Impacts of NEWMAP Intervention on Sustainable Development of South East Nigeria
Description
This dataset contains household-level panel data used to analyze the impact of environmental interventions on clean energy adoption in Nigeria. It supports the study titled “Environmental Intervention and Household Clean Energy Adoption: Assessing the Impact of Nigeria’s NEWMAP on Cooking and Lighting Energy Use.” The dataset enables examination of how participation in the Nigeria Erosion and Watershed Management Project (NEWMAP) influences household transitions to cleaner cooking and lighting energy sources. The data originate from household surveys conducted in communities participating in the NEWMAP program, a World Bank-supported environmental intervention designed to address land degradation, erosion, and watershed management in Nigeria. Surveys were administered at two time points, baseline (pre-intervention) and endline (post-intervention) allowing for a quasi-experimental evaluation of program impacts. The dataset is structured as a balanced household panel. The unit of observation is the household, tracked across two periods (baseline and endline). The final analytical sample consists of 364 households observed in both rounds, yielding 728 total observations. Households are uniquely identified using a consistent identifier across survey waves. The dataset includes several categories of variables. Outcome variables capture household energy use behavior: clean cooking adoption (equal to 1 if the household primarily uses LPG or electricity for cooking, and 0 otherwise) and clean lighting adoption (equal to 1 if the household uses electricity or solar energy for lighting, and 0 otherwise). Key explanatory variables include a treatment indicator identifying households located in areas that received NEWMAP physical environmental interventions, and a post-intervention indicator distinguishing endline from baseline observations. Their interaction term enables estimation of causal effects using a difference-in-differences framework. The dataset also contains a rich set of control variables. Socioeconomic and demographic variables include age, gender, education, literacy status, household size, and employment status. Economic indicators include household income, total expenditure, housing ownership, and housing characteristics. Infrastructure and access variables include electricity access, distance to water sources, and market access. Geographic variables classify households as rural, peri-urban, or urban and include community or watershed identifiers. Data were collected using structured questionnaires administered by trained enumerators following standardized survey protocols. Data processing involved cleaning and validation of responses, harmonization of variables across survey rounds, construction of derived variables (including clean energy indicators), and merging of baseline and endline datasets using household identifiers. Observations with missing or inconsistent identifiers were excluded to ensure data quality and consistency.
Files
Steps to reproduce
Steps to Reproduce the Results The results presented in the main paper and Appendix were generated using household-level panel data from the Nigeria Erosion and Watershed Management Project (NEWMAP). The following steps describe how to reproduce the analysis. 1. Data Preparation and Sample Construction Load the raw dataset containing baseline and endline household survey data (840 observations). Merge baseline and endline datasets using a unique household identifier. Restrict the sample to households observed in both survey rounds to construct a balanced panel. Remove observations with inconsistent or missing household identifiers. The final analytical sample consists of 364 households observed over two periods (728 observations) . 2. Variable Construction Outcome Variables Clean Cooking Adoption: Binary variable equal to 1 if the household primarily uses LPG or electricity for cooking, 0 otherwise. Clean Lighting Adoption: Binary variable equal to 1 if the household uses electricity or solar for lighting, 0 otherwise . Treatment Variable Physical Intervention (Treatment): Binary indicator equal to 1 for households located in watersheds receiving physical restoration interventions under NEWMAP, 0 otherwise. Post Period: Binary indicator equal to 1 for endline survey, 0 for baseline. Interaction Term (DiD): Treatment × Post captures the causal effect of the intervention . Control Variables Construct the following covariates: Socioeconomic: age of household head, literacy, education, household size; Wealth: income, total expenditure, housing ownership, housing type Infrastructure: electricity access, distance to water, market access; Location: rural, peri-urban, urban categories 3. Descriptive and Preliminary Analysis Generate summary statistics for all variables. Compare baseline characteristics between treatment and control groups. Plot mean adoption rates over time to assess parallel trends assumption . 4. Estimation of Determinants (Logistic Regression) Estimate cross-sectional models of clean energy adoption: Model specification: CleanEnergyᵢ = f(Socioeconomicᵢ, Wealthᵢ, Infrastructureᵢ, Locationᵢ); Use logistic regression (logit model) separately for: Clean cooking Clean lighting Report coefficients, standard errors, and significance levels as shown in Table 1 and Appendix Table A1 . 5. Difference-in-Differences (DiD) Estimation Estimate the causal impact of NEWMAP using a panel DiD model: Model specification: Energy_it = β₀ + β₁Physicalᵢ + β₂Post_t + β₃(Physicalᵢ × Post_t) + X_it + μᵢ + τ_t + ε_it Where: μᵢ = household fixed effects; τ_t = time fixed effects X_it = time-varying controls (income, household size, expenditure) Steps: Estimate DiD models for: Clean cooking adoption Clean lighting adoption Cluster standard errors at the watershed level . Interpret the interaction term as the treatment effect. 6. Robustness Checks To reproduce Appendix results: Re-estimate models: With and without control variables Including infrastructure
Institutions
- University of NigeriaEnugu State, Nsukka