“None of the Above?” Casting Blank Ballots in Ontario Provincial Elections

Abstract

Are blank or unmarked ballots in an election a random event? The comparative literature on elections suggests otherwise, linking blank ballots to political protest or social and demographic characteristics of voters. Rational choice theory offers explanations rooted in civic duty and net benefits associated with uncompetitive elections. Similarly, when a voter decides to vote, the literature on party competition and electoral turnout offers the possibility that a blank ballot is another choice beyond voting for a specific candidate or party. While little research exists on blank ballots in election systems lacking compulsory voting, this article presents an exploratory analysis using data from the 1999 and 2003 provincial elections in Ontario. The study finds that while demographic characteristics matter, party competition also provides an explanation for casting blank ballots consistent with expectations of rational choice theory.

What Affects Invalid Voting? A Review and Meta-Analysis

“…Despite that, the aggregate research continues to dominate the field. We counted 26 aggregate-level studies employing regression models of the determinants of invalid voting (Ackaert et al 2011; Aldashev and Mastrobuoni 2013; Cisneros 2013; Cisneros and Freigedo 2014; Cohen 2018; Damore et al 2012; Dejaeghere and Vanhoutte 2016; Fatke and Heinsohn 2017; Galatas 2008; Gendźwiłł 2015; Herron 2011a; Hooghe et al 2009; Kimball and Kropf 2005; Kouba and Lysek 2016; McAllister and Makkai 1993; Nihuys 2014; Pachón et al 2017; Pion 2010; Power and Garand 2007; Power and Roberts 1995; Reynolds and Steenbergen 2006; Rosenthal and Sen 1973; Singh 2017; Socia and Brown 2014; Superti 2015; Uggla 2008), while 11 studies utilize data from individual survey responses (Arbache et al 2014; Borba 2008; Carlin 2006; Cisneros 2016; Cohen 2017; Hill and Rutledge-Prior 2016; Hooghe et al 2009; Katz and Levin 2016; Moral 2016; Singh 2017; Solvak and Vassil 2015). Two studies combine both (Driscoll and Nelson 2014; Hill and Rutledge-Prior 2016).…”

“…The closer the electoral results, the lower the numbers of invalid votes as the voters perceive that their vote has a bigger value. Other measurements include the vote share of the first party (Uggla 2008), the specific competitiveness index (Galatas 2008) and a square root of victory margin (Fatke and Heinsohn 2017). One of the two deviant cases is Italy, where the administrative error of the electoral authority is assumed to operate (Aldashev and Mastrobuoni 2013), but all other studies report the assumed positive relationship of closeness to invalid voting.…”

Invalid Votes, Deliberate Abstentions, and the Brazilian Crisis of Representation

“…Around the world, other studies address the concern of protest votes as an expression of personal preferences. Related studies include: Power and Garand (), who associate socioeconomic variables with the number of invalid votes, blank ballots, and spoiled ballots in 18 Latin American democracies; Schumacher (2014), who attributes the considerable support of protest voters for the Green Party of Germany to its alternative view of the world; and Galatas (2008), who analyzes party competition in Ontario provincial elections as a way to influence the submission of blank ballots.…”

¿Protesta o cambio institucional? Determinantes del voto nulo y blanco en las elecciones departamentales de 2010 en Uruguay en perspectiva comparada

“…En la mayoría de los estudios, las primeras dos variables tienen una asociación negativa, mientras que el analfabetismo y la migración tienen una asociación positiva con el voto inválido, respectivamente. Sobre esto último, McAllister y Makkai (1993), al igual que Galatas (2008), encontraron que el sufragio inválido está determinado por el origen étnico.…”

“…Con respecto a la escolaridad, Power y Garand (2007) señalaron que altos niveles de educación están asociados con bajos niveles de voto nulo. Apoyando lo anterior, Galatas (2008) encontró que los votos anulados por error tienden a asociarse negativamente con bajos grados de instrucción. Ante esto, la relación esperada por este enfoque entre las variables voto nulo y escolaridad es negativa.…”