Modeling Voter Apathy With Demographics In Different Corruption Contexts In South-East Asian Democracies
Name
Aryan Poonacha
Major
Data Science
Class
2023
About
Aryan Poonacha
Signature Work Project Overview
Voter apathy is often pointed to as a central failing characteristic of many advanced and developing democracies alike. I analyze the different demographics that predict voter apathy by modeling the likelihood for an individual to vote in varyingly corrupt Asian democracies using the Asian Barometer survey dataset, adjusting for other numerous demographic factors, and the Corruption Perceptions Index for data on relative country corruption. I use demographics and perceived corruption to predict voter apathy. I first test a series of different kinds of models and use linear SVMs after they are shown to be the best performing models.
All the analysis and visualizations for this project can be recreated using the code from here: https://github.com/Aryan-Poonacha/sigwork
I found that ultimately, there is a correlation between voter apathy and corruption in a very general sense, and that this correlation is more or less similar for countries when split into groups according to their absolute/true corruption levels. However, there is a much more significant relationship between corruption of certain kinds more than others, like institutionalized corruption having a stronger correlation, which affirms the findings in Chang (2019); but overall, more dimensions of corruption need to be explored in a similar manner so that I can better view which aspects of it specifically affect voter apathy, and ultimately, the raw demographic data of individuals remains a stronger and and more accurate predictor of willingness to vote and voter apathy than perceived corruption, which I could model with 84% accuracy.
