본문 바로가기
  • Home

Spatial Relationship Analysis between the 18th Presidential Election Turnout, Support Rates, and Demographic Variables using Geographically Weighted Regression (GWR)

Hong Il Young 1 Bo Ae Chun 2 Lee, Gyoung-ju 3

1남서울대학교
2가톨릭관동대학교
3한국교통대학교

Accredited

ABSTRACT

In this study, we analyzed the spatial relationship between the 2012 presidential election turnouts, support rates, and the demographic factors using Geographically Weighted Regression (GWR). Until recently, understanding he voter behavior of election has been pursued through the traditional regression analysis with the selected sample data. On the contrary, this study investigated the 2012 election with the aggregated data at the national level using GWR with spatial variation. It is well known that the voter turnout and support rate are generally associated with demographic factors such as age and education level. Thus, we select these as the independent variables and turnout and support rate as the dependent variables. We also explore the regional variation of the model with conservative groups and progressive groups. The data for the dependent variable were acquired from the National Election Commission web site and the data for independent variables from Statistics Korea web site. The results indicated that the GWR models predicted the presidential election turnout and the each candidate's support rates better than the traditional OLS (Ordinary Least Squares) model at the goodness of fit of the model and also provided useful information regarding the spatial variation of the demographic factors such as the age and education level affecting the regional turnout rates and the each candidate's support rates.

Citation status

* References for papers published after 2022 are currently being built.