Underrepresented transition factors from social sciences in energy and emissions modeling

Problem Statement
Energy system models and integrated assessment models (IAMs) are pivotal tools for simulating climate change mitigation and socio-economic welfare. However, these models face persistent criticism for relying predominantly on techno-economic considerations and assuming perfectly rational, centrally coordinated decision-makers. This approach fails to capture the "messy reality" of transitions, which are driven by heterogeneous political and economic actors with diverse interests. While previous reviews have examined specific drivers (e.g., psychological or socio-economic factors) within narrow sectors, there is a lack of comprehensive reviews that explore the full spectrum of micro-to-macro drivers across various social science disciplines. Consequently, the integration of social science evidence into energy and emissions modeling remains fragmented, limiting the realism and policy relevance of projections.

Methodology
The authors employed a theory-driven, machine-assisted literature review to identify and evaluate the representation of social science factors in energy modeling. The methodology proceeded in three main stages:

1. Theory Review: The authors reviewed 43 major social science theories (spanning psychology, political science, sociology, economics, and transition studies) to extract 24 distinct "transition factors." These factors were defined as cause-effect mechanisms explaining intentional behavior or unintentional outcomes. They were categorized into four types:

   • Cognitive-motivational: (e.g., emotions, bounded rationality, intentions).
   • Institutional-structural: (e.g., institutional quality, path-dependency, critical junctures).
   • Socio-normative: (e.g., social practices, trust, social conformity).
   • Strategic-rational: (e.g., expected costs/benefits, lobby power, geopolitical power).

2. Machine-Assisted Literature Screening: Using the supervised machine-learning tool ASReview, the authors screened two distinct literature bases:

   • Social Science Literature: 106,178 records related to energy transitions.
   • Modeling Literature: 4,892 records related to energy system and IAMs.
     For each of the 24 transition factors, the top 100 most relevant articles from each base were manually screened and coded for analytical scope (micro/meso/macro), main actor, sector, resolution, geography, and research type.

3. Gap Analysis and Integration Strategy: The authors derived a novel "representation gap index" by contrasting the frequency of each factor in the modeling literature against its frequency in the social science literature. They also categorized existing collaborative efforts into three integration strategies:

   • Merging: Fully endogenizing social factors into quantitative models.
   • Iterating: Joint analysis where social science informs scenario assumptions for models.
   • Bridging: Disciplinary exchange to interpret model outcomes without altering model structure.

Key Results
The analysis revealed significant disparities in how social science factors are treated in modeling:

• Representation Gap: A clear hierarchy of underrepresentation exists. The highest gaps (Index score ≥ 0.90) were found for Emotions, Social movement capacity, Intentions, Trust, Social resources, Lobby power, Social conformity, and Social tipping points. These factors are well-researched in social sciences but rarely integrated into models.
• Well-Represented Factors: Factors with low gap scores (≤ 0.45) include Expected costs and benefits (0.00), Political economy (0.02), Path-dependency (0.21), and Incremental learning (0.24). These are predominantly institutional-structural or strategic-rational factors that align with existing cost-optimization and linear modeling frameworks.
• Integration Strategies:
  • Merging (32% of cases) is dominated by strategic-rational and institutional-structural factors (e.g., cost optimization, learning rates). Cognitive-motivational and socio-normative factors are rarely endogenized.
  • Iterating (33% of cases) is frequently used for institutional-structural factors to test policy impacts or external shocks (e.g., geopolitical tension).
  • Bridging (20% of cases) is often applied to socio-normative factors to contextualize results (e.g., assessing justice or resilience post-modeling).
• Geographic and Sectoral Bias: Both literature bases show a strong bias toward high-income countries (Europe/North America) and national/global scales. Low-income economies and sub-national resolutions are significantly underrepresented, particularly in modeling literature.

Significance and Recommendations
The paper argues that improving the realism of energy transitions requires moving beyond techno-economic assumptions to incorporate empirically grounded social dynamics. The authors propose four specific recommendations for future research:

1. Prioritize Underrepresented Factors: Research should focus on high-gap factors with high relevance and feasibility, specifically Institutional quality, Lobby power, and Social movement capacity. These offer opportunities to model the interplay between institutional settings, vested interests, and policy implementation.
2. Match Strategy to Epistemology: Integration strategies should be chosen based on the theoretical tradition of the factor. While merging is ideal for factors with strong quantitative traditions (e.g., behavioral economics), iterating and bridging are essential for factors rooted in humanities or qualitative social sciences (e.g., social movement theory) where quantification is difficult.
3. Generate Quantitative Evidence: Social scientists should design studies to provide simple, causally accurate, and quantitative estimations (e.g., triangulating evidence) to facilitate merging strategies for factors like Lobby power and Social movement capacity.
4. Expand Geographic Scope: Future research must address the underrepresentation of low-income and industrializing countries to avoid biases in how transition factors (such as justice or cultural norms) are perceived and operationalized.

The paper concludes that a pluralistic approach to integration, acknowledging the limits of full quantification while leveraging diverse collaboration models, is necessary to build a more robust evidence base for low-carbon transitions.