Universities can systematically measure societal impact beyond teaching and research using a four-part framework—research transfer, teaching outreach, campus operations and enabling infrastructure—combining case studies, structured surveys and automated indicators; policymakers should fund institutional governance and guard against gaming of simple metrics.

Main Finding

The study develops a four-dimension typology and a practical assessment framework to measure higher-education institutions’ (HEIs) “third-mission” societal contributions (beyond research and teaching). The typology—research-based impact, teaching-based impact, campus-management-anchored impact, and administrative infrastructure—captures a broad set of impact pathways, including often-overlooked areas like campus operations and institutional support. The proposed mixed-method assessment approach (case studies, evidence-based questionnaires, automated multi-year metrics) balances rigor and feasibility and generates actionable recommendations for academic staff, institutional leaders, and national agencies while cautioning against overstating conclusions.

Key Points

• Typology: four dimensions
  • Research-based impact (knowledge transfer, commercialization, public engagement)
  • Teaching-based impact (community education, skills development, lifelong learning)
  • Campus-management-anchored impact (sustainability, local procurement, campus as living lab)
  • Administrative infrastructure (policies, support units, incentives that enable third-mission activity)
• Fills gaps in literature by highlighting campus operations and institutional support as legitimate impact channels.
• Assessment framework combines qualitative and quantitative methods:
  • Individual case studies for depth and narratives
  • Evidence-based questionnaires for comparability
  • Automated metrics for scalability and repeatability in multi-year cycles
• Identifies central measurement tensions to manage:
  • Breadth versus depth
  • General-purpose versus domain-specific metrics
  • Input (resources/activities) versus output/outcome indicators
  • Levels of analysis (individual, unit, institutional, regional)
  • Assessment burden versus outcome quality
• Practical orientation: provides a methodological toolkit and recommendations for integrating third-mission measures into evaluation and funding systems, with caution about interpretation limits.

Data & Methods

• Literature synthesis of existing third-mission/impact measurement work.
• Analysis of national and international best practices in HEI impact assessment.
• Framework design that operationalizes the typology into an assessment approach:
  • Mixed methods: qualitative case studies + structured questionnaires
  • Automated metrics: selection and deployment of measurable indicators tracked over multiple years
  • Iterative cycle: repeated measurement to capture temporal dynamics and allow benchmarking
• Emphasis on balancing methodological rigor with feasibility to reduce reporting burden and enable adoption by institutions and agencies.

Implications for AI Economics

• Measuring AI-related societal impact in HEIs:
  • The typology maps well to AI activities: research-based impact (AI research transfer, startups), teaching-based impact (AI curricula, workforce reskilling), campus-management (AI for campus efficiency, living-lab deployments), and administrative infrastructure (AI ethics boards, data governance).
  • A mixed-methods assessment helps capture both quantifiable outputs (patents, spinouts, training completions) and qualitative outcomes (community trust, ethical safeguards).
• Metrics design and incentives:
  • AI economics needs both general-purpose indicators (e.g., commercialization counts, employment effects) and domain-specific measures (fairness audits, downstream social harms/benefits); the study’s emphasis on trade-offs guides practical indicator selection.
  • Beware incentive distortions: automated, easy-to-measure AI metrics may be gamed or encourage narrow activities; balance with case studies and outcome-focused indicators to reduce perverse incentives.
• Data & automation opportunities:
  • Automated metrics and multi-year cycles align with AI economics’ capacity to leverage large administrative, bibliometric, and altmetric datasets for longitudinal analysis of AI impact.
  • AI tools can augment measurement (text analysis of case narratives, altmetric aggregation), but must be validated to avoid bias and misinterpretation.
• Policy and funding implications:
  • National agencies and funders can incorporate the framework to allocate support for AI research and education that demonstrably benefits society (e.g., funding tied to verified reskilling outcomes or community impact).
  • Institutional infrastructure (data governance, ethics oversight) should be explicitly funded and measured as part of AI governance—failure to do so risks undercounting crucial enabling activities.
• Research agendas for AI economics:
  • Empirically validate which indicators predict long-run societal outcomes of AI deployments (productivity, inequality, labor market displacement/creation).
  • Study distributional effects of HEI-led AI activities across regions and groups, using the multi-level approach the framework advocates.
  • Develop robust automated indicators for ethical and social outcomes (e.g., measurement of bias mitigation in deployed systems), combined with qualitative case evidence.
• Cautions:
  • Measurement cannot fully capture long-run or diffuse externalities of AI; avoid overinterpreting short-term automated indicators.
  • Standardization should allow for domain specificity—one-size-fits-all AI metrics will miss important contextual impacts.

Overall, this framework offers a practical, balanced approach HEIs and policymakers can adopt to measure AI-related societal contributions while highlighting methodological trade-offs and governance needs that are central to AI economics.