ABSTRACT
Post-war cities with a legacy of socio-spatial rupture, ethnic fragmentation and chronic institutional distrust are in critical need of innovative urban planning tools. This study aims to test Gamified Spatial-Participatory Framework (GAM-SC) in four neighbourhoods of Kabul, with the objective of enhancing social cohesion and reducing spatial segregation. This study employed a rigorous sequential mixed-methods design, incorporating a survey of 450 respondents, spatial autocorrelation analysis (GIS), Partial Least Squares Structural Equation Modeling (PLS-SEM), and Agent-Based Modeling (ABM). Intervention implemented through purposeful deployment of game stations at ethnic boundaries of Dasht-e-Barchi (Hazara-dominated) and Karte-Sakhi (Tajik-dominated) neighbourhoods, with active mediation by local cultural brokers. Quantitative findings exhibited that the model significantly increased the cohesion index and reduced the fragmentation index. Furthermore, the model’s direct path on trust and the mediating role of digital participation were confirmed. Simulation also demonstrated higher resilience in intervened neighbourhoods with notable reduction in stability-recovery time. Qualitative analysis revealed four main mechanisms: the successful formation of third spaces (neutral public arenas) at ethnic boundaries; the active co-creation of shared narratives; inclusive participation via low-cost tools; and the perceptible enhancement of perceived justice through transparency. However, women’s participation was limited to 58 percent of men’s, underscoring the model’s effectiveness contingent on the socio-security context. This research, integrating three paradigms - social capital, spatial justice, and gamification - illustrates the strategic applicability potential of GAM-SC in fragile settings; yet it posits that its success hinges on the presence of trusted cultural brokers and continuous rule adaptability.

Extended Abstract
Introduction
Post-war cities are not merely damaged physical artifacts; they are profoundly traumatized socio-spatial organisms where enduring ruptures - territorial segregation, ethnic polarization, collapsed public trust, and civic bonds’ disintegration - lock communities into cycles of vulnerability. Conventional planning tools, predicated on stable institutional frameworks and baseline intergroup trust, often prove counterproductive because they ignore the psychological and spatial legacies of violence. Against this backdrop, this study develops and rigorously tests the Gamified Spatial-Participatory Cohesion (GAM-SC) framework - a transdisciplinary intervention fusing spatial design, culturally mediated participation, and low-cost gamification into a unified operational model. The empirical ground is provided by four multi-ethnic neighbourhoods in Kabul, Afghanistan a prototypical fragile city where over four decades of armed conflict have systematically deepened ethnic segregation, eroded intercommunal trust to near-zero levels, and engendered a profound suspicion of formal planning. In this setting, the dual aim is to measurably strengthen social cohesion and achieve demonstrable reductions in spatial segregation - outcomes that have remained stubbornly elusive in post-conflict urban interventions. A rich body of literature demonstrates that well-designed public spaces and inclusive participatory governance can catalyse social cohesion in stable, resource-secure cities. Yet this knowledge has not been translated into an operational framework addressing the simultaneous physical, social, and digital determinants of cohesion in post-conflict settings marked by deep intergroup mistrust, resource scarcity, and digital divides. The field thus confronts a critical vacuum: the absence of an empirically validated, integrative model that harnesses gamification’s motivational affordances, cultural mediators’ bridging capacities, and spatial-computational analytical precision. This study fills that gap, offering a theoretical advance and a practical, transferable tool.
The GAM-SC framework is underpinned by a central hypothesis: the impact of public spaces on social cohesion in Kabul’s ethnically fractured fabric is significantly moderated by the active involvement of cultural mediators trusted local figures who bridge divides, translate across identity boundaries, and facilitate safe interaction. It is proposed that gamified interventions are effective in mitigating spatial fragmentation only insofar as they actively mobilize and integrate such mediators into the participatory framework. In the absence of this mediating function, gamification runs the risk of being technologically compelling yet socially vacuous. This study addresses two research questions: 1. How does the GAM-SC framework increase social cohesion? 2. To what extent does it reduce spatial fragmentation? The study pursues three objectives. First, it articulates a theoretically grounded, context-sensitive model weaving together spatial configuration analysis, cultural mediation, and gamified public engagement. Second, it subjects the model to rigorous empirical validation through an advanced multi-phase mixed-methods protocol integrating GIS-based spatial analysis, Structural Equation Modeling (SEM) to test causal pathways, and Agent-Based Modeling (ABM) to simulate emergent socio-spatial dynamics under varying intervention scenarios. Third, it evaluates the framework’s measurable impact on pre-registered indicators: intergroup trust (adapted Putnam-scale with cultural-linguistic calibration and Cronbach’s alpha reliability testing), place attachment (five-point Likert scale), and spatial justice metrics (service distance index and spatial segregation index). The study makes a threefold contribution. It offers a pioneering integrative framework bridging spatial justice, social capital, and urban gamification within fragile post-war cities. It develops an advanced mixed-methods methodology uniquely suited to capturing complex, nonlinear socio-spatial dynamics in environments of extreme mistrust. It furnishes robust, granular empirical evidence from Kabul, a context critically underrepresented in planning scholarship despite its paradigmatic significance. The insights generated are intended to be directly implementable with local resources and transferable to other fractured urban landscapes grappling with rebuilding trust and reshaping divided space. The remainder of the article elaborates theoretical foundations, details the methodology, presents findings, and concludes with a discussion synthesizing theoretical and practical implications and directions for future research.
Methodology
A sequential mixed methods design was employed, grounded in a pragmatic epistemology that permitted the systematic integration of quantitative, spatial, and qualitative evidence within a single investigative framework. The quantitative strand began with a pre /post intervention survey of 450 residents drawn from four ethnically heterogeneous neighbourhoods of Kabul - Rikakhana, Dasht e Barchi, Karte Sakhi, and Qala e Zaman Khan - selected to capture the city’s full socio spatial gradient and ethnic diversity. The sample size was determined through an a priori power analysis (G*Power; Cohen’s f² = 0.15, α = 0.05, power = 0.80) and further inflated to account for a design effect of 1.5–2.0, ensuring robust statistical power for the subsequent Partial Least Squares Structural Equation Modeling (PLS SEM). The intervention itself consisted of a physical digital (phygital) platform deliberately designed to circumvent Kabul’s severe digital divide. Game stations were situated along the most salient ethnic frontiers, most critically between the predominantly Hazara Dasht e Barchi and the Tajik Karte Sakhi, on neutral ground previously identified through space syntax analysis. Each station integrated large format paper based interactive maps, QR codes affixed to physical markers, and tablet devices operated exclusively by trained local facilitators, thereby enabling participation without personal internet access.
The core activity was a weekly team challenge, ‘Hunt for Shared Signs,’ in which mixed ethnicity groups located objects of shared cultural significance such as - tandoor ovens and historic fountains - with bonus points awarded only to ethnically mixed teams and difficulty levels escalating gradually to sustain engagement. A publicly displayed scoreboard provided procedurally transparent rewards, reinforcing perceived fairness. Crucially, the entire process was mediated by locally trusted cultural brokers, including elders and religious figures, who modelled collaborative behaviour and reduced the initial distrust that typically impedes cross group contact in fragile settings.
All spatial analyses were conducted in ArcGIS Pro and GeoDa using base layers extracted from OpenStreetMap and recent satellite imagery, verified through extensive ground truthing. Spatial autocorrelation was assessed via Global and Local Moran’s I with 999 random permutations and a K nearest neighbours weight matrix (k = 8), while geographically weighted regression (GWR) examined how the relationship between access to improved public spaces and trust varied across the urban fabric. A composite Social Cohesion Index (SCI) was constructed from four sub scales between group trust, place attachment, civic participation, and ethnic cohesion each measured on a five point Likert scale through a 27 item researcher developed questionnaire that had been validated by an eight expert panel (Content Validity Index = 0.91) and demonstrated high internal consistency (Cronbach’s α = 0.89). The Spatial Fragmentation Index (SFI) was derived from established segregation and isolation metrics computed in GIS, with lower values indicating less segregated space.
The causal architecture of the GAM SC model was tested with PLS SEM in SmartPLS 4, with all constructs modelled as reflective and significance levels obtained through bootstrapping with 5,000 resamples. Model fit was evaluated using SRMR, NFI, and RMS_theta, while predictive relevance was confirmed through Stone Geisser’s Q² obtained via blindfolding. To project long term systemic behaviour, an agent based model (ABM) was constructed in NetLogo 6.3.0, populated with 1,000 heterogeneous agents calibrated on baseline trust levels, ethnic identities, and spatial mobility patterns from the survey and GIS data. Agents navigated the actual street network, and their interaction probabilities were governed by rules mirroring the game’s incentive structure. After a burn in period, an exogenous shock was imposed a sudden 40% loss in generalised trust across all agents and the system’s resilience was measured as both the time to return to the pre shock steady state and the maximum amplitude of the trust loss, with 100 independent simulation runs performed per condition. The qualitative strand comprised 48 semi structured interviews and four focus groups, purposively sampled for maximum variation in ethnicity, gender, age, and socio economic status. All sessions were conducted in Dari or Pashto, audio recorded with informed consent, transcribed verbatim, and translated.
Systematic grounded theory procedures (Strauss & Corbin, 1998) guided the analysis in MAXQDA 2022, with open, axial, and selective coding proceeding iteratively until thematic saturation was reached after 41 interviews; seven additional interviews confirmed the stability of the findings. An independent second coder analysed 20% of the transcripts, yielding a Cohen’s κ of 0.87, indicating excellent inter coder reliability. The qualitative findings were subsequently integrated with the quantitative and spatial results within a joint display matrix, enabling the extraction of meta inferences that explained the mechanisms underlying the observed changes in the SCI, SFI, and ABM dynamics and strengthening the internal validity of the overall GAM SC framework.

Results and Discussion
The intervention produced statistically significant and practically meaningful shifts. The SCI rose from 58.3 to 74.0 (∆ +15.7 units, +26.9%, p < 0.001), and the SFI decreased from 0.68 to 0.52 (∆ –0.16 units, –23.5%, p = 0.008). Between group trust indicated the largest relative gain (+61.9%, from 2.1 to 3.4 on a 5 point Likert scale), while place attachment grew by 39.3% (2.8 → 3.9). Global Moran’s I for the SCI fell from 0.42 to 0.21 a 50% reduction indicating a significantly more even distribution of cohesion across the urban fabric. PLS SEM confirmed a strong direct effect of GAM SC on between group trust (β = 0.428, p < 0.001) and a significant indirect effect mediated by digital participation (β = 0.152, p = 0.011). The overall model explained 62% of the variance in social cohesion (R² = 0.62). ABM simulations revealed that treated neighbourhoods returned to pre shock stability 43.3% faster than control neighbourhoods (Cohen’s d = 1.85) and sustained higher inter group cooperation throughout the recovery. Qualitative analysis uncovered underlying generative mechanisms shaping these patterns: (1) the creation of third spaces neutral, co managed public arenas at ethnic boundaries reduced perceived threat and enabled positive informal contact; (2) gamified team based activities (e.g., ‘shared heritage hunts’ with collective rewards) re scripted collective memory, forging a superordinate neighbourhood identity that partially displaced exclusionary ethnic categories; (3) low cost phygital tools lowered participation barriers for women, the elderly, and low income groups, fostering inclusive planning; and (4) transparent public display of project progress and resource allocation heightened perceived procedural justice, mitigating feelings of spatial discrimination. Critically, women’s participation reached only 58% of men’s, and the model’s effectiveness was attenuated in peripheral neighbourhoods where acute security constraints restricted full implementation. These findings resolve a persistent paradox: public space can either foster cohesion or intensify conflict depending on its configuration specifically neutral design and game rules rewarding mixed ethnic collaboration and on the active presence of trusted cultural mediators. The moderating role of cultural brokers (f² = 0.15) and the three reinforcing feedback loops cooperation based trust, progress based belonging, and transparency based justice explain how gamified spatial interventions convert latent social capital into resilient inter group cooperation, operationalising Allport’s contact hypothesis and the common ingroup identity model in a context devoid of formal institutional trust.

Conclusion
This study supplies robust, multi layered evidence that the GAM SC framework can substantially enhance social cohesion and reduce spatial fragmentation in post war, multi ethnic neighbourhoods. By integrating social capital, spatial justice, and gamification paradigms, the model offers a replicable, low cost strategy aligned with SDG 11 targets on participatory planning and inclusive public spaces. Nevertheless, generalisability is conditional upon three factors: the availability of trusted cultural intermediaries, the feasibility of establishing physically neutral third spaces, and the continuous, context sensitive adaptation of game rules. The limitations security driven under implementation in certain peripheries, constrained female participation, and a short follow up period call for longitudinal, comparative trials in similarly fragile cities (e.g., Mogadishu, Baghdad). In policy terms, GAM SC demonstrates that, even in deeply fractured urban settings, strategically designed playful participation can repair socio spatial bonds - provided that it remains anchored in local social fabric and committed to iterative, ethically vigilant refinement.

Funding
This research received no external funding.

Authors’ Contribution
Aimal Formolly conceived and designed the study, collected and analysed all data, conducted the spatial and computational modelling, interpreted the findings, and wrote the entire manuscript, including the Introduction, Theoretical Framework, Methodology, Results, Discussion, and Conclusion. Hojatollah Rahimi supervised the research process, critically reviewed the draft and final manuscript, and provided substantive revisions and editorial refinements.

Conflict of Interest
The authors explicitly state that there is no conflict of interest regarding the authorship and publication of this article.

Acknowledgments
We are grateful to all the scientific consultants of this paper.