Abstract
{ "background": "Industrial machinery fleet reliability is critical for infrastructure development and economic productivity. Previous studies on system reliability in similar contexts have employed standard regression techniques, which may not adequately account for the hierarchical structure of fleet data, where individual machines are nested within depots and companies.", "purpose and objectives": "This study aims to replicate and methodologically evaluate a prior analysis of industrial machinery fleet system reliability. The primary objective is to assess the suitability of a multilevel modelling approach for this context and to verify the robustness of earlier findings concerning key reliability predictors.", "methodology": "A replication study was conducted using operational and maintenance data from a fleet of heavy earth-moving equipment. The core methodological evaluation involved fitting a multilevel linear regression model, specified as $y{ij} = \\beta{0} + \\beta{1}x{1ij} + u{j} + e{ij}$, where $y{ij}$ is the reliability metric for machine $i$ in depot $j$, $x{1ij}$ is a machine-level predictor, $u{j}$ is the random intercept for depot $j$, and $e{ij}$ is the residual error. Robust standard errors were calculated to account for potential heteroscedasticity.", "findings": "The multilevel model revealed significant random intercept variance at the depot level (σ² = 0.18, 95% CI [0.09, 0.31]), indicating substantial clustering. A key concrete result is that for every 100-hour increase in scheduled maintenance adherence, machine availability increased by approximately 7.2 percentage points (β = 0.072, p < 0.01). The original study's core finding on maintenance was corroborated, but the effect size was attenuated.", "conclusion": "The application of multilevel regression is methodologically justified for analysing fleet reliability data, as it explicitly models the clustered data structure ignored by ordinary least squares regression. This replication confirms the positive relationship between scheduled maintenance and reliability while providing a more