Civil Service People Survey: wellbeing and engagement items
Identity
Version: Annual survey; 2024 wave (fielded October 2024, results published 2025). Engagement Index is a fixed 5-item module; wellbeing module embeds the ONS4 personal wellbeing questions plus additional health/wellbeing items.
Structure: Engagement Index: 5 items. Wellbeing: ONS4 (4 single items: life satisfaction, worthwhile, happiness yesterday, anxiety yesterday) plus a small number of additional wellbeing/health items. Full instrument is roughly 60 to 70 questions across all themes.
Original citation: Engagement Index construction and question set documented in Cabinet Office technical/quality-and-methodology guides (csps_qmi2024). ONS4 wellbeing items originate from Dolan and Metcalfe 2012 (10.1017/s0047279411000833) and Hicks, Tinkler and Allin 2013 (10.1007/s11205-013-0384-x).
Steward / publisher: UK Cabinet Office (Government People Group), on behalf of participating civil service organisations.
Constructs claimed
Employee engagement (a composite 'Engagement Index' capturing pride, advocacy, attachment, inspiration and motivation) and employee wellbeing (subjective wellbeing via ONS4, plus perceived organisational health, stress and related states) among UK civil servants.
Evidence
Structural validity Very lowthinevidence form: canonical
indirect evidence is Cabinet Office grey literature on UK civil servants; no independent factor analysis located.
No independent peer-reviewed factor-analytic validation of the CSPS engagement index or wellbeing module was located in the latest review pass. The Engagement Index is constructed by Cabinet Office as an unweighted mean of five theme questions treated as a single dimension, described in the technical and quality-and-methodology guidance rather than derived from published confirmatory factor analysis (csps_qmi2024). The design draws conceptually on the broader employee-engagement literature synthesised by Bailey et al. 2015 (10.1111/ijmr.12077), but that synthesis concerns the construct generally, not this five-item index. As an item-set the record does not assert a single latent scale; structural evidence for the engagement index specifically is untested in the peer-reviewed record.
Convergent and discriminant validity Very lowthinevidence form: mixed
indirect ONS4 convergent evidence is general-population; index-level convergent evidence is grey-literature and internal.
No peer-reviewed convergent or discriminant validity study of the CSPS engagement index or wellbeing items was located in the latest review pass. Cabinet Office reporting relates the engagement index to other survey themes (for example leadership, my work, pay and benefits) descriptively, but these are internal cross-tabulations in grey literature, not validated convergent evidence (csps_qmi2024). The ONS4 items embedded in the wellbeing module have established convergent relationships with other wellbeing measures in the national-survey literature (dolan2012, hicks2013), but that evidence is general-population, not civil-service-specific.
Criterion validity: reference standard Absent (a finding about the literature)untestedevidence form: canonical
indirect no reference-standard study exists for this workplace instrument.
No validity evidence against any diagnostic or clinical health reference standard was located for the CSPS engagement index or wellbeing items. The instrument is not designed or validated as a health screener. The embedded anxiety-yesterday ONS4 item is a single affect rating, not a validated anxiety-disorder screen.
Criterion validity: organisational Lowthinevidence form: canonical
indirect associations are internal cross-sectional grey-literature findings on UK civil servants, not validated predictive studies.
This is the criterion for which the instrument was designed, yet published validation is thin. Cabinet Office and departmental reporting routinely relate engagement-index scores to organisational outcomes such as intention to leave, sickness-related themes and performance narratives, but these appear as descriptive cross-sectional associations in grey-literature reports (csps_qmi2024, csps_highlights2024) rather than as peer-reviewed predictive-validity studies with sickness absence, turnover or performance as measured outcomes. No independent longitudinal criterion study linking CSPS engagement or wellbeing scores to verified work outcomes was located in the latest review pass. The general proposition that engagement predicts work outcomes is supported in the wider literature (bailey2015), but not specifically validated for this index.
Internal consistency Absent (a finding about the literature)thinevidence form: canonical
indirect no reliability coefficient located in steward grey literature; general-population ONS4 items are single items.
No peer-reviewed internal-consistency coefficient (Cronbach alpha or omega) for the five-item CSPS Engagement Index was located in the latest review pass; the published technical and quality-and-methodology guidance describes index construction but was not found to report an alpha (csps_qmi2024). Because the record is treated as an item-set spanning heterogeneous single items (including the deliberately multidimensional ONS4 questions, which are not intended to form one scale), a single scale-level alpha is only meaningful for the engagement sub-index, and even there it is Absent in the located sources rather than a category error.
Test-retest reliability Absent (a finding about the literature)untested
indirect no retest study of CSPS items in a civil-service or any workplace sample located.
No test-retest study of the CSPS engagement index or wellbeing items was located in the latest review pass. The survey is fielded annually to (largely) different respondents, so year-on-year change reflects both real change and sampling, not measurement stability. Absence of retest evidence is itself the finding for this item-set.
Measurement invariance Absent (a finding about the literature)untestedevidence form: canonical
indirect cross-group comparisons are made in grey literature without invariance testing.
No published measurement-invariance analysis of the CSPS engagement index or wellbeing module across departments, grades, occupations or over time was located in the latest review pass. Cabinet Office benchmarking compares raw scores across organisations and demographic groups without establishing measurement invariance, so cross-group score differences cannot be assumed to be free of measurement bias (csps_qmi2024). This is a material gap given the instrument's primary use is cross-organisation and cross-group comparison.
Responsiveness and MIC Absent (a finding about the literature)untestedevidence form: canonical
indirect year-on-year movement is reported without a validated change threshold.
No responsiveness or minimal-important-change evidence for the CSPS engagement or wellbeing items was located in the latest review pass. Annual score movements are reported and interpreted by Cabinet Office, but no anchor-based or distribution-based minimal-important-change threshold has been established in the located sources (csps_highlights2024).
Populations, languages and norms
Extensive UK civil-service norms exist as an artefact of the census-style annual survey: several hundred thousand civil servants respond each year, giving very large organisational and demographic benchmarks published by Cabinet Office (csps_highlights2024, csps_qmi2024). These norms are highly representative of the UK civil service specifically, which is the intended population, but they are not general-UK-working-population norms and do not transfer to other sectors. The instrument is fielded in English.
Criticisms and controversies
Principal criticisms are that the CSPS engagement index is a bespoke administrative instrument whose psychometric properties have not been independently validated in the peer-reviewed literature, and that heavy reliance on Cabinet Office grey literature makes external scrutiny difficult. Cross-organisation league-table use of engagement scores proceeds without demonstrated measurement invariance. Response-rate variation across departments and years, and the shift to online administration, complicate longitudinal interpretation. The ONS4 wellbeing items, while nationally established, are single-item affect and evaluation measures with known ceiling effects and limited sensitivity, and were not developed for organisational benchmarking.
References (7)
- Benson T, Sladen J, Liles A, Potts HWW (2019). Personal Wellbeing Score (PWS) - a short version of ONS4: development and validation in social prescribing https://doi.org/10.1136/bmjoq-2018-000394
- Dolan P, Metcalfe R (2012). Measuring Subjective Wellbeing: Recommendations on Measures for use by National Governments https://doi.org/10.1017/s0047279411000833
- Hicks S, Tinkler L, Allin P (2013). Measuring Subjective Well-Being and its Potential Role in Policy: Perspectives from the UK Office for National Statistics https://doi.org/10.1007/s11205-013-0384-x
- OECD (2013). OECD Guidelines on Measuring Subjective Well-being https://doi.org/10.1787/9789264191655-en
- Bailey C, Madden A, Alfes K, Fletcher L (2015). The Meaning, Antecedents and Outcomes of Employee Engagement: A Narrative Synthesis https://doi.org/10.1111/ijmr.12077
- Cabinet Office (2025). Quality and Methodology Information for the Civil Service People Survey 2024 https://www.gov.uk/government/publications/civil-service-people-survey-2024-results/quality-and-methodology-information-for-the-civil-service-people-survey-2024
- Cabinet Office (2025). Civil Service People Survey 2024: Results Highlights https://www.gov.uk/government/publications/civil-service-people-survey-2024-results/civil-service-people-survey-2024-results-highlights
Record notes
Overall confidence: Very low to Low across most psychometric properties, driven by near-total reliance on Cabinet Office grey literature and an absence of independent peer-reviewed validation of the engagement index. This is a genuine, gradeable finding, not a search failure: the CSPS is an operational monitoring instrument, and its evidentiary strength lies in very large representative civil-service norms rather than in demonstrated measurement properties. The v0.2 item-set treatment fits well: the ONS4 items carry transferable (but general-population) evidence, while the engagement index is essentially untested in the peer-reviewed record. Licence verified against the current GOV.UK OGL footer, not founding papers.