Stakeholder engagement and the effectiveness of the Parish Development Model in Bushenyi District, Uganda: a cross-sectional analysis with a measurement caution

Abstract

Background Uganda’s Parish Development Model (PDM) is a flagship decentralised rural-development programme premised on coordinated engagement among civil society organisations (CSOs), local communities, and government institutions. Whether these three engagement dimensions are empirically associated with perceived programme effectiveness has rarely been tested with a fully diagnosed regression model at district level. This study tested that relationship in Bushenyi District. Methods A cross-sectional, correlational design was used. A structured, five-point Likert questionnaire was self-administered to 189 respondents drawn by simple random and purposive sampling from a target population of 358 district officials, technical staff, civil society representatives, and PDM beneficiaries; 169 usable responses were returned (89.4%). CSO, local community (LCE), and government (GE) engagement were each measured with six items, and perceived programme effectiveness with six further items. Data were analysed using Pearson correlation and multiple regression tested three pre-registered hypotheses. Results Reliability screening showed two of four constructs far below the 0.70 threshold (CSO, α = 0.09; effectiveness, α = −0.04), with the remaining two marginal (LCE, α = 0.45; GE, α = 0.58). Item-level inspection traced this to substantial response overlap across construct blocks, which also produced severe multicollinearity (VIF up to 7.65). The overall regression model was not significant (R 2  = 0.009, adjusted R 2  = −0.010; F(3,165) = 0.47, p = .702), and none of CSO (β = 0.20, p = .281), LCE (β = −0.17, p = .438), or GE engagement (β = −0.02, p = .823) significantly predicted perceived effectiveness. Conclusions the study concluded that, no statistically defensible relationship between the three engagement dimensions and perceived PDM effectiveness could be established. This contrasts with prior Uganda-focused evidence and should be read alongside the measurement limitation identified: it does not show stakeholder engagement is irrelevant.