Replication Data for: Determinants of the Digital Divide in Somaliland
Description
Background: In the contemporary global economy, digital connectivity serves as a foundational catalyst for socio-economic development, equitable knowledge dissemination, and poverty alleviation. However, in fragile and post-conflict contexts such as Somaliland, digital inclusion remains profoundly uneven, governed by an intricate matrix of infrastructural, geographical, and socio-economic stratifications. Objective: This study empirically investigates the structural determinants of the digital divide in Somaliland. By aligning the analysis with the United Nations’ Sustainable Development Goals (SDGs) and Somaliland’s National Development Plan III (NDPIII), the research seeks to elucidate the intersecting physical, cognitive, and spatial barriers dictating household internet utilization. Methodology: The study utilizes a binary logistic regression model applied to a nationally representative dataset derived from the Somalia Integrated Household Budget Survey (SIHBS 2022) (N = 2,771). The model tests the predictive power of household electricity, mobile ownership, educational attainment, geographic location, distance to road networks, employment status, remittances, age, and sex. Results: The findings reveal that physical infrastructure acts as an absolute gatekeeper; mobile phone ownership and household electrification exponentially increase the odds of internet utilization. Educational attainment significantly bridges the "second-level" cognitive divide. Furthermore, severe spatial penalties exist: urban households benefit from a pronounced digital premium, whereas nomadic populations and those isolated from all-season roads face systemic exclusion. Transnational remittances were found to act as a vital compensatory mechanism to overcome affordability barriers in a liquidity-constrained labor market. Conclusion: The digital divide in Somaliland is deeply entrenched in broader structural inequalities. Achieving the inclusive digital ecosystems envisioned by the SDGs and NDPIII necessitates multi-sectoral policy interventions that transcend the mere expansion of telecommunications networks to concurrently address infrastructural resilience, spatial marginalization, and human capital development.
Files
Steps to reproduce
Steps to Reproduce: To fully replicate the binary logistic regression results (Table 3) and the Average Marginal Effects (Table 4) from the manuscript "Determinants of the Digital Divide in Somaliland," please follow these steps: 1. Software Requirements: Ensure you have R and RStudio installed. You will need the following R packages: dplyr (for data manipulation), broom (for tidying regression outputs), and margins or marginaleffects (for calculating Average Marginal Effects). 2. Data Setup: Download the dataset (SIHBS_2022_clean.csv) and the R script (digital_divide_analysis.R) from this repository. Open RStudio, set your working directory to the folder containing these files, and load the dataset. 3. Variable Formatting (Preprocessing): Ensure the dependent variable Internet_Utilization is coded as a binary numeric variable (0 = No, 1 = Yes). Ensure categorical independent variables are set as factors with the correct reference groups as specified in the manuscript (e.g., Geographical_Location should have "Urban" set as the reference level using the relevel() function). 4. Replicating Table 3 (Binary Logistic Regression & Odds Ratios): Run the standard logistic regression model using the glm() function with family = binomial(link = "logit"). To obtain the Odds Ratios (OR) and 95% Confidence Intervals as seen in Table 3, exponentiate the coefficients of the model using exp(coef(model)) and exp(confint(model)). 5. Replicating Table 4 (Average Marginal Effects - AME): Apply the margins() function from the margins package to the fitted glm model object. Use summary(margins(model)) to output the Average Marginal Effects (dY/dX), standard errors, z-values, p-values, and 95% confidence intervals exactly as reported in Table 4.
Institutions
- Amoud UniversityAwdal, Boorama