A Conceptual Foundation for the Development of a Lifelong Learning Measurement Framework

7. Toward furthering data and indicator development

Abstract

This chapter examines how data and indicators can improve lifelong learning (LLL) systems by providing better information about learning needs, participation, and outcomes throughout life. It discusses the limitations of current data sources, which often provide only a partial picture of learning activities and overlook important forms of learning that take place outside schools and workplaces. The chapter proposes ways to improve data collection, including broader survey coverage, better use of administrative information, and stronger links between different sources of data. It also explores how digital tools and new technologies may help individuals, organisations, and policymakers better understand learning pathways and identify emerging needs.

The LLL framework’s three pillars – coordination, demand, and supply – inform consolidated data and indicator strategies, building on the conceptual integration of objectives, interventions, and outcomes and the distinct roles of data in research, policy design, and coordination.

As mentioned at the outset, the framework frames learning data points (i.e. concrete instances of learning occurrence at the intersection of demand and supply) as the foundational unit of measurement. These points, generated through both actual coordination (real-world resource flows and pragmatic alignments) and envisioned co-ordination (policy-driven system design), enable the tracking of proximal outcomes (e.g. skill acquisition) and distal outcomes (e.g. social mobility, civic resilience).

Current data systems, dominated by cross-sectional surveys such as the Programme for International Student Assessment (PISA), the Programme for the International Assessment of Adult Competencies (PIAAC) and the Adult Education Survey (AES), alongside siloed administrative records, produce retrospective, fragmented, and primarily economic insights. They under-represent informal and hybrid learning, non-work outcomes, and life stages outside the 25–64 age range. They also rarely distinguish between coordination as it actually occurs (real resource flows and pragmatic alignments) and coordination as it is envisaged in policy (system design), which limits their usefulness for dynamic policymaking. Specific weaknesses include poor tracking of educational biographies; under-representation of non-traditional learners; incomplete evidence on drivers and barriers to participation; limited coverage of non-formal and informal learning among young people; and weak linkages to welfare policy domains (Rubenson and Desjardins, 2009[1]; Boeren, 2016[2]; Desjardins, 2017[3]; Rubenson, 2018[4]; Boeren and Kalenda, 2025[5]).

To address these limitations, this section proposes experimental, integrative, and adaptive data development strategies that are aligned with the report's objectives:

• Cross-sectional survey development.

• Longitudinal survey developments.

• Administrative data development.

• Cumulative records: Digital profiles (e.g. skills passports) to track learning across life stages.

• Innovative methods: AI-guided pilots, participatory logging, and real-time indicators.

• Multi-source integration: Linking surveys (PIAAC, AES), administrative records, and big data (platform analytics).

Further, Annex C outlines a range of exploratory indicators tied to each pillar of the proposed conceptual framework. These are only exploratory indicators at this stage, meant to foster an experimental approach that could pilot the integration of new data to enhance information and indicators for policy purposes. The proposals aim to address gaps in LLL measurement and indicators, pending further validation and development.

7.1. Cross-sectional surveys development

To strengthen the LLL data ecosystem, existing survey instruments (e.g. AES, PIAAC, PISA, Continuing Vocational Training Survey – CVTS) should be upgraded, either through core revisions or targeted modules, building on the conceptual groundwork in Sections 2–4. Key areas for development:

1. Broaden age coverage: Extend all LLL cross-sectional surveys to include youth aged 16–18 and older adults (65+, ideally 70+), capturing the full range of learning activities (see Figure 4.2). This secures a genuinely life-wide perspective on early- and later-life learning.

2. Capture youth non-formal and informal learning: Develop PISA modules to measure non-formal and informal learning among 15-year-olds (e.g. participation in civic associations, digital platforms, and cultural activities) to reveal skill acquisition outside school and support a holistic view of youth learning.

3. Map today’s learning landscape: Introduce new operationalisations of learning activities that identify mixed modes of learning activities (formal with non-formal; non-formal with informal) (Kalenda and Boeren, 2025[6]). This is particularly relevant for semi-qualifications (e.g. micro-credentials), recognition of prior learning, and emerging modalities such as learning via AI applications and chatbots.

4. Measure digital delivery with nuance: Develop concise, sharper measures that move beyond the online/offline binary to capture synchronous vs asynchronous formats, hybrid provision, platform-mediated delivery, AI-supported tutoring, device context, and intensity of use.

5. Extend measurement of informal learning: Current cross-sectional surveys rely on few items, long recall periods, and a narrow focus on intentional informal learning, overlooking incidental and accidental forms of learning that are vital for skill formation (Eraut *, 2004[7]; Dochy et al., 2021[8]; OECD, 2025[9]). Adopt a new temporal framework centred on everyday learning in natural settings. Diary-based designs and mobile app data collection methods that track learning across the day offer promising routes (e.g. Kahneman et al. (2004[10]) and Killingsworth and Gilbert (2010[11])).

6. Strengthen employer-side coverage: As employers are principal LLL providers in advanced economies, large-scale measures should better capture workplace learning cultures, particularly the distinction between expansive and restrictive environments (Fuller and Unwin, 2004[12]). Expand the CVTS accordingly and add complementary measures for the public and non-profit sectors, as well as for workers in precarious employment, who remain underrepresented.

7. Refine participation drivers: Develop scales that integrate job-related, non-job-related, and skill-specific motives for engagement in LLL (e.g. DeSeCo competencies (2003[13])) Building on motivational scales, like Boshier’s Education Participation Scale (EPS) (Boshier, 1971[14]), distinguish between intrinsic and extrinsic motivations to inform responsive policy design.

8. Track support mechanisms: Extend cross-sectional surveys to measure financial support, information and guidance, and backing from unions and NGOs, linking these directly to welfare and employer policies. This will clarify how stakeholder support shapes access and participation.

9. Assess barriers comprehensively: Adopt a barrier framework that covers dispositional, institutional, and situational factors, with transversal constraints of finance, time, and information (Cross, 1981[15]; Kalenda, Vaculíková and Kočvarová, 2022[16]). Items should target both participants and non-participants to surface latent demand and guide equitable interventions.

10. Record educational biographies in depth: Enhance PIAAC and AES with modules on pathways to qualifications, including formal adult education at non-traditional ages. New items should capture flexible transitions, such as second-chance routes, to show how trajectories evolve over time.

These prospective enhancements directly address existing gaps and could be tested through OECD-led pilot surveys in five to ten countries, refining modules for coherence and scalability. Iterative feedback could drive global standardisation, fostering a data ecosystem that supports adaptive, inclusive, and evidence-informed LLL policies.

7.2. Longitudinal survey development

To develop longitudinal surveys, key research priorities are to:

1. Enhance existing longitudinal datasets by integrating additional LLL modules – particularly on learning needs, opportunities, and participation. These modules should align with the recommendations set out for cross-sectional surveys above to ensure conceptual and measurement coherence.

2. Synchronise national longitudinal surveys to enable tracking of long-term LLL dynamics across diverse institutional contexts, with harmonised timelines, core variables, and metadata standards to support cross-country comparability.

3. Link international cross-sectional surveys with national longitudinal and administrative data (e.g. labour-market outcomes), and with other relevant international surveys (for example, connecting PISA and PIAAC). Such linkages should use secure, privacy-preserving protocols and common identifiers to support robust, policy-relevant analysis over the life course.

7.3. Administrative data development in international contexts: negotiations, concepts, and comparative indicators

Administrative data, routinely collected by governments through sources such as enrolment records and training participation, are essential for constructing indicators of LLL. However, they remain challenging to harmonise for international research and policy learning, with comparability limited by divergent definitions, uneven quality, and incomplete coverage. Even so, there are notable examples of progress, such as the joint United Nations Educational, Scientific and Cultural Organization (UNESCO), OECD, and Eurostat database, which provides comparable indicators derived from administrative data and could serve as a foundation for further development. Advancing this field requires sustained expert negotiations, for instance, within OECD and UNESCO working groups, to clarify concepts, standardise indicators, and strengthen monitoring systems. Such efforts would ensure that data underpin comparative analysis and guide effective policy design.

A particular weakness lies in the treatment of Active Labour Market Policies (ALMPs), which encompass job-search assistance, training schemes, and subsidies. Currently, administrative datasets typically report only aggregated expenditure, obscuring key distinctions between relatively simple ALMPs, such as basic counselling with limited impact, and qualification-linked ALMPs, such as vocational training that leads to certified skills (for the exception see Albrecht, van den Berg and Vroman (2005[17])). This lack of granularity undermines robust evaluation of effectiveness. While simple ALMPs may generate short-term improvements, qualification-linked measures are more likely to support durable labour-market transitions and deliver wider social outcomes, including enhanced inclusion (Desjardins and Ioannidou, 2020[18]). This shift is crucial, as, according to Bonoli, Emmenegger and Felder‐Stindt (2025[19]), current welfare policies in many countries are moving towards more adult-education-based measures. Progress in this area depends on expert-led negotiations, such as those convened by the OECD’s Working Party on Employment, to define “qualification-linked ALMPs” as ISCED-aligned programs and to establish standardised indicators, such as “ALMP qualification completion rates.”

In practice, developing administrative data for LLL will require sustained international deliberation to harmonise concepts (e.g. measuring "training intensity" in terms of hours leading to certified outcomes) and to use shared platforms such as Eurostat’s Labour Market Policies (LMP) database as the basis for integrated datasets that combine ALMP records with survey data from sources such as AES or PIAAC. By bringing together statisticians, policymakers, and researchers through mechanisms like OECD working groups, these negotiations can foster comparable indicators, overcome governance fragmentation, and build more equitable and resilient LLL systems.

7.4. Cumulative records of qualifications and capabilities: managing transitions

Cumulative records in the form of ongoing digital profiles that integrate qualifications, skills, and capabilities represent a powerful data tool for tracking and managing LLL transitions along the life-course, such as career pivots or retirement. They allow individuals, organisations, and policymakers to monitor progress and align learning pathways with emerging needs. Unlike static credentials, these records can capture dynamic data across formal qualifications (e.g. degrees), non-formal certifications (e.g. micro-credentials), and informal capabilities (e.g. self-assessed or platform-verified skills).

Illustrative examples of the current implementations of these include:

1. European Skills Passports: The Europass portfolio combines CVs, language passports, and mobility documents into a comprehensive record that enables learners to present their qualifications and capabilities across borders. Recent pilot projects applying blockchain in education suggest new ways to secure, manage, and verify credentials (El Koshiry et al., 2023[20]). Malta, for example, has launched national initiatives to develop blockchain-based systems for educational and professional certification (Grech and Camilleri, 2017[21]), facilitating transitions such as job mobility for migrants.

2. Digital Learning Wallets: Singapore’s MySkillsFuture portal serves as a digital wallet, managing credits for approved training courses and allowing users to track competencies, such as digital literacy, throughout their learning journey. The platform offers personalised training recommendations through its Skills and Training Advisory tool, directly supporting mid-career reskilling. Data sharing between employers, training providers, and government agencies, underpinned by the Singapore Skills Taxonomy, ensures coordinated workforce development. By combining self-assessment tools with administrative records, the initiative enhances employability for mid-career and lower-wage workers, widening equitable access to skills development (SkillsFuture Singapore, 2025[22]).

3. National Capability Registries: Australia’s SkillsAware platform, launched in 2024 by SkillsIQ and Edalex, employs AI to aggregate competencies from formal education, non-formal training, and informal experiences into evidence-based skill profiles. Integrated with the National Skills Taxonomy, under development by Jobs and Skills Australia, it enables individuals and organisations to identify skills, close gaps, and plan career pathways. SkillsAware also contributes to a shared national language on skills, connecting education and employment systems, and fostering more equitable and resilient labour-market outcomes (SkillsAware, 2025[23]).

Nonetheless, challenges remain. Digital divides and data security concerns highlight the need for international agreements on standardised capability definitions and equitable access. Addressing these issues is essential to ensure that cumulative records contribute not only to employability but also to wider social outcomes such as resilience, inclusion, and well-being.

7.5. Data integration: combining survey, administrative, and big data sources

Current LLL data systems remain fragmented, relying predominantly on siloed sources such as large-scale surveys, administrative registries, and various small-scale qualitative studies of varied quality (e.g. Boeren (2017[24]; 2019[25]), Boeren and Kalenda (2025[5]); Elfert and Rubenson (2023[26])). This fragmentation makes it difficult to capture the fluid and multi-domain nature of learning across the life course. Data integration offers a promising approach, combining diverse sources into richer, more comprehensive datasets that enhance accuracy, coverage, and usability.

Each type of data brings distinctive strengths. Surveys can illuminate motivations and barriers; administrative records supply objective, longitudinal measures of participation and outcomes; while big data sources offer the capacity for real-time monitoring of informal learning. When integrated, these sources enable novel forms of analysis.

Illustrative examples include:

1. LLL Surveys and Administrative Data Combinations: Combining, e.g. AES or PIAAC measures with administrative labour‐registry records (e.g. unemployment benefits, earnings) enables rigorous impact evaluations of ALMPs on skills acquisition and employment outcomes. Following this approach in Switzerland, Denzler et al. (Denzler, Ruhose and Wolter, 2025[27]) recently linked microcensus data on adult education with labour-market registers and found that work-related education was associated with higher annual earnings and a lower risk of unemployment two and three years after treatment. Beyond analytical power, such integration reduces respondent burden through data reuse.

2. Big Data and Administrative Data Combinations: Combining administrative qualification records with big data from learning platforms (such as Coursera user logs) enables predictive analytics of capability development. In this regard, the European Skills, Competences, Qualifications and Occupations (ESCO) framework illustrates this potential, integrating job portal data with administrative employment records to identify emerging skills, such as AI literacy, and to refine real-time labour-market forecasts (Cedefop, 2024[28]). In practice, such data linkages could be implemented through API-driven dashboards for policymakers, allowing them to monitor platform-based learning and address gaps in formal qualifications.

3. LLL Surveys and Big Data Combinations: Merging AES survey data with big data from social media or app usage (e.g. YouTube engagement metrics) can shed new light on some of the patterns of informal learning. Such datasets, characteristic of big data, would capture information on respondents' demographic profiles alongside their engagement logs. Applying machine learning and predictive analytics to these combined sources could generate valuable insights into participation and outcomes, particularly when tracked over a longer period (two to four years).

In practice, integration may be implemented through secure, international systems, such as the EU’s European Blockchain Services Infrastructure (EBSI) blockchain platforms, where anonymised survey responses are enriched with administrative timestamps and big data metadata. Conversational AI agents can facilitate user-driven data input while ensuring privacy protection. At the international level, OECD and UNESCO working groups can negotiate shared standards (e.g. for measuring “informal learning engagement”), ensuring comparable indicators that promote agency and equitable outcomes across diverse contexts.

References

[17] Albrecht, J., G. van den Berg and S. Vroman (2005), “The Knowledge Lift: The Swedish Adult Education Program that Aimed to Eliminate Low Worker Skill Levels”, SSRN Electronic Journal, https://doi.org/10.2139/ssrn.673516.

[25] Boeren, E. (2019), Quantitative Research in Research on the Education and Learning of Adults, In A. Fejes & E. Nylander. (Eds.). Mapping out the Research Field of Adult Education and Learning (pp. 139–155). Cham: Springer.

[24] Boeren, E. (2017), “The Methodological Underdog: A Review of Quantitative Research in the Key Adult Education Journals”, Adult Education Quarterly, Vol. 68/1, pp. 63-79, https://doi.org/10.1177/0741713617739347.

[2] Boeren, E. (2016), Lifelong Learning Participation in a Changing Policy Context. An Interdisciplinary Theory, London: Palgrave Macmillan.

[5] Boeren, E. and J. Kalenda (2025), “Adult participation and large-scale survey data: conceptual and methodological debates”, International Journal of Lifelong Education, Vol. 44/3, pp. 249-256, https://doi.org/10.1080/02601370.2025.2491194.

[19] Bonoli, G., P. Emmenegger and A. Felder‐Stindt (2025), “Re‐Skilling in the Age of Skill Shortage: Adult Education Rather Than Active Labor Market Policy”, Regulation & Governance, Vol. 20/2, pp. 482-494, https://doi.org/10.1111/rego.70065.

[14] Boshier, R. (1971), “Motivational Orientations of Adult Education Participants: a Factor Analytic Exploration of Houle’s Typology”, Adult Education, Vol. 21/2, pp. 3-26, https://doi.org/10.1177/074171367102100201.

[28] Cedefop (2024), “European skills index 2024 technical report”, Publications Office of the European Union, https://www.cedefop.europa.eu/files/esi_technical_report_2024.pdf.

[15] Cross, K. (1981), Adults as learners: Increasing participation and facilitating learning, Jossey-Bass.

[27] Denzler, S., J. Ruhose and S. Wolter (2025), “Labour market effects of work-related continuous education in Switzerland – evidence from administrative data”, Economics of Education Review, Vol. 107, p. 102683, https://doi.org/10.1016/j.econedurev.2025.102683.

[13] DeSeCo (2003), Key Competencies for A Successful Life and, Hogrefe & Huber Publishers, https://www.deseco.ch/Key_Competencies_for_a_successful_life.pdf.

[3] Desjardins, R. (2017), “Political economy of adult learning systems: Comparative analysis of strategies, policies and constraints”, Bloomsbury Academic.

[18] Desjardins, R. and A. Ioannidou (2020), “The political economy of adult learning systems—some institutional features that promote adult learning participation, Die politische Ökonomie von Weiterbildungssystemen – einige institutionelle Merkmale, die Weiterbildungsbeteiligung fördern”, Zeitschrift für Weiterbildungsforschung, Vol. 43/2, pp. 143-168, https://doi.org/10.1007/s40955-020-00159-y.

[8] Dochy, F. et al. (2021), Theories of Workplace Learning in Changing Times, Routledge.

[20] El Koshiry, A. et al. (2023), “Unlocking the power of blockchain in education: An overview of innovations and outcomes”, Blockchain: Research and Applications, Vol. 4/4, p. 100165, https://doi.org/10.1016/j.bcra.2023.100165.

[26] Elfert, M. and K. Rubenson (2023), Lifelong learning: Researching a contested concept in the twenty-first century, In K. Evans, W. O. Lee, J. Markowitsch, & M. Zukas (Eds.), Third international handbook of lifelong learning (pp. 1219–1234). Springer Netherlands.

[7] Eraut *, M. (2004), “Informal learning in the workplace”, Studies in Continuing Education, Vol. 26/2, pp. 247-273, https://doi.org/10.1080/158037042000225245.

[12] Fuller, A. and L. Unwin (2004), Expansive learning environments: Integrating organizational and personal development, In H. Rainbird, A. Fuller, & A. Munro (Eds.), Workplace learning in context (pp. 126–144). Routledge.

[21] Grech, A. and A. Camilleri (2017), “Blockchain in Education”, EUR 28778 EN, Publications Office of the European Union, https://doi.org/10.2760/60649.

[10] Kahneman, D. et al. (2004), “A Survey Method for Characterizing Daily Life Experience: The Day Reconstruction Method”, Science, Vol. 306/5702, pp. 1776-1780, https://doi.org/10.1126/science.1103572.

[6] Kalenda, J. and E. Boeren (2025), “Revisiting the triadic classification of learning activities: Rethinking their measurement”, International Journal of Lifelong Education, Vol. 44/3, pp. 257–272.

[16] Kalenda, J., J. Vaculíková and I. Kočvarová (2022), “Barriers to the participation of low-educated workers in non-formal education”, Journal of Education and Work, Vol. 35/5, pp. 455-469, https://doi.org/10.1080/13639080.2022.2091118.

[11] Killingsworth, M. and D. Gilbert (2010), “A Wandering Mind Is an Unhappy Mind”, Science, Vol. 330/6006, pp. 932-932, https://doi.org/10.1126/science.1192439.

[9] OECD (2025), Rethinking Informal Learning. Assessment of the Latest Literature and Data, OECD Publishing.

[4] Rubenson, K. (2018), Conceptualising participation in adult learning and education: Equity issues, In M. Milana, S. Webb, J. Holford, R. Waller, & P. Jarvis (Eds.), The Palgrave international handbook on adult and lifelong education and learning (pp. 337–357). Palgrave Macmillan.

[1] Rubenson, K. and R. Desjardins (2009), “The Impact of Welfare State Regimes on Barriers to Participation in Adult Education”, Adult Education Quarterly, Vol. 59/3, pp. 187-207, https://doi.org/10.1177/0741713609331548.

[23] SkillsAware (2025), SkillsAware [Website], https://skillsaware.com/about-us/.

[22] SkillsFuture Singapore (2025), My Skills future [Website], https://www.myskillsfuture.gov.sg/content/portal/en/index.html.