"background": "Industrial machinery fleets in Nigeria are critical to economic output, yet systematic methodologies for diagnosing and quantifying efficiency gains in their operational systems are lacking. Current evaluations often rely on retrospective, non-experimental analyses, which are susceptible to confounding and bias. ", "purpose and objectives": "This article presents a novel methodological framework for conducting randomised field trials (RFTs) to robustly measure efficiency improvements within industrial machinery systems. The objective is to provide a structured protocol for the design, implementation, and analysis of such trials in an active industrial setting. ", "methodology": "The proposed RFT methodology employs a cluster-randomised, stepped-wedge design where fleets or sites are randomised to receive diagnostic interventions at different sequential steps. The core statistical model for estimating the treatment effect on a continuous efficiency metric (Y) is a linear mixed model: Y{ijt = \ + \ Tijt + \ + \ +, where Tijt is the treatment indicator for cluster i, machine j at time t, \ and \ are random cluster and fixed time effects, respectively, and \₈₉ₓ is the error term. Inference is based on cluster-robust standard errors. ", "findings": "As a methodology article, this paper presents no empirical results from a completed trial. However, a pilot simulation based on typical fleet data indicates that the proposed design can detect a minimum 12% relative improvement in mean time between failures (MTBF) with 80% power, assuming an intra-cluster correlation coefficient of 0. 15. ", "conclusion": "The stepped-wedge RFT framework provides a rigorous and feasible approach for obtaining causal evidence on efficiency diagnostics in real-world industrial machinery operations, overcoming key limitations of observational studies. ", "recommendations": "Researchers and industrial engineers should adopt this experimental methodology to generate high-quality evidence for operational decisions. Future
Okonkwo et al. (Mon,) studied this question.