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

6. Distinguishing data purposes and needs: informing research, policy design, and coordination needs

Abstract

This chapter examines how data and evidence can support lifelong learning (LLL) systems by helping researchers, policymakers, and practitioners better understand learning needs, opportunities, and outcomes. It looks at how data can be used for three main purposes: improving knowledge about how people learn throughout life, informing the design and evaluation of policies, and helping to match learning opportunities with learners’ needs. The chapter highlights the strengths and limitations of current data sources and explores how new technologies and better data integration could provide more timely and useful information.

A central insight of the LLL framework is the distinct yet deeply interconnected roles that data must play to support its core logic. Coordination functions as the primary driver of the system, requiring data that are purpose-specific, integrated, and adaptive in order to align learning needs (demand) with available opportunities (supply) effectively across the entire life course. Building on the multidimensional data space and the synthesis of stakeholder objectives, interventions, and outcomes, this section delineates three essential functions that data serve within the framework:

• Research: Uncovering patterns, relationships, and causal mechanisms embedded at the nexus among learning occurrences, their associated demand, supply, and coordination, as well as their associated outcomes.

• Policy Design: Evaluating interventions and informing strategic choices to strengthen coordination pathways, transitions, and equity.

• Coordination: Facilitating both real-time alignment of demand and supply through feedback loops and mismatch detection, and system-level alignment linked to policy design through adaptive governance.

By differentiating research, policy design, and coordination purposes while advocating data enhancements and integration, the framework transforms data from a diagnostic afterthought into a proactive instrument for building adaptive, inclusive, and evidence-informed LLL ecosystems.

6.1. Research: uncovering patterns, relationships, insights and nuances

In large-scale comparative research, data help uncover both correlations and causal relationships, as well as recurring patterns such as the erosion of skills through disuse (Section 2.1). Large-scale, representative cross-sectional surveys, including the Adult Education Survey (AES) and the Programme for International Assessment of Adult Competencies (PIAAC), enable comparative analysis of LLL across countries and cohorts at the population level. Complementing these, longitudinal studies – such as the Longitudinal Surveys of Australian Youth, Canada’s Youth in Transition Survey and National Graduate Survey, National Educational Panel Study (NEPS) and PIAAC-L in Germany (Rammstedt et al., 2017[1]; Wicht et al., 2021[2]), the United Kingdom’s 1970 British Cohort Study (BCS70) and English Longitudinal Study of Ageing (ELSA), and in the United States the Education Longitudinal Study of 2002 and the High School and Beyond Longitudinal Study of 2022 – trace the evolution of educational careers and transitions over the life-course.

The principal advantage of comparable datasets such as PIAAC and AES lies in their provision of standardised metrics for skill proficiency and LLL practices. These offer reliable baselines for identifying trends in skill formation across diverse national contexts (Desjardins, 2020[3]; OECD, 2025[4]). However, such datasets are not designed to generate causal inferences. To address this, they must be linked with longitudinal data, whether administrative records, national panels, or experimental big data from private sources, thereby enabling more robust causal analysis and richer evidence on the effects of LLL policy measures (Field, 2011[5]; Grotlüschen et al., 2023[6]).

Together, these steps would strengthen the evidence base on causal relationships in LLL and provide a more reliable foundation for policy design. Advancing data infrastructure in this direction requires collaboration at both national and international levels.

6.2. Policy design: informing the policy process

For policy design, data provide the foundation for the entire process: informing analysis, deliberation, and debate with the best available evidence, while enabling both bottom-up formulation and top-down evaluation of interventions (OECD, 2023[7]). For instance, enhanced data on barriers such as time, cost, or access can support the design of equity-focused subsidies. Integrated strategies, such as linking responses from the AES survey with administrative records at aggregated geographical and sociodemographic levels, could improve the quality of evidence available for policymaking. Likewise, combining cross-sectional data from international surveys (e.g. PISA) with platform-based sources (e.g. LinkedIn, YouTube) at aggregated levels may generate more precise insights into learning needs and behaviours.

Future-oriented approaches are also emerging. AI-guided micro-credential pilots and simulation technologies, for example, offer policymakers tools to anticipate and test potential outcomes in controlled settings. AI-guided pilots involve small-scale, experimental programmes that use AI to design, deliver, or evaluate micro-credentials – short, targeted learning modules that certify specific skills or competencies (e.g. digital literacy, data analysis). AI can personalise learning pathways, recommend relevant modules based on individual or labour market needs, and analyse participant outcomes to assess effectiveness.

Similarly, simulation technologies that model the impacts of automation on labour markets and learning demands may soon become more widely available. Such simulations could replicate workplace scenarios where automation (e.g. AI-driven processes or robotics) displaces jobs in a particular economic sector or generates new skill requirements, enabling policymakers to observe how learners respond to training interventions in virtual environments.

Both AI-guided pilots and simulation tools create controlled environments in which hypotheses about LLL interventions can be tested, outcomes such as participation rates or skill gains can be observed, and policies can be adjusted iteratively.

6.3. Coordination: aligning needs and opportunities

Current coordination of LLL relies largely on lagged datasets: research-driven estimates of mismatches and policy analyses tailored to local needs. While effective to an extent, both research and policymaking are constrained by retrospective and static information, typically gathered through annual or periodic reporting cycles. Data sources are often fragmented or siloed, and further complicated by persistent privacy concerns.

Looking ahead, technology-driven systems hold the promise of more dynamic forms of alignment. Advances in AI and blockchain could, in principle, enable near-real-time coordination. Yet ethical and technical barriers mean that such a vision remains aspirational rather than immediately attainable.

In the shorter term, meaningful progress is possible within existing constraints. Stakeholders can strengthen cumulative records by integrating administrative data with consented, aggregated insights from other data sources that also cover less formal and organised types of learning. This would support both research, for example, by refining alignment models, and policy analysis, such as evaluations grounded in local needs. Aggregated linkages, such as combining regional administrative data with enrolment patterns from non-formal education providers and learning platforms, could improve coordination around concrete priorities (e.g. green reskilling initiatives or civic literacy programmes). Crucially, this approach connects survey or administrative data to aggregated local information without breaching anonymity through direct survey-platform linkages.

More ambitiously, a coherent data infrastructure, built around integration as a guiding principle, could pave the way for near-real-time coordination. Cumulative records might then combine administrative datasets with skills passports and aggregated big data, updated through secure platforms. For example, a blockchain-based skills passport could validate micro-credentials from platforms such as Coursera and link them to employment records, offering foresight on skill gaps for workers affected by automation. Such systems would enrich both research models and policy analyses through structured stakeholder dialogue.

AI could further expand these capacities. Machine-learning tools, for example, may link regional administrative data with anonymised platform trends to identify needs in areas such as civic or health literacy, thereby strengthening tripartite coordination and supporting more resilient ecosystems. Well-designed national pilots could provide a testing ground for such innovations, creating pathways towards a more adaptive and LLL infrastructure.

Examples such as Australia’s My Skills platform, which consolidates qualifications and workplace assessments into digital profiles, demonstrate scalable models (MySkills, 2025[8]). These could be extended into cross-national systems through OECD-led consortia. While achieving full real-time coordination will require overcoming ethical challenges (such as algorithmic bias) and investing in infrastructure, near-real-time systems may be within reach by developing standardised protocols and federated platforms.

References

[3] Desjardins, R. (2020), “PIAAC thematic review on adult learning”, OECD Publishing.

[5] Field, J. (2011), “Researching the benefits of learning: the persuasive power of longitudinal studies”, London Review of Education. Advance online publication, https://doi.org/10.1080/14748460.2011.61632.

[6] Grotlüschen, A. et al. (2023), “Adult learning and education within the framework of lifelong learning”, International Perspectives in Adult Education, No. 81, https://www.dvv-international.de/en/adult-education-and-development/international-perspectives-in-adult-education.

[8] MySkills (2025), MySkills Australia [Website], https://myskillsaustralia.edu.au/.

[4] OECD (2025), Trends in Adult Learning: New Data from the 2023 Survey of Adult Skills, Getting Skills Right, OECD Publishing, Paris, https://doi.org/10.1787/ec0624a6-en.

[7] OECD (2023), Flexible adult learning: What it is, why it matters and how to make it work, OECD Publishing.

[1] Rammstedt, B. et al. (2017), “The PIAAC Longitudinal Study in Germany: Rationale and design”, Large-scale Assessments in Education, Vol. 5, p. Article 4.

[2] Wicht, A. et al. (2021), “Low Literacy is not Set in Stone”, Zeitschrift Für Pädagogik 67. Beiheft 2021. Advance online publication, https://doi.org/10.3262/ZPB210110.