Assessment of the Role of the Deming Cycle in Enhancing the Implementation of Constituency Development Fund (CDF) Projects in Lusaka

Abstract

This study applied the Deming Cycle (Plan-Do-Check-Act) to improve continual improvement practices in Constituency Development Fund (CDF) projects in Zambia. Three objectives guided the research: (i) assess how PDCA stages can strengthen planning, monitoring, and evaluation; (ii) analyze the extent PDCA addresses performance gaps; and (iii) develop a PDCA-based framework tailored to CDF projects. Despite CDF allocations increasing from K1.6 million per constituency in 2021 to K25.7 million in 2023, only 58% of funds are disbursed and 65% of projects face delays or abandonment. Independent variables comprised PDCA stage practices (PLAN, DO, CHECK, ACT). Dependent variables included project performance indicators (delays, cost overruns, community satisfaction, quality, functionality). Monitoring effectiveness, learning application, and planning quality served as mediating constructs. Explanatory sequential mixed-methods design across three Lusaka constituencies (Mandevu urban, Matero peri-urban, Chongwe rural-urban mix). Surveys (n=140, 94.0% response rate), 15 key informant interviews, 3 focus groups (n=30), and 40 project documents. Quantitative analysis used SPSS v26 (descriptive statistics, Pearson correlation, one-way ANOVA, multiple linear regression: Y = β₀ + β₁X₁ + β₂X₂ + β₃X₃ + ε). Qualitative analysis used NVivo 12 (thematic analysis). Systemic failures permeate all PDCA stages: PLAN (mean=2.83 consultation, 2.51 technical input), DO (mean=2.12 fund disbursement; 42.9% identify funding delays as primary cause), CHECK (mean=2.38 monitoring; 60.0% rate ineffective), ACT (mean=2.13; only 10.7% report formal corrective processes). Performance gaps are pervasive: 70.7% report frequent delays, 56.4% cost overruns, 60.0% dissatisfaction with community involvement, 57.1% indicate <50% project functionality post-completion. Correlation analysis showed monitoring effectiveness most strongly associated with outcomes (r=-0.62 with delays; r=0.59 with satisfaction). Regression confirmed monitoring as strongest success predictor (β=0.41, p<0.001), followed by learning application (β=0.29, p<0.001) and planning quality (β=0.21, p=0.003), explaining 50.4% of variance (R²=0.504). ANOVA revealed significant constituency differences (p<0.01), with Chongwe performing worse than Mandevu. Stakeholders strongly support PDCA principles (>85% agreement) but express capacity concerns (28.6%) and system flexibility doubts (25.0%). This study makes three original contributions. First, it bridges quality management theory (TQM/PDCA) with decentralized development practice, empirically validating PDCA's applicability in resource-constrained, politically sensitive grassroots project environments an underexplored context. Second, it provides the first empirical ranking of PDCA stage effects in development projects, demonstrating monitoring (CHECK) as the strongest success predictor (β=0.41), followed by learning (ACT) (β=0.29) a hierarchical finding with resource allocation implications. Third, it develops the CDF-PDCA Continuum Model, a practical framework with constituency-specific adaptations (urban political accountability, peri-urban capacity building, rural accessibility solutions), operationalizing continuous improvement for Zambia's largest decentralized funding mechanism.