Journal Design Engineering Masthead
African Structural Engineering | 12 January 2015

A Difference-in-Differences Analysis of Water Treatment System Adoption in Tanzania

A Methodological Evaluation, 2000–2026
A, m, i, n, a, M, w, i, n, y, i, ,, J, u, m, a, M, w, a, k, a, l, i, n, g, a, ,, N, e, e, m, a, M, t, e, i
Causal InferenceInfrastructure AdoptionQuasi-ExperimentalImpact Evaluation
DiD analysis shows 18pp increase in water treatment adoption with intervention.
Parallel trends assumption held in pre-intervention period.
Causal estimates are sensitive to alternative model specifications.
Highlights need for rigorous testing of identification assumptions.

Abstract

{ "background": "Evaluating the impact of large-scale infrastructure interventions, such as water treatment systems, requires robust quasi-experimental methods to establish causal inference, a persistent challenge in structural engineering project assessment.", "purpose and objectives": "This study provides a methodological evaluation of the difference-in-differences (DiD) model for measuring the adoption rates of community-scale water treatment facilities, assessing its applicability and robustness within a longitudinal infrastructure development context.", "methodology": "A longitudinal panel dataset was constructed from national utility records and field surveys. The primary analysis employed a two-way fixed effects DiD model: $Y{it} = \\alpha + \\beta (Treati \\times Postt) + \\gammai + \\deltat + \\epsilon{it}$, where adoption $Y_{it}$ is the outcome for community $i$ in period $t$. Inference was based on cluster-robust standard errors at the district level.", "findings": "The DiD estimator identified a statistically significant positive effect, with system adoption increasing by approximately 18 percentage points in intervention communities relative to controls (95% CI: 12 to 24). The parallel trends assumption held pre-intervention, but sensitivity analyses revealed the result was not robust to certain alternative model specifications.", "conclusion": "The DiD framework is a valuable but context-sensitive tool for evaluating engineering infrastructure adoption, capable of generating credible causal estimates when key identification assumptions are rigorously tested.", "recommendations": "Future engineering impact evaluations should incorporate pre-intervention parallel trends testing and sensitivity analyses as standard practice. Data collection protocols must be designed to facilitate the construction of suitable control groups from project inception.", "key words": "difference-in-differences, causal inference, water infrastructure, adoption rates, impact evaluation, quasi-experimental design", "contribution statement": "This paper provides the first formal application and critical assessment of the DiD methodology for evaluating the rollout of decentralised water treatment infrastructure at a national scale, demonstrating its utility and limitations for structural engineers."