"background": "Industrial machinery fleets in Nigeria operate under significant constraints, including erratic power supply and limited maintenance resources, leading to chronic inefficiencies. Current diagnostic methods often rely on retrospective data or theoretical models, lacking the rigour to isolate causal factors of performance variation in real-world operational environments. ", "purpose and objectives": "This article presents a novel quasi-experimental framework designed to measure and attribute efficiency gains within industrial machinery fleets. Its primary objective is to provide a structured methodology for conducting causal inference on operational interventions without randomised control trials. ", "methodology": "The framework employs a difference-in-differences design, comparing a treatment group (machinery undergoing a targeted predictive maintenance protocol) with a matched control group. The core statistical model is Y{it = \0 + \1 + \2 + \ (\) +, where Yit is the efficiency metric. Robust standard errors are clustered at the machine-unit level to account for serial correlation. ", "findings": "As this is a methodology article, no empirical results are presented. However, the framework's diagnostic power is demonstrated through a simulated case study, where it correctly identified a 12% net efficiency gain attributable to the intervention, with a 95% confidence interval of 8. 5%, 15. 7%. ", "conclusion": "The proposed framework provides a robust, field-applicable method for engineering managers to quantitatively evaluate the impact of operational changes on machinery fleet efficiency, moving beyond descriptive analytics. ", "recommendations": "Practitioners should adopt this quasi-experimental approach for pre- and post-implementation analysis of new maintenance regimes or technology retrofits. Further research should adapt the framework for cross-sectoral application. ", "key words": "quasi-experimental design, efficiency diagnostics, industrial machinery, predictive maintenance, causal inference
Nwosu et al. (Wed,) studied this question.