- Cross-Sector Systems Change: An analysis of 46 practical guides published between 2014 and 2025 across philanthropic, governmental, financial, and nonprofit sectors established a definition, an evaluation review, and a five-phase implementation model for systems change initiatives.
- AI Policy Frameworks: A systematic review of 229 social science articles found that public policy scholarship predominantly treats artificial intelligence as an administrative problem-solving mechanism or regulatory target rather than a socio-technical system embedded in power relations.
- Generative Agent-Based Modeling: The RePair framework operationalizes candidate mechanisms as natural-language rules within agent-based models, using matched interventions and behavioral traces across four simulation environments to measure how individual-level actions produce macro-level phenomena.
- Cyber-Systemic Human Resource Development: A book traces three decades of organizational learning research, using Stafford Beer's Viable System Model and Soft Systems Methodology to structure models for organizational self-regulation and adaptation.
- Power Dynamics in Housing Policy: Research combining systems mapping with power theories examined 38 interview-based maps of the English social housing sector, documenting how structural, instrumental, and discursive power shape municipal decisions regarding estate demolition versus refurbishment.
- K-12 Systems Pedagogy: A systematic review of ten empirical studies categorized five curriculum designs for integrating systems thinking into K-12 game creation, identifying System-Model-Game and Game-as-System inquiry orientations alongside collaborative, audience-oriented, and technical skill pathways.
- Engineering Problem Framing: A qualitative study of thirteen engineering students showed that systems mapping coursework shifted student approaches away from premature solution convergence toward structural and causal exploration of sociotechnical problems.
- Design for Transitions: Conceptual work analyzes the Multi-Level Perspective as an assemblage framework that coordinates theoretical knowledge across scales, actors, and contexts in systems-oriented transition design.
- Sustainability Archetypes: Researchers applied systems archetypes—including policy resistance, tragedy of the commons, and growth and underinvestment—to diagnose behavioral and organizational barriers within energy transitions.
- Sociotechnical Media Infrastructure: A conceptual model synthesizes platform studies, political economy, and infrastructure studies to evaluate digital journalism as a multilayered network shaped by platform dependencies and structural co-optation.
- Resilience Modeling in Science: A proposed framework applies tools from continuum damage mechanics and ecological resilience theory to model the scientific enterprise, using early-warning signals and agent-based simulations to monitor institutional erosion and potential tipping points.
- Mine Seismicity Precursors: A spatiotemporal dynamic prediction model combined an auto-reservoir neural network with Takens' embedding theorem to detect indicators of critical slowing down prior to seismic events in underground coal mines.
- Oscillator Network Synchronization: Permutation entropy quantifiers applied to spatial and temporal sequences signaled impending explosive synchronization transitions in Kuramoto oscillator networks across distinct topologies.
- Metropolitan Forest Resilience: Monitoring of vegetation index autocorrelation between 2004 and 2023 indicated that 60.54 percent of forests in the Wuhan Metropolitan Region exhibited declining ecological resilience, driven by patch shape complexity and potential evapotranspiration variation.
- Macroprudential Food Risk: A policy analysis evaluated transmission mechanisms linking food price shocks to inflation and bank non-performing loans, demonstrating that combining adaptive macroprudential instruments with food security programs curbs financial stability risks.
- Financial Contagion Dynamics: A dynamic epidemic framework integrated time-resolved transmission paths and Katz-Bonacich centrality into a multilayer SEIR network to model cascading risk spillovers and regulatory intervention thresholds across financial markets.
- Dementia Care Simulations: An agent-based model tracked 1,000 synthetic community dementia care units over 60 months across 100 iterations, quantifying longitudinal trends in caregiver burden, formal care transitions, and behavioral symptoms.
- Mental Health Policy Rollout: An evaluation of 31 interviews regarding the UK Future in Mind initiative mapped five feedback loops that explain local administrative adaptations, school responsiveness, and cross-system co-evolution between educational and health authorities.
- Additional Applications: Recent publications also include an interactive agent-based model of Milgram obedience experiments developed with the Mesa framework, an analysis of cultured beef adoption pathways through sociotechnical systems theory, and a published text examining critical thinking frameworks in literature, art, and film.
Sources
- Decoding a Decade of Systems Change Guidance, 2014–2025: Building a Shared Framework for Collaboration
ABSTRACT Systems change is gaining momentum across sectors as a powerful way to tackle today's most complex social and environmental challenges. Over the past decade, philanthropic, government, finance, development, nonprofit and research groups have published a wide range of practical guides, each tailored to specific areas. We review 46 frameworks published between 2014 and 2025 to identify patterns in how they define and implement systems change. We synthesize this knowledge to propose a shared definition and rationale, introduce a five‐phase framework to guide practice, unpack evaluation…
- Problem-solving ontologies on steroids? The space for critique in policy research on artificial intelligence
This article critically examines social sciences research on AI and policy as a politics of knowledge production, offering three key contributions. First, it debates the merit of systematic reviews in critical policy research in revealing epistemic hierarchies and biases. Second, it introduces a heuristic to map the disjointed methodologies of AI-policy scholarship across two ontological dimensions: whether AI features as a neutral problem-solving tool or a socio-technical system shaping power relations; and whether technology serves as a governance instrument or a regulatory target. This…
- Exploring Causal Mechanisms with Generative Agent-Based Models
In this paper, we explore using generative agent-based models for a classical ABM application: testing how individual-level behavioral rules produce collective phenomena. We introduce RePair, a method that calibrates simulation worlds, operationalizes candidate mechanisms as natural-language rules, estimates their effects through matched interventions, and examines behavioral traces. We assess the method by testing it in four simulation worlds grounded in established social-science models and empirical studies. Our results reveal that (1) natural-language rules can produce measurable…
- Realising the Learning Organisation through Systems Thinking
This book revisits one of the management theory's most enduring aspirations, the learning organisation, through three decades of research, practice, and reflection. Structured in three temporal layers (1993, 2013, and 2026), this book traces the evolution of cyber-systemic human resource development from its origins in Stafford Beer's Viable System Model and Soft Systems Methodologies, through empirical field research with leading German corporations, to a deepened contemporary understanding of how organisations can genuinely learn, self-regulate, and regenerate. This book offers a rigorous…
- Perceptions of Power in English Social Housing: A Systems Thinking Approach to Housing Politics
The erosion of affordable and secure housing in England prompts questions about power and its role in the “housing crisis”. While crucial for understanding housing politics, theory-grounded, actor-based power analyses remain rare. This study combines theories of power and systems thinking to examine perceptions of power in the English social housing sector. Using 38 interview-based systems maps, we trace instrumental, structural, and discursive power, and their material and social outcomes. Selected maps exemplify power dynamics in the decision to demolish or refurbish social housing estates.…
- Cultivating systems thinking through game-making: A systematic review on curriculum design
This systematic review examines how K–12 game-making curricula are designed in relation to systems thinking (ST). Prior research documents varied learning outcomes from game-making, but less is known about the curriculum designs through which ST is addressed. PRISMA 2020 was used to structure reporting, and ten empirical reports met the eligibility criteria. All ten contributed to the curriculum design synthesis, although the directness of their evidence in relation to ST varied. The analysis identified five pedagogical approaches. Two function as foundational orientations that define the…
- Learning to See the System: An Exploratory Qualitative Study of Systems Mapping Pedagogy and Sociotechnical Thinking in Undergraduate Engineering Students
Engineering curricula offer students few structured opportunities to practice reasoning about problems whose social, political, and economic dimensions are as consequential as their technical ones. This paper reports an exploratory qualitative study of an undergraduate engineering elective in which students investigate a wicked problem of their choice using multiple systems mapping tools, with an explicit emphasis on problem exploration over solution convergence. Drawing on semi-structured post-course interviews with 13 purposively sampled students across three cohorts, including a case-based…
- Multiplicity Through the Multi-Level Perspective in Design for Transitions
This short paper acts as a provocation for the role of the Multi-Level Perspective (MLP) in Design for Transitions. It explores Design for Transitions as a practice of assemblage, in which multiplicity (a key enabler of this complexity-led approach) is realised through a kind of “stickiness” that the MLP provides. The paper presents a number of ways in which multiplicity is engaged, with particular attention paid to the role of knowledge through theoretical multiplicity. This focus aims to demonstrate the wide range of theoretical knowledge that is frequently applied through, within, or…
- Uncovering the systems archetypes of sustainability transitions
Abstract Moving towards sustainable and resilient energy practices is a challenge with profound organisational and behavioural implications. While the importance of systems approaches in transition management is recognised, key principles of systems thinking such as understanding how complex systems behave, and our role within them, are often overlooked. This paper highlights interconnections, feedback mechanisms, and misperceptions shaping transitions at the intersection of business and societal risks. It uses Senge and Meadows’ work on systems archetypes, such as ‘policy resistance’,…
- JOURNALISM IN THE SHADOW OF INVISIBLE NETWORKS: SOCIOTECHNICAL SYSTEMS AND STRUCTURAL CO-OPTATION
This study aims to explore the theoretical and analytical potential of an infrastructure-oriented approach to journalism within the context of sociotechnical systems and to propose a framework for understanding the continuously evolving media environment. The study adopts an interdisciplinary qualitative conceptual research design that brings together the literatures on infrastructure studies, journalism studies, platform studies, and the political economy of media, analyzing these bodies of literature relationally across the micro, meso, and macro levels. Traditionally grounded in physical…
- Below the Surface: Creep, Corrosion, and Tipping Points in the Social Resilience of Science
By its most visible quantitative metrics: numbers of researchers, papers, journals, and institutions, science has never looked healthier. Yet a set of subsurface indicators points the other way: the disruptiveness of the average paper and patent is falling, progress slows in the largest fields, replication is rare and often unsuccessful. Incentive systems can select for poor methods even when no one is actually cheating. Lay society shows a curious mixture of admiration and lack of trust for scientists and their work. We argue that a healthy-looking surface can coexist with, and mask, an…
- Early warning of mine seismic events via spatiotemporal information transformation learning
Abstract Early warning of mine seismic events is critical for the safety of underground mining operations, yet remains challenging due to the high dimensionality and inherent noise of monitoring data, which obscure the latent dynamics preceding instability. Here, we propose a spatiotemporal dynamic prediction model (SDPM) that integrates an auto-reservoir neural network (ARNN) with spatiotemporal information (STI) transformation grounded in Takens’ embedding theorem. The framework converts simultaneous multi-sensor observations into low-dimensional temporal trajectories describing system…
- Anticipating explosive synchronization in networks of Kuramoto oscillators with network-based permutation entropy
Abstract Complex systems can experience gradual or sudden changes in their dynamics when a system parameter changes. For real world systems such as animal populations, vegetation fields, and climate systems, it is important to develop analysis tools able to infer that a change of behavior is approaching. Here we study networks of Kuramoto oscillators that experience a sudden transition to synchronization, referred to as explosive synchronization (ES), when the oscillators' coupling strength increases above a critical value. We analyze the performance of several quantifiers based on the…
- Climate-mediated associations between landscape fragmentation and metropolitan forest resilience in subtropical China
Forest ecosystems are increasingly exposed to climatic and anthropogenic disturbances under accelerating global change, making their resilience crucial for sustaining ecological functions. While resilience has become a key concept in forest management, its associations under combined pressures of climate stressors and fragmentation remain insufficiently understood, especially in metropolitan forested landscapes. We quantified and mapped forest resilience in the Core area of the Wuhan Metropolitan Region (CWMR), a rapidly urbanizing region in subtropical China, from 2004 to 2023, and…
- Macroprudential Policy in Mitigating Food Sector Shocks and Safeguarding Financial System Stability
Ideally, macroprudential policy is designed to operate in a forward-looking and anticipatory manner to safeguard financial system stability from various sources of risk, including sectoral shocks. In reality, however, policy design remains predominantly focused on internal credit cycles, while food sector shocks (food shocks) with potential systemic implications have not been fully integrated into the macro-financial surveillance framework. This study aims to analyze the role of macroprudential policy in anticipating the impact of food sector shocks on financial system stability through…
- Research on Risk Contagion Mechanism of Financial Markets Based on Complex Network Theory
Against deepening global financial integration, traditional linear risk tools cannot capture time-lagged cascading risk spillovers, while existing epidemic-network studies separately analyze multilayer topology, temporal transmission and node centrality without unified frameworks. This paper develops a temporal multilayer Susceptible-Exposed-Infected-Recovered (SEIR) coupled contagion model integrating time-resolved paths and statistically validated Katz-Bonacich centrality. It sorts three branches of contagion literature and identifies three key defects in prior static, fragmented models.…
- Evidence-based construction and validation of an agent-based model of community integrated care for older adults living with dementia
Objective Community dementia care involves interacting patients, caregivers, and formal care providers whose long-term dynamics are difficult to characterize with single-component empirical studies. We constructed and assessed a baseline agent-based model (ABM) to establish a transparent, inspectable reference scenario against which future computational experiments on organizational configurations can be evaluated. Methods A fully rule-based ABM of community dementia care was developed using prior empirical studies, published literature, policy documents, and structured expert consultation.…
- Applying a Complexity Lens to Policy Implementation: How Feedback Loops Help to Understand Systems Change in Integrated Healthcare
BACKGROUND: Although a complexity theory lens can help to understand national policy implementation at a local government level, application of this lens often remains at a metaphorical level. We illustrate how complexity concepts, such as adaptation, feedback, emergence and co-evolution can be used practically in research on policy implementation using a UK case study (Future in Mind; FiM), aimed at improving school children's access to mental health services in a metropolitan borough in the North England. METHODS: We re-analysed interview data with staff from local government, National…
Also this week
Full transcript
Decades of initiatives to transform institutions have produced dozens of separate guides without a common operating framework. Tracing the shared phases of that work is our focus on ComplexityPod, examining systems research across disciplines. Here is this week's survey.
Across forty-six practitioner guides published over the last decade, researchers identified a recurring five-phase pattern for cross-sector systemic work: initiation, system diagnosis, strategy formulation, execution, and continuous evaluation.
Yet the review also established persistent blind spots in how those phases get assessed, especially when shifting from theoretical roadmaps to real institutional settings.
That friction shows up clearly in public policy literature on artificial intelligence. Out of two hundred and twenty-nine social science publications, three-quarters restrict their scope to regulatory frameworks rather than observing administrative deployment.
They treat the algorithms as neutral optimization instruments or discrete risk calculations, sidestepping how automated tools function within existing political hierarchies.
Power dynamics dictate municipal outcomes in tangible ways. In the English social housing sector, researchers mapped thirty-eight decision networks regarding whether municipal estates undergo refurbishment or outright demolition.
The maps showed institutional responsibilities aligning directly with market discourse. Instrumental, structural, and discursive forms of power funnel local authorities toward demolition regardless of stated preservation targets.
Counter-pressures can emerge, though. An analysis of the United Kingdom's Future in Mind child mental health policy identified two reinforcing and three balancing feedback loops across thirty-one interviews.
Those specific balancing and reinforcing loops forced schools and healthcare bureaucracies to co-evolve, producing localized adaptations rather than top-down compliance.
When systems fail to adapt, classic archetypes explain the stalling patterns. Researchers mapped energy transition delays using Donella Meadows and Peter Senge's models—specifically policy resistance, growth and underinvestment, and the tragedy of the commons.
Qualitative archetypes counter the blind spots of purely technical grid models, showing why economic incentives alone stall decarbonization.
Testing those individual behavioral mechanisms at scale is shifting toward generative agent-based setups. The new RePair framework encodes behavioral hypotheses as natural-language instructions within simulation environments.
Because it runs matched counterfactual scenarios against baseline literature, you can isolate whether a macro-level social shift came from a specific micro-level rule change.
That direct link from micro-behaviors to macro-outcomes is evident in a five-year simulation of community dementia care tracking one thousand patient-caregiver dyads.
As simulated behavioral symptoms rose from seventy-five point six percent to ninety-six point four percent, the quantifiable strain on caregivers triggered a steep rise in transitions to institutional facilities.
Contagion mechanics apply to balance sheets as well. A multilayer epidemic model incorporating temporal network paths and Katz-Bonacich centrality mapped liquidity-endogenous recovery in interbank networks.
It tracks phase transitions directly, arguing that baseline microprudential rules fail unless regulators implement spectral monitoring of systemically central nodes.
A companion macro study linked that vulnerability to resource shocks, demonstrating that agricultural shortfalls generate sustained inflation and non-performing loan spikes.
Standard credit surveillance ignores supply shocks, which is why the authors pair adaptive capital buffers directly with domestic food reserves.
On network phase shifts, researchers applied permutation entropy to Kuramoto oscillator networks, tracking ordinal patterns across node sequences under increasing coupling strengths.
That gave an early warning signal for explosive synchronization before critical slowing down became visible.
Similar early-warning mathematics appeared in underground mining safety. Coupling auto-reservoir neural networks with Takens delay embedding theorem mapped multi-sensor streams into low-dimensional trajectories.
Which flagged critical slowing down before seismic failure occurred across three working coal basins.
Degradation dynamics also registered in ecological mapping across the Wuhan Metropolitan Region between 2004 and 2023. Satellite vegetation indices showed temporal autocorrelation increases across sixty point five four percent of metropolitan forests.
Structural equation modeling pinpointed patch shape fragmentation and evapotranspiration swings as the primary drivers of that declining stability.
Researchers extended those exact continuum damage models to academia itself, treating declining reproducibility and distorted publication incentives as structural fatigue.
They propose agent-based tracking to determine how close the research ecosystem sits to an irreversible threshold.
Systemic training at the student level shows mixed results. An assessment of thirteen engineering undergraduates using systems mapping observed shifts toward holistic problem framing, but students struggled with temporal feedback loops compared to structural diagrams.
A review of ten K-12 game design curricula revealed similar divisions, splitting between 'System-Model-Game' and 'Game-as-System' approaches supported by three specific facilitation models.
In management theory, a cyber-systemic framework synthesized Stafford Beer's Viable System Model and Soft Systems Methodology across three distinct benchmark years—1993, 2013, and projections for 2026.
Simultaneously, the Multi-Level Perspective was re-evaluated as an assemblage methodology, linking disparate actors across operational scales rather than acting as a static transition model.
Communication networks received a similar treatment, with digital journalism conceptualized as a three-tier infrastructure shaped by algorithmic mediation and institutional co-optation.
Also tracked this cycle: Maia Lamarque examined arts engagement within cognitive well-being models; researchers mapped socio-technical hurdles for cultured beef commercialization; and an interactive simulation reconstructed the Milgram experiments using the Mesa library.
We return next week with more research on complex systems. From ComplexityPod, thanks for listening.
From our community
At the Intersections — latest issue
Welcome to this issue of At the Intersections, where investigations connect across physical structures, human collaboration, and the wider cosmos. This week brings together research on the built environment, linking resource reuse in constr