PREDICTING VOTER CHOICES: THE CASE OF MONGOLIA

Authors

  • Золзаяа Г. ХААИС, Эдийн засаг, бизнесийн сургуулийн Статистик ба эконометриксийн тэнхим
  • Өсөхбаяр Б. ХААИС, Хэрэглээний шинжлэх ухааны сургууль, Компьютерийн шинжлэх ухааны тэнхим

DOI:

https://doi.org/10.22353/PS20261.10

Keywords:

Multinomial logistic regression for qualitative data, Electoral forecasting

Abstract

In democratic societies, the electoral process constitutes the primary mechanism for the legitimate formation of political authority, with voters’ political choices directly influencing socio-economic development and policy outcomes. Research on voter behavior in Mongolia remains limited, with most studies relying on descriptive statistical summaries. This study employs a Multinomial Logistic Regression (MLR) model to analyze the combined effects of demographic characteristics, socio-economic indicators, and party evaluations among voters from two electoral districts. The results indicate that age, education, income, sources of political information, political activism, previous voting behavior, and party evaluations are statistically significant predictors of voter choice (p < 0.05). The model accurately predicts support for the Mongolian People’s Party (MPP) and Democratic Party (DP) with 84–98% precision, whereas predictions for the ‘Other’ category are weaker, likely due to heterogeneity and a small number of observations. Model fit indices (Cox & Snell R² = 0.745–0.791; Nagelkerke R² = 0.803–0.852) demonstrate a high explanatory capacity. The findings underscore the theoretical and methodological relevance of data-driven electoral forecasting in the Mongolian context and highlight the need for further research into small-party heterogeneity and latent voter structures.

Downloads

Download data is not yet available.

Downloads

Published

2026-03-26

How to Cite

Г., З., & Б., Ө. (2026). PREDICTING VOTER CHOICES: THE CASE OF MONGOLIA. Politology, 24(633), 162–177. https://doi.org/10.22353/PS20261.10