Scottish Mentoring and Leadership Programme: impact and value for money evaluation
Findings of the impact and value for money evaluation of the Scottish Mentoring and Leadership Programme (SMLP), looking at the MCR Pathways element of the programme for care experienced young people.
Footnotes
1 Based on all care experienced pupils including those who have participated in the MCR Pathways programme, who account for around 6% of all care experienced pupils.
2 This is discussed further in Chapter 2.
3 Columba 1400 targets young people from disadvantaged backgrounds or facing ˜tough realities, identified by schools and support teams for their potential to benefit from the programme. Similarly, MCR Pathways Group 2 targets young people facing ˜tough realities.
4 Around 0.3% to 0.5% of records per year are duplicate pupil records, mostly due to pupils moving school. In these cases, we use the school in the later record to attach school-level data.
5 Biennial except for period 2008/2009 to 2010/2011 where annual data was available. Data from the 2020/2021 wave was not used to due discontinuity because of the covid pandemic.
6 Biennial except for period 2008/2009 to 2010/2011 where annual data was available.
7 Data on 2020/2021 attendance was not used due to the impact of Covid.
8 For Approaches A, B1 and B2, analysis was based on a comparison of point estimates of the control and treatment groups. Sample sizes are reported separately for these groups. For Approaches C and D, regression models are used and the total number of cases used by the regression model is reported.
9 Based on full attendance being 190 days, an increase in attendance by 4.4pp and a population size of 3,452 care experienced pupils in schools offering the MCR Pathways programme.
10 For all approaches, subgroup categories may not sum due to rounding. Additionally, the effect sizes for Approaches C and D are calculated from regression coefficients of the effects rather than as the difference between levels for control and participant groups. This means that they are estimated with errors associated with each coefficient, and therefore effects over multiple categories will not necessarily sum for these approaches.
11 Supporting tables are available alongside this publication in Excel format. These also show the point estimates for the treatment and control groups for Approaches A, B1 and B2 and the point estimates of outcomes among the population of non-care experienced and care experienced pupils for context. The supporting tables are available here: supporting tables.
12 The following destinations are considered ˜positive: Higher Education; Further Education; Employment; Training Voluntary Work; Personal Skill Development and Activity Agreements. Unemployed and Unknown are not considered ˜positive.
13 Section 5.4 of HM Treasury Green Book notes that social discounting should not be applied retrospectively for costs and benefits in the past for appraisal
14 Link for version of GDP deflator used is here: GDP deflators at market prices, and money GDP September 2025 (Quarterly National Accounts) - GOV.UK
15 MCR School Coordinator Job adverts (Accessed 12/05/2025): Clackmannanshire 2022, East Dunbartonshire 2024, Aberdeenshire 2023).
16 GCSE Attainment and Lifetime Earnings, Department for Education (2021)
17 The DfE tables show sizeable variation by subject within the same grade step. In general, STEM subjects have higher returns per grade than many arts/creative subjects. For example, the returns from a grade improvement in maths is £16,404 for males and £12,637 for females while returns from a grade improvement in music are £4,820 for males and £6,052 for females. Subject choice is therefore important which is why a subject‘weighted average is used.
18 Further details on calculation of unit values can be found in Appendix D
19 What Works Wellbeing (2017) Unemployment, (Re)employment and Wellbeing
20 HM Treasury (2021) Wellbeing Guidance for Appraisal
21The costs to the Exchequer of an unemployed person from Universal Credit were also considered when deciding the correct values to use; however, following the HM Treasury Green Book guidance, because transfers like Universal Credit pass purchasing power from one person to another and do not involve the consumption of resources they should not be included in Social Cost Benefit Analysis.
23 English Partnerships (2008) Additionality Guide
25 Section 5.4 of HM Treasury Green Book notes that social discounting should not be applied retrospectively for costs and benefits in the past for appraisal
26 PWC (2016) Economic Returns to Education in the United Kingdom
27 Based on individuals born in 1993 to 1994 and educated in England between 2003 and 2017. Source Office for National Statistics (2022) The education background of looked-after children who interact with the criminal justice system
28 Home Office (2018) The economic and social costs of crime: Second Edition
29 MCR Pathways provide first hand testament from beneficiaries of the programme. See example from Lucy: From homelessness to medical school
30 Crisis (2016) Better than Cure?
31 Centre for excellence for Childrens Care and Protection (2019) Homelessness and care experience
32 Department for Education (2025) Record attendance gains unlock over £2bn in future earnings
33 Brookes, M. Goodall, E. Heady.L (2007) Misspent youth: The costs of truancy and exclusions
35 There were a small number of minor changes to these series of data over time. Where this was the case, data was amended to allow measures to be consistent over time.
36 Intention-to-Treat is a principle in which participants in an intervention are treated within their originally assigned groups, regardless of adherence, withdrawals or deviations. In this context, this estimator treats all care experience pupils in MCR participating schools as treated.
37 As with Approach A, care experience was defined using Child Looked After Statistics (CLAS) data, while details of participation were provided by MCR.
38 This differs from the approach described for Analysis Approach B. In Approach B, once a pupil in the control group was matched to a treated pupil they were removed from the pool and could not be selected again. For Approach C with replacement, once a control school is matched, they can be matched again if they are also the most similar school to another treated school. This helps reduce bias where small samples are involved but does mean that it is necessary to apply weights to any post estimation of effects.
39 For further detail see STATA manual.
42 Supplementary tables referred to in this appendix are available alongside this publication in Excel format. The supporting tables are available here: supporting tables.