Abstract
{ "background": "The adoption of industrial machinery fleet management systems (FMS) is critical for enhancing productivity and safety in developing economies. However, robust methodological frameworks for analysing their adoption drivers, particularly within the complex hierarchical structures of industrial sectors, are lacking.", "purpose and objectives": "This study presents a methodological evaluation of multilevel regression modelling for measuring FMS adoption rates. Its objective is to assess the model's efficacy in capturing the nested data structure inherent in industrial sectors and to identify key adoption determinants.", "methodology": "A cross-sectional survey collected data from 347 machinery operators and managers nested within 42 firms. A two-level random intercept logistic regression model was specified: $\\logit(p{ij}) = \\beta{0} + \\beta X{ij} + u{j}$ where $p{ij}$ is the probability of adoption for operator $i$ in firm $j$, $X{ij}$ are covariates, and $u_{j}$ is the firm-level random effect. Model fit was evaluated using the intraclass correlation coefficient and likelihood ratio tests.", "findings": "The multilevel model demonstrated superior fit over a standard logistic model, with 31% of the variance in adoption odds attributable to firm-level differences. A key concrete result is that firms with formal maintenance protocols were 2.4 times more likely to adopt an FMS (95% CI: 1.7 to 3.5).", "conclusion": "Multilevel regression provides a statistically sound methodological framework for analysing FMS adoption, effectively accounting for clustered industrial data. It confirms that adoption is influenced by both operator-level and firm-level characteristics.", "recommendations": "Future adoption studies in engineering should employ multilevel modelling where hierarchical data structures exist. Policymakers and industry associations should prioritise interventions that standardise firm-level operational protocols to encourage adoption.", "key words": "fleet management systems, multilevel modelling, technology adoption, industrial engineering, Kenya", "contribution statement": "This