"background": "Manufacturing systems in developing economies face unique challenges in yield optimisation, yet comparative analyses of their efficacy using advanced statistical techniques are scarce. This gap limits evidence-based decision-making for plant managers and policymakers. ", "purpose and objectives": "This study aims to methodologically evaluate and compare the effectiveness of three prevalent manufacturing systems—Total Productive Maintenance (TPM), Lean Manufacturing, and computerised maintenance management systems (CMMS) —in optimising production yield within an industrial context. ", "methodology": "A comparative multilevel regression analysis was conducted using operational data from a stratified sample of plants. The core statistical model is a three-level hierarchical linear model: Y{ij = \0j + \1jX1ij + \2X2ij + rij, where Yij is the yield for batch i in plant j, X1ij denotes system type, X2ij represents capital investment, and r₈₉ is the residual. Inference was based on robust standard errors. ", "findings": "The analysis indicates a statistically significant positive association between TPM implementation and yield, with an estimated coefficient of 0. 15 (95% CI: 0. 09, 0. 21). This effect persisted after controlling for plant size and capital expenditure, whereas the effects of Lean and CMMS were not statistically distinguishable from zero. ", "conclusion": "The methodological application of multilevel regression provides a robust framework for isolating the effect of manufacturing systems from confounding plant-level variables. TPM demonstrates a superior, quantifiable impact on yield optimisation in this context compared to the other systems evaluated. ", "recommendations": "Plant managers should prioritise investment in TPM systems for yield improvement. Policymakers and industry bodies are encouraged to develop support programmes facilitating TPM adoption, including training and technical guidance. ", "key words": "multilevel modelling, hierarchical linear model, yield
Kwame Gyeabour Asante (2024) studied this question.