing training in accounting and finance in particular (Figure 2).

Figure 2: Composition of LCSI sector, by year.

The income structure, measured using NCVO’s size bands, of the LCSI sector has changed over time, with a shrinking proportion of organisations in the lower income size bands (<£100k). Most organisations (c.80%) have annual income of between £100k and £10m - Figure 3 (Figures have been adjusted for inflation using 2023 GDP Deflator). There are no LCSI organisations in the top two income bands (>£10mn).

Figure 3: Income structure of LCSI sector, by year.

Note: Period begins in 1995 due to unavailability of financial information for earlier years.

The age structure of the LCSI sector has changed over time, with a shrinking proportion of organisations in the lower age bands (between 1-10 years old). This reflects two dynamics in particular: the sizeable number of LCSI organisations registered between 1997 and 2010; and the lower risk of dissolution for these organisations compared to those registered before 1997 (figure 4).

Figure 4: Age structure of LCSI sector, by year.

LCSI organisations can operate outside of the local authority they have a registered / head office in. Analysing the period for which this data is available (2012-2023):

• A little over 70% (518) of the 725 LCSI organisations active at some point across 2012-2023 reported operating in a local authority outside of where they have a registered / head office in.

• 70% of the 725 LCSI organisations report operating in a single local authority outside of where they have a registered / head office in, 10% in two areas, and 20% in three or more areas. Three LCSI organisations report operating in 12-15 areas.

LCSI Coverage

This section presents results of descriptive and statistical analyses of the geographical distribution of LCSI organisations in England over the period 1991-2023. The focus is on the extent to which LCSI organisations are present in local authorities, how coverage varies over time, and the local authority characteristics that explain why some areas are better served than others.

Present or not

Figure 5 shows the proportion of local authorities that have at least one LCSI organisation present – whether registered or operating in - and how this varies over time. There was slight but steady growth in the proportion of local authorities with LCSI organisations in the 1990s, followed by a flat trajectory until 2009, and subsequently a steady decline until 2023. Coverage in 2023 is the lowest across the observation period and 11 percentage points lower than the early 2000s peak. The decline in coverage has largely occurred in district councils, where the proportion has decreased from a peak of .92 in 2001 to a low of .67 in 2023 – coverage is largely consistent in other local authority types. The decline in coverage is likely a function of the dissolution of LCSI organisations that cover a single area i.e., those that are registered in a particular local authority and do not indicate they cover adjacent or other authorities.

Figure 5: Proportion of areas with at least one LCSI organisation, by year.

Local authority LCSI coverage is determined by the presence of one or more of the following LCSI types:

1. LCSI organisations registered in the authority.

2. LCSI organisations registered elsewhere but report operating in the authority.

Examining the period and areas for which both types of presence can be calculated, it is clear how important the second type of LCSI coverage is. Figure 6 shows that roughly 40% of all LCSI organisations present in an area are registered in a different local authority.

Figure 6: Mix of LCSI coverage, by year.

Note: Information on LCSI organisations registered elsewhere is first available from 2012.

The previous analyses show whether LCSI organisations are present at any point in a local authority over the period 1991-2023. Presence at particular time points can also be analysed. In general, most local authorities always have LCSI organisations present but there is a minority of areas that only have LCSIs from the late 1990s onwards. This analysis suggests there are three types of local authorities in regards to coverage:

1. A majority that has always had LCSI organisations present.
2. A minority (c. 40) that gained LCSI organisations over time.
3. A minority (c. 60) that had LCSI organisations initially but then lost them over time (and in some cases regained LCSI presence).

Note these patterns refer to LCSI organisations registered in local authorities. For the local authorities served by LCSI organisations registered elsewhere, the presence of those organisations is effectively constant over the period 2012-2023. Put another way, local authorities have a constant presence of LCSI organisations that are registered elsewhere.

Numbers of organisations

While the presence of LCSI organisations is common to all local authorities in England – at least at some point between 1991-2023 – there are inequalities in the number of these organisations. The typical number of LCSI organisations present in a local authority is two, with considerable variation around this: 75% of local authorities have three or fewer LCSI organisations, while 1% have eleven or more.

Figure 7 shows changes in the mean number of LCSI organisations present in a local authority, disaggregated by LCSI type. The number of organisations registered elsewhere but operating in a given local authority has remained largely steady since 2012, while for other LCSI organisations the number has declined slowly and consistently since 2006.

Figure 1: Number of LCSI organisations, by year.

Note: Information on LCSI organisations registered elsewhere is first available from 2012.

Correlates of coverage

The analysis now considers the area-level context in which LCSI coverage takes place.

Local civil society

LCSI coverage should relate in a logical way to the size and composition of the local civil society sector. In general, areas with larger numbers of civil society organisations tend to have more LCSI organisations present (figure 8). This pattern is weaker for LCSI organisations registered elsewhere, where the general upward trend tapers for areas with over 1000 civil society organisations. In summary, local authorities with larger numbers of civil society organisations have more LCSI infrastructure, but mainly in the form of more locally registered LCSI organisations.

Figure 8: Relationship between number of LCSI and civil society organisations.

A similar pattern is found for the relationship between the number of LCSI organisations and size of local civil society sector (measured as total income of all charities registered in the area) – see figure 9.

Figure 9: Relationship between number of LCSI organisations and size of local civil society sector.

There is no evidence that higher numbers of LCSI organisations registered in a local authority area are associated with higher foundation rates of civil society organisations (figure 10). There is a slight association between the number of LCSI organisations registered elsewhere and the foundation rate.

Figure 10: Relationship between number of LCSI organisations and foundation rate of civil society organisations.

Similarly, there is little-to-no association between the number of LCSI organisations in a local authority and the survival rate of civil society organisations (figure 11).

Figure 11: Relationship between number of LCSI organisations and survival rate of civil society organisations.

Area characteristics

LCSI organisations are found across local authorities of varying degrees of deprivation, but they are typically found in higher numbers in areas with highest deprivation, particularly LCSI organisations registered elsewhere (figure 12).

Figure 12: Relationship between number of LCSI organisations and local area deprivation.

Statistical modelling of coverage

The above analyses uncover area-level measures that are associated with higher or lower LCSI coverage. However, they are limited in that they represent summaries by one measure at a time (e.g., deprivation) and are calculated over all years in the data (e.g., the association over 1991-2023). Therefore, it is important to isolate the association between a specific area-level measure and LCSI coverage by controlling for all other relevant measures using a statistical model. The following characteristics of local authorities were used in the predictive component of the modelling:

• Index of Multiple Deprivation (2019) decile

• Region

• Urban-Rural classification

• Population size

• Expenditure on public services per capita (2019) decile

• Council type (i.e., London borough)

• Number of CSOs

• Year of observation, categorized as Westminster political periods (i.e., 1991-1996, 1997-2009)

Multilevel count regression models were used to identify characteristics associated with the number of LCSI organisations registered in a local authority. The multilevel structure arises from the 30 annual observation points (1991-2023) nested within each local authority. Local authorities were defined by their 2023 boundaries, and it was necessary to treat these boundaries as fixed over time, i.e. LCSI organisations were assigned to their 2023 boundary definitions for all years pooling across any previous local authority boundary definitions to create the 2023 definition. Figure 13 reports the results of the regression model. The IRR (incidence rate ratio) represents the change in the rate of occurrence of the outcome attributed to a given local authority characteristic:

• Values greater than 1 (dots to the right of dashed line) represent a higher rate of occurrence of the outcome (i.e., more LCSI organisations).

• Values less than 1 (dots to the left of the dashed line) represent a lower rate of occurrence of the outcome (i.e., fewer LCSI organisations).

The model shows the following relationships between local authority characteristics and number of LCSI organisations:

• LCSI organisations are more numerous in areas with higher numbers of civil society organisations.

• Relative to 1991-1996, the New Labour period of 1997-2009 was associated with higher numbers of LCSI organisations, while the post-Brexit referendum period was associated with fewer organisations.

• Relative to district councils, all other council types have fewer LCSI organisations. Note that this finding is net of the influence of other measures in the model, in particular local authority population size. This suggests that while district councils have fewer LCSI organisations in general, if you compare them to other council types with the same local authority characteristics they have higher numbers. In summary, district councils have lower levels of coverage as a result of factors other than council type.

• There are considerable regional disparities: relative to the South East, the East Midlands and North East have more LCSI organisations, while all other regions have lower numbers.

• There are also noticeable differences between rural and urban settings: relative to those classed as Urban with City and Town, Mainly Rural areas have higher numbers of LCSI organisations.

Figure 13: Count regression model results for number of LCSI organisations registered in area.

LCSI Sufficiency

The ability to define sufficiency in a conceptually clear and policy relevant manner is limited by the data available. Whether LCSI provision is determined as sufficient or not is dependent on the aims of infrastructure (support existing vs developing new civil society organisations), the efficiency of LCSI organisations’ operations (some with lower expenditure may be more effective or productive than those with greater financial resources), and the specific needs and mix of local civil society organisations. Therefore, to minimise the potential for drawing misleading conclusions, two simple measures of sufficiency have been operationalised. These focus on the spending capacity of LCSI organisations, as a proxy for capacity to deliver support relative to the size of the local population:

1. Typical (median) expenditure of LCSI organisations with a registered / head office address in the local authority, standardised by population size.

2. Total expenditure of all LCSI organisations with a registered / head office address in the local authority, standardised by population size.

For example, an area with 100k residents and a median LCSI expenditure of £50k would have a ratio of 50p per person. The actual values are less important than the inequalities in this ratio across local authority characteristics, as it is not literally the case that LCSI organisations are spending a certain amount on a given person (or every person). Analysis will particularly focus on showing these ratios for each local authority, as well as national summaries over time. All analyses are restricted to the period 2000-2022 as financial information is not available in sufficient quantity for other years.^{[footnote 1]} An additional caveat is expenditure is allocated to the area where a LCSI organisation has a registered / head office address. This likely undercounts expenditure for areas where there are some LCSI organisations present but registered elsewhere.

LCSI expenditure

There is little change over time in the level of LCSI expenditure per capita (figure 14). However, there is significant variation between local authorities in a given year. For example, in 2006 one local authority had total LCSI expenditure of almost £125 per head of population, while 25% of local authorities had £1.75 or less.

Figure 14: LCSI expenditure per capita, by year.

Figure 15 analyses changes in these expenditure measures for each local authority. Each line is the predicted trend in expenditure for a specific local authority. There is some variation in trends across local authorities:

• The general trend is flat or slightly increasing levels of expenditure over time.

• A number of local authorities experienced consistently increasing levels of expenditure (particularly total expenditure).

• Some local authorities experienced slow but steady declines in expenditure over time.

Figure 15: LCSI expenditure per capita trends over time.

Correlates of sufficiency

The analysis now considers the area-level context of LCSI sufficiency.

Local civil society

LCSI sufficiency should relate in a logical way to the size and composition of the local civil society sector. In general, areas with larger numbers of civil society organisations tend to have slightly higher levels of total LCSI expenditure per capita, though this is not the case for median expenditure (figure 16).

Figure 16: Relationship between LCSI expenditure and civil society organisations.

A similar pattern is found for the relationship between LCSI expenditure and size of local civil society sector (measured as total income of all charities registered in area) – see figure 17.

Figure 17: Relationship between LCSI expenditure and size of local civil society sector.

There is no evidence that higher LCSI expenditure is associated with higher foundation rates of civil society organisations (figure 18).

Figure 18: Relationship between LCSI expenditure and foundation rate of civil society organisations.

Similarly there is no association between LCSI expenditure and the survival rate of civil society organisations (figure 19).

Figure 19: Relationship between LCSI expenditure and survival rate of civil society organisations.

Area characteristics

LCSI total expenditure per capita tends to be higher in more deprived areas though this is not the case for median expenditure (figure 20). This likely reflects the fact that there are more LCSI organisations in deprived areas, which means the typical organisation can be the same size as in less deprived areas, but total expenditure is higher due to summing across more organisations.

Figure 20: Relationship between LCSI expenditure and local area deprivation.

Statistical modelling of sufficiency

The above analyses uncover area-level measures that are associated with higher or lower levels of LCSI expenditure per capita. However, they are limited in that they represent summaries by one measure at a time (e.g., deprivation) and are calculated over all years in the data (e.g., the association over 2000-2022). Therefore, it is important to isolate the association between a specific area-level measure and LCSI sufficiency by controlling for all other relevant measures using a statistical model. The following characteristics of local authorities were used in the predictive component of the modelling:

• Index of Multiple Deprivation (2019) decile

• Region

• Urban-Rural classification

• Expenditure on public services per capita (2019) decile

• Council type (i.e., London borough)

• Number of CSOs

• Year of observation in linear form^{[footnote 2]}

Multilevel linear regression models were used to identify characteristics associated with LCSI expenditure per capita in a local authority. Figure 21 reports the results of the regression model. The coefficient represents the change in the level of the outcome attributed to a given local authority characteristic:

• Values greater than 0 (dots to the right of dashed line) represent a predicted increase in the outcome (i.e., higher LCSI expenditure per capita).

• Values less than 0 (dots to the left of the dashed line) represent a predicted decrease in the outcome (i.e., lower LCSI expenditure per capita).

The model shows the following relationships between local authority characteristics and LCSI expenditure per capita:

• LCSI expenditure per capita is higher in areas with higher numbers of civil society organisations.

• There is no linear temporal trend in expenditure (as seen in figure 14).

• Relative to district councils, all other council types have lower levels of LCSI expenditure per capita. Note that this finding is net of the influence of other measures in the model. This suggests that while district councils have lower levels in general, if you compare them to other council types with the same local authority characteristics they have higher expenditure per capita. In summary, district councils have lower levels of expenditure as a result of factors other than council type.

• There are considerable regional disparities: relative to the South East, all regions have higher levels of LCSI expenditure per capita, in particular the East Midlands, North East, and Yorkshire and The Humber.

• There are also noticeable differences between rural and urban settings: relative to those classed as Urban with City and Town, all types of rural areas have lower levels of LCSI expenditure.

Figure 21: Multilevel linear regression model results for total LCSI expenditure per area.

Data Gaps and Recommendations

Data gaps

Lack of comparable data for community interest companies (CICs) – it was not possible, at scale, to identify LCSI organisations registered as CICs. This is due to:

1. limited reporting requirements of these organisations to their regulator Companies House; and

2. the unavailability of machine-readable annual reports.

As a result, it is difficult to measure the financial profile, even headline income and expenditure, of CICs. This matters as it prohibits analysis of whether there is sufficient LCSI provision relative to the composition of the local civil society sector. As the number of CICs grows, this data gap will worsen and in certain local authorities more than others due to the uneven distribution of this organisational form. The regulator (Companies House) has begun to release some sets of annual accounts for CICs but coverage is patchy and the data are difficult to work with from a technical and quality perspective.

Lack of area of operation data for district councils – a methodological strength of work package 4 was the incorporation of ‘Area of Operation’ data for LCSI organisations. This dataset, derived from annual snapshots of the Charity Commission for England and Wales’ data download, contains a self-reported list of the local authorities where an organisation operates in, and thus supplements analyses based purely on where an organisation is registered. In addition to the restricted time period of this data (2012-2023), organisations do not report which district council they operate in and therefore LCSI presence is likely underestimated. Organisations do report which county council they operate in but the research team are of the opinion that this would provide an inaccurate picture of LCSI presence i.e., it would lead to overestimates of coverage and underestimates of sufficiency.

Recommendations

Improve the coverage and quality of community interest company data – this may involve working with relevant organisations expert in the data (e.g., MyCake), and liaising with Companies House to understand how digitalised accounts are being rolled out. There is also a dearth of non-financial data. Therefore, it may be of value to consider developing a representative sample of these organisations - similar to the NCVO Almanac approach – and capturing more detailed information such as the organisation’s charitable / social purpose objects and activities. Implementing this recommendation would support more accurate estimates of LCSI provision as well as improve understanding of local civil society sectors.

Appendix: Methodology

Aims

The following three objectives were the focus of work package 4:

• How can we measure whether sufficient (in terms of coverage and quality) local civil society infrastructure is present in a local area?

• What are the characteristics associated with stronger and weaker LCSI presence in the area?

• What occurs to civil society in a local area following a change in LCSI provision?

Approach

List of LCSI organisations

Work package 4 commenced with a mapping exercise to determine the scope of entities to be defined as LCSI organisations. We started with the list of infrastructure organisations underpinning 360Giving’s Sector Infrastructure Funding Analysis report (Kane and Cohen, 2023).^{[footnote 3]} This included 1,782 local and national civil society infrastructure organisations registered as charities in England, Scotland or Wales. To align with the other work packages, only organisations classified as operating at a local level and based in England were retained. The 360Giving list had not been updated since 2021, therefore their methodology was applied to more recent extracts of the Charity Register for England and Wales to identify newly registered LCSI organisations (see previous footnote for process). Finally, a number of additional (existing or newly registered) LCSI organisations were identified through the membership lists of national infrastructure bodies such as NAVCA, Community Foundations UK, and Action with Communities in Rural England. Table 1 summarises the main steps in the construction of the final sample of CSI organisations.

A number of organisation types were excluded from the final sample:

1. Community Interest Companies (CICs): after discussion with David Kane, it was determined that the likelihood of a CIC being a significant CSI organisation was low. In addition, there is little easily accessible information on the activities / objects of CICs, which is needed to correctly identify them as CSI organisations. The exception would be CICs that are listed as members of NAVCA which are counted as valid CSI organisations.

2. Grantmakers / foundations: with the exception of members of the UK Community Foundations network, organisations whose only support for civil society organisations is in the form of grant funding were excluded. This is in line with the 360Giving approach though a list of these organisations is easy to obtain and include in the research if DCMS is interested.

3. Community facilities: the research team discussed the inclusion of village halls and community / youth centres as instances of CSI, but it was felt that this was not their primary purpose – they provide space for the local community in general, not CS organisations in particular. They are an ‘edge case’ whose inclusion could be warranted.

Classifying LCSI organisations

In order to analyse the composition of the LCSI sector, a measure of LCSI type was adopted from the 360Giving report:

1. Local – general

2. Local – service provider

3. Local – specialist

4. Local – volunteer centre

The research considered developing a bespoke typology based on keyword classification of LCSI organisation’s charitable objects but a) these would not be mutually exclusive and lead to the same organisation assigned multiple tags (e.g., youth-focused, financial support); b) would not be measured over time and therefore potentially present a misleading sense of stability of LCSI type.

List of local authorities

The base dataset containing LA level demographic data is the NOMIS mid-year population estimates for upper and lower-tier local authorities in England from 1991-2023. County councils were excluded from the analysis as information on LCSI organisations cannot be aggregated sensibly to this level and in a way that is comparable to other council types (e.g., district councils). The result is a sample of 296 local authorities observed 33 times over the period 1991-2023. In addition to population size, the following key contextual information was gathered for the local authorities from the following sources:

• Material deprivation – 2019 upper-tier and lower-tier local authority summaries

• Rurality – 2011 Rural Urban Classification (RUC)

• Local government expenditure – ONS local government expenditure 2019 dataset

• Region – November 2023 National Statistics Postcode Lookup

• Council type – UK Government list of councils in England by type

• Local civil society – Administrative registers of charities and community interest companies in England; measures include number of civil society organisations (CSOs), and income composition of local CSO sector

1. Headline income and expenditure is not available yet for all charities with a financial year end in 2023, while there are similar concerns with the coverage of financial information prior to 2000. ↩

2. Due to the restricted observation period (2000-2022), the Westminster political period operationalisation used in the statistical model of coverage was deemed invalid. ↩

3. Extract from the Register of Charities in England and Wales a list of charities meeting particular criteria, based on how they have classified themselves. The criteria are: a. “Activities” includes “Acts As An Umbrella Or Resource Body” or b. “Beneficiaries” includes “Other Charities Or Voluntary Bodies”. After this, a long list of charities was created, the list was manually checked, starting with the largest organisations to check whether they met the definition above. For smaller organisations, they were assumed to not be suitable for inclusion unless they contained a particular set of keywords, for example, “Council for Voluntary Service” or “Volunteer Centre” or included key phrases related to support for charitable organisations. ↩