{ "background": "The adoption of advanced manufacturing systems in developing economies is a critical engineering challenge, yet robust methodological frameworks for analysing plant-level adoption drivers are scarce. Existing studies often fail to account for the hierarchical structure of industrial data, where plants are nested within sectors and regions.", "purpose and objectives": "This case study presents and evaluates a multilevel regression methodology to analyse the determinants of manufacturing systems adoption. It aims to demonstrate the application of this technique using Ugandan plant data and to assess its superiority over conventional single-level models for capturing contextual influences.", "methodology": "A methodological case study was conducted using a longitudinal dataset from a sample of manufacturing plants. The analysis employed a two-level random intercept model, formalised as $y{ij} = \\beta{0j} + \\beta X{ij} + \\epsilon{ij}$, where $\\beta{0j} = \\gamma{00} + u_{0j}$. Robust standard errors were used for inference, and model fit was compared using the deviance information criterion.", "findings": "The multilevel model demonstrated a significantly better fit than a pooled ordinary least squares regression, with a reduction in deviance exceeding 15%. A key substantive finding was that plant size had a positive but diminishing marginal effect on adoption likelihood, whereas sector-level technological infrastructure was a stronger predictor than individual plant capital expenditure. The intra-class correlation indicated that nearly 30% of the variance in adoption rates was attributable to sector-level differences.", "conclusion": "The case study confirms that multilevel regression is a methodologically rigorous approach for engineering adoption studies in an African industrial context, effectively disentangling plant-specific and sectoral influences.", "recommendations": "Future engineering management research on technology adoption in similar settings should employ multilevel modelling to avoid ecological fallacies and aggregation biases. Data collection efforts must be designed to capture relevant contextual variables at both plant and sector levels.", "key words": "multilevel modelling, manufacturing systems, technology adoption, industrial engineering, sub-Saharan Africa,