Multiple Regression Analysis of Sustainable Stakeholder Engagement and Project Success: Evidence from Road Infrastructure Projects in Rwanda

Infrastructure agencies increasingly treat stakeholder engagement as a governance requirement rather than a measurable predictor of delivery outcomes, leaving practitioners with limited quantitative guidance on which engagement practices actually move the needle on project success. This gap is particularly acute in Sub-Saharan African infrastructure programs, where engagement research has relied largely on descriptive or single-case accounts rather than tested predictive models. This study addresses that gap by modelling the statistical relationship between sustainable stakeholder engagement (SSE) and project success across three World Bank-supported urban road corridors in Rwanda (Huye, Musanze, and Nyarugenge/Kigali). Cross-sectional survey data were collected from 107 respondents, 77 technical stakeholders drawn through stratified proportional sampling from a population of 95, and 30 purposively selected community representatives, and complemented by 34 key-informant interviews. Sustainable stakeholder engagement was operationalized across four dimensions (transparent communication, community engagement and inclusion, environmental responsibility, and long-term relationship building) and entered as predictors of project success in a multiple linear regression model, alongside project size and location as controls. The model was statistically significant, F (6, 89) = 19.354, p < .001, and explained 56.6% of the variance in project success (adjusted R² = .537). Long-term relationship building (β = .504, p = .001) and community engagement (β = .274, p = .001) emerged as the only statistically significant behavioral predictors, while transparency and environmental responsibility, despite positive bivariate correlations, did not retain independent predictive value once the other variables were controlled. Diagnostic checks (VIF 1.2-3.1; Durbin-Watson = 2.151) confirmed the model met the assumptions required for valid inference. The findings indicate that durable, relationship-based engagement, not communication volume alone, is the dominant statistical driver of infrastructure project success in this setting, with direct implications for how agencies such as RTDA and MININFRA design and resource engagement strategies.