This project investigates the research question of what factors motivate people to participate in national elections. Hypotheses are drawn for personal motivations as well as for the interplay between these individual-level drivers and structural country-level factors. The expectations are tested with Bayesian logistic and multilevel logistic regressions. The Bayesian MCMC sampling is achieved via model formulation in Stan files as well as a manual implementation with the Metropolis Hastings algorithm.
To run the project, simply execute the master script. This script loads custom functions from functions and executes all individual R scripts collected under scripts.
Data from the European Social Survey (ESS) is used to measure the degree of political participation. Information on country-level degrees of social expenditures, corruption levels and youth unemployment is obtained by the OECD data portal and the World Bank.
All plots created for the report saved as png-files.
All models are implemented as Stan files which are stored in this folder. Additionally, to get a better understanding of the sampling process, each model is also implemented via the Metropolis Hastings algorithm in an additional R script.
Contains the posterior matrices for all models.
Here you can find the final analysis report as a PDF-file.
All R scripts for data preprocessing, model loading, diagnostics and visualizations. These files get loaded when running the master file.