Voided Ballots in the 2012 Parliamentary Election in Serbia: Abstinence, Protest, Irrationality, or Manipulation?

Abstract

The "white ballots" campaign in the 2012 parliamentary election in Serbia was accompanied by a change in the portion of invalid ballots, but also in their correlation with socio-demographic variables. When the percentage of invalid votes in the three previous elections is analyzed using the generalized least squares (GLS) multiple regression analysis, the percentage of inhabitants with incomplete primary education represents the best predictor. The percentage of individuals with primary education is also statistically relevant, just in the opposite (negative) direction, and less predictive. When GLS is applied to invalid votes in the 2012 election, incomplete primary education is positively, and primary education is negatively correlated with the criterion. The percentage of illiterates, the share of people over 65 years of age and the degree of urbanization are significantly, but negatively related to the criterion. The consequences of the 2012 election outcome were considered in the light of the regression to competitive authoritarianism.