Linking datasets: Charity Risk Data
Published:
Overview
This blog post outlines an example analysis carried out by linking organisations in our foundational spine to an important source of covariate data. Our project aims to link a diverse range of data sources to our organisational spine, including the ONS Business Index, 360Giving grant data, and trustee data held by the charity regulators. In this blog, we link our spine to the Serious Incident Reporting (SIR) scheme operated by the Office of the Scottish Charity Regulator (OSCR) between April 2016 and April 2024. This scheme offered charities a mechanism for demonstrating accountability in response to organisational incidents and crises; the intended impact of this scheme was to prevent and mitigate the effect of serious incidents on individual charities and the wider sector. In the sections that follow, we will provide some descriptive analyses on the type of events reported, who reports them, the type of charities affected, and a time series analysis to show the reporting trend over the period of our data.
Risk data
Beginning in April 2016, OSCR requested that charities report certain types of serious incidents through its SIR scheme; these were incidents, ongoing and historical, that threatened to have a significant impact on the charity or its assets. The SIR data we use covers the entire period of this scheme and includes 1185 reported events. These events relate to 760 distinct Scottish charities that we link to our organisational spine. Figure 1 below provides the first overview of this data and shows the most frequent type of events reported to OSCR.
Over 30 per cent of events relate to fraud and theft or financial issues/loss, with another 23.3 per cent related to safeguarding or personal behaviour. It was also possible for charities to identify Covid-19 as a threat to the organisation after the outbreak of the pandemic in March 2020. We can see that 5.7% of all events reported in this period were related to the pandemic. This data also allows us to see who reports these events to OSCR.1 In the vast majority of cases, this is either a trustee (52.6 per cent) or a staff member (41.5 per cent). The remaining six per cent of cases are reported by solicitors or independent examiners. Table 1 adds additional detail on how these events are distributed across categories of charitable activity using a version of the International Classification of Non-Profit Organisations scheme. This allows us to classify organisations into broader categories. Table 1 also provides data on how these categories are distributed across the Scottish charitable sector as whole to allow for comparison
| ICNPO category | % of SIR events | % of sector | |
|---|---|---|---|
| Social services | 19.7 | 10.6 | +9.1 |
| Civic, advocacy, political and international activities | 16.0 | 13.6 | +2.4 |
| Culture, communication and recreation activities | 14.8 | 17.8 | -3.0 |
| Religious congregations and associations | 11.1 | 15.6 | -4.5 |
| Community and economic development, and housing activities | 10.5 | 13.4 | -2.9 |
| Philanthropic intermediaries and voluntarism promotion | 9.5 | 14.5 | -5.0 |
| Education services | 8.0 | 8.1 | -0.1 |
| Human health services | 7.6 | 4.2 | +3.4 |
| Professional, scientific, accounting and administrative services | 1.4 | 1.2 | +0.2 |
| Business, professional and labour organisations | 1.0 | 0.4 | +0.6 |
| Environmental protection and animal welfare activities | 0.3 | 0.5 | -0.2 |
| Other activities | 0.2 | 0.2 | 0.0 |
Note: Figures for percentage of sector by ICNPO category is based on register of active charities in April 2023.
An interesting descriptive finding here is that charities classified as being in the ‘social services’ categories account for almost a fifth of all reported events, despite comprising just 10.6% of the sector. Below we add one additional layer of analysis. Figure 2 shows the time trend of reported events over the full period of the data. The numbers on the y-axis represent a count of reported events while the x-axis represents the period April 2016 to April 2024 in monthly intervals. Figure 2 and its associated statistical tests suggest that there is no significant time trend over the period of our data. In other words, the number of events reported per month does not seem to be increasing or decreasing significantly when we consider the full time period of our data. However, there is considerable and significant variation in this measure over time and we may find statistically significant changes in reporting rate by splitting our time period into sections of substantive interest.
Conclusion
In this blog post, we have linked an important source of covariate data (the OSCR SIR scheme) to our spine and have demonstrated how this data can be used to explore features of the Scottish charitable sector. By linking our spine to several additional data sources such as SIR schemes from other jurisdictions, the ONS Business Index, 360Giving grant data, and trustee data held by the charity regulators, we aim to produce a foundational database of UK third sector organisations that will allow users to build a holistic and integrated picture of the UK charitable sector. We hope this resource will make a significant contribution to the evidence base of the field and will be used to improve analysis of the third sector.
This data is missing for 162 events in the data so we base these descriptives on the 1,023 non-missing values. ↩