No One Left Behind: evaluability assessment (national impact and economic evaluation)
This report presents the findings of an evaluability assessment for impact and economic evaluation of No One Left Behind at a national level. It considers feasibility, suitable methodological approaches and data and analytical requirements for future evaluations.
8. Annex 2: Long List
Below, we present the long list of data sources that were reviewed, along with the reasons why they were excluded from the refined short list for the Scotland-wide No One Left Behind evaluation:
Longitudinal Education Outcomes (LEO): LEO is a linked administrative dataset that tracks individuals’ education and labour market outcomes over time. For Scotland the dataset combines information from the Scottish Funding Council (SFC), the Higher Education Statistics Agency (HESA), Skills Development Scotland (SDS), the Department for Education (DfE), HM Revenue and Customs (HMRC), and the Department for Work and Pensions (DWP), enabling longitudinal analysis of post-16 education participation, employment, earnings, and benefit receipt, with a linkage to schools education data planned for 2027 to expand this datasets analytical potential.
LEO is an evolving dataset, and the extent to which it can be used to identify outcomes for No One Left Behind participants who go on to access education routes is unclear. While a preliminary investigation of LEO has been carried out, we recommend that the Scottish Government investigate further its feasibility for the evaluation of No One Left Behind.
Scottish Index of Multiple Deprivation (SIMD): SIMD is a composite index that brings together information from several key domains, including income, employment, health, education, crime, and housing. These domains are then combined to produce an overall ranking for each geographical area in Scotland, enabling comparisons of deprivation levels across the country. This was excluded from the short list as SIMD is an area-based measure rather than an individual-level dataset. It reflects the average characteristics of people living in an area rather than the circumstances or outcomes of specific individuals. As a result, it is not suitable for tracking individual-level changes over time among non-participants or for assessing changes in service users' outcomes.
Scottish Household Survey (SHS): a national survey conducted by the SG that involves over 10,000 households across Scotland. The survey covers a range of different topics, including housing conditions, housing costs, employment, income, finances, and transport. Although SHS contains relevant information on employment and income, it is primarily designed as a general social survey rather than a dedicated labour market dataset. As a result, its labour market variables are relatively limited compared to specialist labour market surveys. Additionally, the SHS sample size is smaller than that of APS, which is included in the refined short list.
Business Register and Employment Survey (BRES): BRES is an annual survey conducted by the ONS which collects information directly from businesses about the number of people they employ at each location. This enables the production of local-level employment statistics, including breakdowns by sector, employment type, and geography, such as Local Authorities and regions. Despite its high-quality data on aggregate employment statistics at the workplace level, BRES was excluded from the final shortlist because it is an organisation-based dataset rather than an individual-level dataset. This means that it does not track individuals over time and therefore cannot be linked to programme participation or individual outcomes. Secondly, since it focuses on jobs rather than individuals, it cannot distinguish among the different jobs held by the same person or capture individual employment histories. This makes it less suitable for evaluating participant-level outcomes, such as progression into sustained employment post-receiving No One Left Behind support. For these reasons, BRES was excluded from the shortlist despite being a useful source for understanding the structure of employment in Scotland.
Scottish Local Authorities’ Development (SLAED) Indicators Framework: The SLAED indicators framework is a national performance and benchmarking system used by Scotland’s 32 Local Authorities to measure and report on their economic development activity. It provides a basis for collecting consistent data on activities, outputs, and outcomes from Local Authorities. SLAED was excluded because it is essentially a reporting system that focuses on measuring overall activity rather than tracking outcomes for individuals who receive support. Most importantly, SLAED data is recorded and reported inconsistently across different areas, making national-level aggregation challenging.
Skills Development of Scotland Regional Skills Assessments (RSAs): RSAs are a set of reports that provide a detailed overview of labour market conditions and future skills needs across all major geographies in Scotland. It brings together a wide range of existing data sources and forecasting models to build an evidence base of regional labour markets to support future investment and planning. While RSAs are valuable for informing policy and investment decisions, they are primarily labour market intelligence tools rather than evaluation datasets. They are based on a combination of multiple data sources and forecasting models, and as a result, they provide a high-level overview of labour market conditions and future trends, but do not allow for the linkage of information at an individual level. Hence, they have been excluded.
Social Security Scotland: Social Security Scotland holds data on devolved social security benefits. The data includes information on submitted claims and benefit award details for disability benefits (e.g., adult disability payments), carer benefits (e.g., carer support payment), and other benefits (e.g., winter heating). Additionally, it includes variables such as age, gender, sex, religion, and disability. This dataset can be used as a mitigation to address data linkage concerns with RAPID. However, it is not included on the shortlist as it will not be used if linkage with RAPID is feasible. The other datasets reviewed are discussed in greater detail in the Data Assessment Findings section, as they were included in the refined shortlist for the impact evaluation.