"background": "Municipal infrastructure system reliability in developing nations is critical for urban resilience. Previous studies in this context have applied deterministic models, which may not adequately capture the inherent variability and hierarchical structure of asset performance data. ", "purpose and objectives": "This replication study aims to methodologically re-evaluate a prior analysis of infrastructure reliability by implementing a Bayesian hierarchical model. The objective is to assess whether this probabilistic framework provides more robust and nuanced inferences compared to the original approach. ", "methodology": "We replicated the data collection protocol for water supply and road networks across multiple municipalities. System reliability was modelled using a Bayesian hierarchical structure: y{ij \ (, ), (, ) = Xij\ + uj, with uj \ N (0, \²u), where i indexes assets and j municipalities. Posterior distributions were estimated using Markov chain Monte Carlo sampling. ", "findings": "The Bayesian model quantified substantial municipality heterogeneity, with the group-level variance parameter \ having a 95% credible interval of 0. 42, 0. 87 on the log-odds scale. This indicates that over 30% of the variation in reliability metrics is attributable to municipal-level factors not captured by asset-specific covariates, a finding obscured in the original deterministic analysis. ", "conclusion": "The replication confirms that hierarchical Bayesian modelling is a superior methodological choice for municipal infrastructure systems, as it formally accounts for data structure and provides probabilistic uncertainty quantification for decision-making. ", "recommendations": "Future infrastructure reliability assessments should adopt hierarchical probabilistic models. Municipal engineers should prioritise interventions informed by the estimated random effects to target underperforming administrative units. ", "key words": "Bayesian inference, infrastructure resilience, asset management, probabilistic modelling, urban services", "contribution statement": "This study provides a
Kavishe et al. (Tue,) studied this question.