I intend to employ a panel dataset spanning from 1999 to 2023 to analyze long-term trends and patterns of urbanization, economic performance, ICT adoption, and environmental sustainability across 12 West African countries.
Specifically, the objectives are:
1. To examine the interaction effect between urbanization and ICT adoption on environmental sustainability in West Africa.
2. To analyze the dynamic relationship among urbanization, ICT adoption, and environmental sustainability in the region.
Given that all variables are either non-stationary at level or become stationary at first difference, the question arises:
What is the most appropriate estimation technique for this type of panel dataset, and what are the required steps?
You might want to consider big data analytics, use apps to visualize your data set like "orange" or R studio
The most appropriate estimation technique for analysing a panel dataset depends on the data-generating process: fixed-effects estimators are preferred when unobserved time-invariant heterogeneity is correlated with regressors; random-effects if that heterogeneity is uncorrelated (and you want time-invariant predictors retained); dynamic panels require IV/GMM (Arellano–Bond, system GMM) to handle lagged dependent variables and endogeneity; choose tests (Hausman, serial-correlation, instrument validity) to guide the selection.
- Badges
- Science topic