Assessing Ballot Structure and Split Ticket Voting:

Evidence from a Quasi-Experiment

Tiffany D. Barnes, University of Kentucky
Carolina Tchintian, Rice University
Santiago Alles, Rice University

Although a growing number of countries have implemented electronic voting, few scholars have considered the un-
intended consequences of such reforms. We argue that changes in ballot structure imposed by electronic voting, im-
plemented under the exact same electoral rules, can facilitate ballot splitting. Exploiting data from three elections and a
novel ballot reform in Salta, Argentina—electronic voting was incrementally introduced over multiple elections—we
provide an empirical analysis of how ballot structure influences ballot splitting. We use the Geographic Information
System to reconstruct precinct demographics and matching to address threats to random assignment. This empirical
strategy allows us to treat our data as a quasi-experiment. We find that precincts casting electronic ballots under an
Australian ballot, rather than the ballot-and-envelope system, have significantly higher rates of ballot splitting. Our
findings imply that less complicated voting procedures can affect the composition of legislative representation and
manufacture a more inclusive legislature.

I n recent years, many governments have changed their vot-

ing procedures in an attempt to make elections more ef-
ficient and reliable (Alvarez and Hall 2008; Alvarez, Katz,
and Pomares 2011). In particular, dozens of countries around
the world have piloted or adopted electronic voting devices
in many electoral systems and is extremely important for
representation because it influences the distribution of leg-
islative seats—and subsequently which voters, policies, and
preferences are represented in office (Massicotte, Blais, and
Yoshinaka 2004; McAllister 2015). Voters may split their
designed to increase efficiency, confidence, and turnout by ballot to produce ideological or partisan balance (Burden
offering more user-friendly procedures for voters and autom- and Helmke 2009), to support candidates who exhibit de-
atizing vote tallying (Alvarez et al. 2009; Tula 2005). Schol- sirable personal traits (Beck et al. 1992), or to vote strate-
ars have devoted considerable attention to understanding gically to avoid wasting their vote (Cox 1997; Moser and
whether the implementation of electronic voting successfully Scheiner 2009). Despite voters’ incentives to split their bal-
improves electoral processes and bolsters confidence in elec- lots, not all voters are willing to pay the costs associated with
tions. Yet the implementation of electronic voting frequently splitting their ticket. We argue that altering something as
introduces small adjustments to the ballot structure that may simple as the ballot structure, while maintaining the exact
have unintended effects on representation (Calvo, Escolar, same electoral rules, can lower the cost of split ticket vot-
and Pomares 2008; Katz et al. 2011). We contend that seem- ing, thereby significantly increasing the probability of ballot
ingly minor changes have major consequences for how citi- splitting.
zens cast their vote. Specifically, we examine how modifica- Indeed, complicated ballot structures, such as the ballot-
tions to ballot structure imposed by electronic voting influence and-envelope system (i.e., the French system, in which each
ballot splitting. party prints and distributes a unique partisan paper ballot,
Split ticket voting (i.e., voting for a different party for and voters can either vote straight ticket or create their
different contests in the same election) is a common feature own mixture of partisan ballots by physically splitting their

Tiffany D. Barnes ([email protected]) is an assistant professor at the University of Kentucky. Carolina Tchintian ([email protected]) is a
PhD candidate at Rice University. Santiago Alles ([email protected]) is a PhD candidate at Rice University.
Data and supporting materials necessary to reproduce the numerical results in the paper are available in the JOP Dataverse (https://dataverse.harvard.edu
/dataverse/jop). An online appendix with supplementary material is available at http://dx.doi.org/10.1086/688677.

The Journal of Politics, volume 79, number 2. Published online January 26, 2017. http://dx.doi.org/10.1086/688677
q 2017 by the Southern Political Science Association. All rights reserved. 0022-3816/2017/7902-0006$10.00 000

This content downloaded from 132.239.001.230 on February 02, 2017 09:21:58 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
000 / Ballot Structure and Split Ticket Voting Tiffany D. Barnes, Carolina Tchintian, and Santiago Alles

ballots and then placing their vote in an envelope), create are statistically significant differences between the treatment
unnecessary complications that may discourage ballot split- and control groups in 2011 under partial implementation,
ting. By contrast, more user-friendly ballot structures, such there are no differences between the treatment and control
as the Australian ballot (in which the candidates from all groups in the 2007 (no implementation) or 2015 (full im-
political parties are grouped on a single ballot by office), make plementation) elections, indicating that the increase in bal-
it easier and more convenient for voters to split their ballot lot splitting is a product of the ballot reform. Results from a
and may even encourage such behavior (Burden and Kimball precinct-party–level analysis provide some insight into how
2002; Rusk 1970). modifications to the ballot structure may influence ballot
To test our argument, we take advantage of a novel ballot splitting differently across political parties. In particular, we
reform in Salta, Argentina, where electronic voting was in- see that electronic voting led to an uptick in ballot splitting
crementally implemented over multiple elections. In 2007, among smaller political parties where the electoral payoff
the entire electorate used the traditional ballot-and-envelope for split ticket voting was less clear. Taken together, our re-
system. In 2011, for the first time, half of the voters in the sults demonstrate how changes in voting procedures and bal-
district of Salta used electronic devices, while the other half lot structures, within the exact same electoral institutions, can
used the ballot-and-envelope system. Finally, in 2015, all transform a system from being one where voters are effec-
voters used electronic voting.1 The reform only introduced tively discouraged from splitting their votes to one where vot-
one change—with electronic voting voters were given the ers are able to deliberately maximize their interests by more
option to view the ballot either arranged by political party accurately translating their preferences at the polls.
(similar to the ballot-and-envelope system) or by elected This result has important implications for public policy.
positions (similar to the Australian ballot). If voters chose to Although many leaders are resistant to major electoral re-
vote by political party, it was similar to viewing the partisan forms known to influence election outcomes, they are more
paper ballots. But if they chose to vote by elected position, open to minor, seemingly innocuous, adjustments to voting
the voter could easily vote for different parties for each po- procedures and ballot structures. Yet our findings indicate
sition. The availability of this second option makes vote that electoral engineers might be introducing new incen-
splitting easier. tives for split ticket voting even when reforming small, ap-
The quasi-experimental design facilitated by the partial parently neutral voting measures. These minor alterations in
implementation of the reform in 2011 offers a unique op- ballot structures, moreover, have important implications for
portunity to assess the impact of ballot structures on ballot representation linkages in democracies. As executive races
splitting using a difference-in-difference strategy that com- are the locus of electoral competition in presidential sys-
pares the likelihood of ballot splitting in those precincts tems, concurrence between executive and legislative elec-
with and without electronic voting. As quasi-experiments do tions reduces the number of legislative parties (Cox 1997;
not involve an explicit random assignment, we construct an Jones 1997; Shugart and Carey 1992). When ballot struc-
original data set using Geographic Information System (GIS) tures erode the influence the executive race exerts over the
technology to identify precinct demographics and employ legislative race by facilitating ballot splitting, they are likely
matching methods to reproduce the properties of a randomized to increase the effective number of legislative parties and
experiment using observational data. This empirical approach thereby manufacture more inclusive and representative leg-
allows us to explicitly address endogeneity issues associated islative outcomes.
with institutional reforms—that is, the concern that electronic
voting was assigned in precisely those precincts where voters THE (UN)INTENDED CONSEQUENCES OF BALLOT
have a higher proclivity to split their ballots. STRUCTURES AND VOTING PROCEDURE
Results from difference-in-difference analyses demon- In recent decades, the adoption and piloting of different
strate that precincts using electronic voting with the option voting procedures and ballot structures have spurred con-
to view the Australian ballot structure (treatment group), as siderable research on the effect of these reforms on electoral
compared to the ballot-and-envelope system (control group), competition and political outcomes (Alvarez and Hall 2008;
have significantly higher rates of ballot splitting. Whereas there Herrnson et al. 2009). Governments reform their voting
procedures in an effort to increase political participation
(Barnes and Rangel 2014; Kersting and Baldersheim 2004),
1. In 2013, all voters used electronic ballots, and in 2009, all voters
reduce electoral fraud (Fujiwara 2015; Rezende 2003), and
used paper ballots. These, however, were midterm elections, and thus, they bolster confidence in the electoral process (Alvarez et al.
are not useful for studying executive-legislative ballot splitting. 2009, 2011; Thompson Jiménez 2009; Tula 2005).

This content downloaded from 132.239.001.230 on February 02, 2017 09:21:58 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 Volume 79 Number 2 April 2017 / 000

For example, electronic voting (herein e-voting) in Brazil party-centric informational cues, such as party name or logo,
was introduced in part as an effort to lower barriers to po- decrease the likelihood of split ticket voting.
litical participation among less educated voters by elimi- We contribute to this nascent body of literature by ex-
nating the difficulties associated with writing-intensive pa- amining another unintended—but electorally important—
per ballots (Fujiwara 2015; Rezende 2003). Still, in other consequence of ballot structure. Specifically, we examine
countries, concerns about election transparency and fraud how changes in the ballot structure introduced by e-voting
have triggered the adoption of different versions of the Aus- in Argentina affect the way voters translate their prefer-
tralian ballot (Alvarez et al. 2011; Prasad et al. 2010) as a way ences into votes—either by casting a straight ticket ballot
of centralizing ballot provision and guaranteeing the avail- in concurrent executive-legislative elections or by splitting
ability of every ballot in every voting center. In the ballot- their ballot.
and-envelope system—a paper ballot that, until recently, was
common in Latin America, (e.g., Argentina, Panama, Colom- BALLOT SPLITTING AND THE CRUCIAL IMPACT
bia, and Uruguay) and Africa (e.g., Chad, Guinea, Burkina OF BALLOT STRUCTURES
Faso, Mali, Mauritania, Niger, Senegal, and Tunisia)—parties Voters are often called upon to make multiple decisions for
are responsible for printing, distributing, and stocking their different contests in the same election. Although straight ticket
own ballots. Under such rules, smaller parties often lack the voting, in which voters support the same political party for
resources to ensure ballot availability during elections and each contest at stake, is the dominant form of voting, a num-
even run the risk of their ballots being stolen from some ber of voters prefer to split their ballot. Split ticket voting
polling stations (ACE 2011; Pomares et al. 2014b). Con- occurs when voters choose between parties competing to fill
versely, the Australian ballot system and e-voting can both political posts at different levels of government (e.g., vertical
reduce fraud by assuring voters have the option to vote for ticket splitting) or between parties competing for equivalent
each of the parties competing in the electoral contest. Simi- offices (e.g., horizontal ticket splitting; Burden and Helmke
larly, confidence in the electoral process is a major concern 2009; Campbell and Miller 1957).
when it comes to ballot reforms. With respect to e-voting in The level of split ticket voting varies substantially across
particular, whereas earlier studies in the United States have different political and institutional contexts. In the United
found that voters are skeptical about reforms (Alvarez and States, for instance, ticket splitting between presidential and
Hall 2008; Alvarez, Hall, and Llewellyn 2008), more recent House member candidates rose from about 12% of voters
research from the United States, Belgium, Argentina, and Co- in the 1950s to more than 25% during the 1980s, and then
lombia suggests that, conditional on demographic character- it declined again in the 1990s (Burden and Kimball 2002).
istics, citizens’ concerns surrounding e-voting are abating (Al- These values vary across congressional districts, depend-
varez et al. 2011, 2013; Beaulieu 2015; Delwit, Kulachi, and ing in part on how contested the election is (Brunell and
Pilet 2005). Grofman 2009; Burden and Kimball 1998, 2002). Further,
Although research has most often focused on these in- split ticket voting tends to be higher in countries with
tended consequences of ballot reforms, new ballot structures multi-party systems and where political parties are weakly
and voting procedures may also have meaningful unin- institutionalized (Burden and Kimball 2002; Moser and
tended consequences for election outcomes. For example, in Scheiner 2009). In Brazil, for example, where the number of
the United States, e-voting attenuates ballot roll-off (Nich- effective parties is high, approximately 70% of voters split
ols and Strizek 1995). Complicated ballot structures, by com- their ticket between presidential and legislative candidates
parison, may cause confusion and frustration for voters and (Ames, Baker, and Reno 2009). In Mexico, by comparison,
ultimately increase the number of invalid votes (Kimball and where there are far fewer competitive political parties, less
Kropf 2005). Moreover, changes in the voting procedure can than 10% of voters split their ballot during concurrent pres-
bias outcomes by inconsistently influencing turnout across idential elections (Helmke 2009). Horizontal ticket split-
different groups of voters and shifting vote choice in favor of ting in mixed member systems also varies substantially,
different parties or candidates. For instance, Card and Mo- ranging from a low of 8.73% in Hungary to more than 30%
retti (2007) show that the adoption of touch screen machines in New Zealand and South Korea (Rich 2014).
in the United States was associated with lower turnout in Incentives to cast a split ticket are diverse. Voters may
counties with large Hispanic populations. Similarly, both have incentives to split their ballot in an effort to achieve
Calvo et al. (2008) and Katz et al. (2011) report that the in- policy/ideological balancing (Alesina and Rosenthal 1995;
formational cues offered by different ballot structures favor Brunell and Grofman 2009; Shugart 1995), to engage in stra-
some parties to the detriment of others. Ballots that reinforce tegic voting (Cox 1997; Moser and Scheiner 2009), because

This content downloaded from 132.239.001.230 on February 02, 2017 09:21:58 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
000 / Ballot Structure and Split Ticket Voting Tiffany D. Barnes, Carolina Tchintian, and Santiago Alles

they are attracted to an individual candidate’s personal qualities tronic ballot (like the ones adopted in Brazil, Estonia, the
(Beck et al. 1992), or even because they are indifferent between Philippines, India, the United States, and Venezuela) dis-
political parties (Campbell and Miller 1957). Despite these playing all candidates/parties on the same ballot limit the
incentives, the electoral benefits of split ticket voting are not number of steps voters must take to split their ballot. Voters
always clear. Thus, not all voters who prefer to split their vote only need to click a button or check a box to choose an ex-
are willing to pay the costs of doing so. In particular, we ar- ecutive candidate and a legislative candidate from different
gue that some ballot structures can impose unnecessary costs parties. We anticipate that more user-friendly ballot struc-
on voters, making it taxing for them to split their ballot. tures that allow voters to view the ballot organized by office
Ballot structures vary substantially in terms of the lev- substantially reduce the cost of ballot splitting by making
els of difficulty, that is, the cognitive and physical demands it convenient, fast, and easy as compared to the ballot-and-
they place on voters. Some ballots require voters to select or envelope system. To evaluate this proposition, we test the
punch multiple boxes, to cut or tear the ballot, or to use com- following hypothesis:
plex machines that demand considerable attention. Such var-
iation in ballot structure affects voters’ behavior at the polls, H1. The average level of ballot splitting will be lower
making them more or less likely to cast a split ticket vote. Thus, with the ballot-and-envelope system than with the
even when voters prefer to split their ticket, if it is uncertain Australian ballot.
how it will affect the electoral outcome, the costs imposed by
complicated ballot structures may be sufficient to discourage THE VARYING IMPACT OF BALLOT STRUCTURE
voters from doing so. By contrast, other ballot structures can ON SPLIT-TICKET VOTING ACROSS PARTIES
streamline the ticket splitting process, making it easy for mo- Although we anticipate that the Australian ballot structure
tivated voters to split their vote. When voters can split their should, on average, increase ballot splitting as compared to
ballot without incurring any additional costs, they are more the ballot-and-envelope system, we should not expect a uni-
likely to do so, even if they do not anticipate that it will result form effect across all political parties. To understand how
in clear electoral payoffs. changes in the ballot structure will influence ballot split-
Previous research from the United States provides some ting differently across political parties, it is important to take
evidence to support the idea that the ballot structure shapes into consideration voters’ incentives for splitting their bal-
voters’ predisposition to split their vote (Rusk 1970). When lot. Regardless of voters’ motivations to split their ballot (e.g.,
ballots do not offer a straight party option, and thus require policy/ideological balancing, attraction to individual can-
voters to make separate decisions for each contest, splitting didates, indifference toward parties, or strategic voting), the
one’s vote is no more costly than voting straight ticket. physical characteristics of the electoral ballot could explain
Consequently, voters are more likely to split their ballot some of the overall increase or decrease in the proportion
when they are not presented with a straight ticket option of split ticket voting (Campbell et al. 1960). Only strategic
(Campbell et al. 1960; Campbell and Miller 1957; Darcy and voting theory, however, offers insight about which candi-
Schneider 1989; Reynolds and McCormick 1986; Walker dates or parties will see the largest uptick in vote splitting
1966). resulting from different ballot structures.
The ballot-and-envelope system is a good example of how The strategic voting explanation relies on voters’ calcula-
the ballot structure can discourage split ticket voting. The de- tions, arguing that voters refrain from voting for their most
fault is straight ticket voting. To vote straight ticket all voters preferred candidate and instead vote for their most pre-
need to do is select a ballot, put it in an envelope, and place it ferred viable candidate to avoid wasting their vote (Cox
in the ballot box. Split ticket voting requires voters to man- 1997). In concurrent elections, candidates compete under
ually tear a paper ballot in order to split their vote. This is the same party label, but often each race is governed by a
inconvenient, complicated, and requires extra effort and time different set of electoral rules or structured by a different field
in the ballot box. For this reason, we argue that the ballot- of political competition. Consequently, voters face different
and-envelope system discourages split ticket voting, partic- incentives for each candidate selection.
ularly when the electoral payoffs are unclear. In concurrent executive and legislative elections, for ex-
By contrast, streamlined voting procedures such as the ample, the executive candidates compete for a single seat.
Australian ballot simplify the vote splitting process. For Voters may, therefore, be compelled to vote strategically be-
example, a single paper ballot (such as the ones recently cause there are fewer viable candidates. By contrast, in multi-
adopted by Benin, Guinea, Niger, and Tunisia) or an elec- member legislative elections, as the number of seats at stake

This content downloaded from 132.239.001.230 on February 02, 2017 09:21:58 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 Volume 79 Number 2 April 2017 / 000

grows, there are a larger number of viable candidates, and vot- et al. 2011). The governor and the Electoral Tribunal jointly
ers have stronger incentives to cast a sincere vote for their decided to incrementally introduce the new voting pro-
most preferred candidate (Moser and Scheiner 2009). More- cedure to half of the electorate in the legislative district of
over, even when the different races are governed by the same Salta.2 The partial implementation provides a rare oppor-
electoral rules, candidates from the same political party may tunity to evaluate the relationship between voting proce-
be more or less viable for elections at different levels of gov- dures and ballot splitting using a quasi-experimental de-
ernment. When there is more certainty about the electoral sign. We exploit this opportunity by analyzing data from
payoff of split ticket voting, voters are more willing to pay the 2007 (no implementation), 2011 (partial implementa-
the cost of splitting their ballot even when doing so is incon- tion), and 2015 (full implementation) concurrent guber-
venient, as with the ballot-and-envelope system. The strate- natorial and provincial legislative elections.
gic voting literature thus implies that overall levels of ballot We leverage precinct-level observational data from the
splitting will be higher among the most popular/viable exec- district of Salta—the only legislative district in which half
utive candidates than for the less competitive executive can- of the voters used electronic ballots and half used paper
didates regardless of the ballot structure. ballots—to evaluate how different ballot structures facili-
It is our contention that ballot structures that facilitate tate ballot splitting.3 The unit of analysis is the precinct
ticket splitting, such as the Australian ballot (as compared because this is the unit of assignment for the treatment
to the ballot-and-envelope system), increase split ticket condition. The precinct-level analysis combined with par-
voting by lowering the cost of ballot splitting such that the tial implementation allows us to examine different ballot
benefit outweighs the cost. Thus, in thinking about how structures applied in the exact same district, in the exact
ballot structure may cause a larger increase in ballot split- same election, and under the same electoral rules.4 As the
ting for some parties than others, the strategic voting lit- effects of ballot reforms are conditional on other cultural,
erature implies that an uptick in ballot splitting may be institutional, and political factors (e.g., the electoral rules or
larger for smaller parties where the payoff for split ticket the number of parties competing in the election), the pri-
voting is unclear. As voters are already inclined to split mary advantage of a subnational analysis is that it facilitates
their ticket when there are clear electoral payoffs, lowering a controlled comparison across precincts where other po-
the cost of ballot splitting may not substantially increase tentially confounding factors are held constant. In a cross-
the level of split ticket voting in such circumstances. By national setting, by contrast, institutional and political fac-
contrast, a simplified ballot structure may produce an up- tors vary substantially, making it more difficult to isolate
tick in split tickets when the payoff for ticket splitting is the effect of ballot reform. Although this variation presents
ambiguous. That is, if a voter is uncertain how much ticket a host of research design challenges, the subnational setting
splitting will pay off, she may be less likely to pay the costs with partial implementation in the same election makes it
of undertaking the laborious procedure of splitting their possible to identify the causal effects of ballot structures—a
paper ballot. But, when ticket splitting is costless, voters will comparison that is virtually impossible in a cross-national
be more likely to do so, even if the payoff is uncertain. Given analysis.
this, in the executive-legislative context, we expect that the In the quasi-experimental research design, the “control
Australian ballot is most likely to result in an uptick in split group” comprises precincts using the ballot-and-envelope
ballots among smaller parties with nonviable executive can- system in the 2011 election. Precincts using the e-voting
didates and/or those with a lower probability of winning a device in the 2011 election comprise the “treatment group.”
legislative seat.

H2. The adoption of the Australian ballot will lead to 2. In 2011, a total of 33% of the voters across the entire province used
the electronic voting system for the first time. In the following two elec-
a greater increase in ballot splitting among smaller
tions (2013 midterm and the 2015 gubernatorial elections), e-voting was
political parties. used across the entire province.
3. There are 54 precincts in the district of Salta. Each includes many
PARTIAL IMPLEMENTATION OF ELECTRONIC VOTING polling stations (i.e., schools) and multiple voting booths within each
IN SALTA: A QUASI-EXPERIMENT station. Voters within precincts are assigned to voting booths according to
the first letter of their last name (see table B2 in appendix B).
In 2011, e-voting was implemented in Salta, Argentina,
4. The governor is elected by plurality rule, and deputies in the district
replacing the ballot-and-envelope system that was ubiqui- of Salta are elected from a closed list proportional representation system
tous across the entire country (Alvarez et al. 2013; Pomares with a district magnitude of 9.

This content downloaded from 132.239.001.230 on February 02, 2017 09:21:58 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
000 / Ballot Structure and Split Ticket Voting Tiffany D. Barnes, Carolina Tchintian, and Santiago Alles

We briefly explain the voting procedures and ballot struc- lope, and discarding the unwanted pieces of the ballots. If
tures for the two types of ballots to illustrate how they in- the voter makes an error when casting her ballot, such as
fluence ballot splitting. tearing the ballot, including more than one ballot for the
same office, or not including a ballot for some category at
Control condition: Ballot-and-envelope system stake, she wastes her vote by either spoiling her vote or by
Before 2011, the same ballot-and-envelope system was used voting blank for the omitted category.
in virtually every Argentine election across the entire coun-
try. Under this system, each party is represented on a separate Treatment condition: Electronic ballot
ballot, and every party is responsible for printing and dis- The voting procedure under the electronic ballot is very
tributing their ballots to every voting center. Figure 1 shows different and considerably more user-friendly. The voter
an example of the paper ballot from 2011 from the district first approaches a touch screen machine. Then she chooses
of Salta. The paper ballot is a single piece of paper, with the how the candidates for each position at stake will be dis-
party label, the party/coalition registration number, and played on the screen. Candidates can be displayed in one of
sometimes a picture of the candidate. It contains the party’s two ways: (1) by political party or (2) by elected positions.
candidates for all the positions at stake in a given district, If voters choose candidates to be displayed by political party,
divided by a printed line. In the voting booth, the voter finds the e-voting device will first show all gubernatorial can-
the ballots from all the contending parties, sorted by their didates competing in that election. After the voter chooses
registration number. The voter selects the ballot of her choice, the gubernatorial candidate of her choice, the device will
folds it, places it into an envelope provided by the Electoral display the other lists and candidates associated with the
Tribunal, and slips it into the ballot box. gubernatorial candidate. In Argentina, the top gubernatorial
If the voter wants to split her ballot, she is tasked with candidates frequently run with the support of other smaller
a more laborious and risky process. First, the voter must parties that choose not to field their own gubernatorial
gather the different ballots of her preferred parties for each candidates. For those candidates competing with several leg-
office at stake. Then she must physically “split” the ballot islative lists, the electronic device will show a second screen
by cutting out the elected position she supports from each in which the executive ticket is represented on the screen mul-
party, placing all the preferred ballot pieces in the enve- tiple times, once for each of the lists affiliated with the can-

Figure 1. Demo paper ballots from the 2011 election in the district of Salta, Argentina. Each one-third of the ballot corresponds to a different elected position.
Voters choosing to split their ballot can divide their ballot accordingly.

This content downloaded from 132.239.001.230 on February 02, 2017 09:21:58 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 Volume 79 Number 2 April 2017 / 000

didate. Finally, voters may cast a straight ticket ballot by Geo-referenced data
confirming and selecting this option on the screen. By con- Although electoral data are available at the precinct level,
trast, if the voter chooses candidates to be displayed by elected sociodemographic variables are only available at the census-
position, she views all the candidates for each position at tract level, and census tracts do not map onto precincts. The
once, and selects the party/candidate she supports for each lack of information on precinct boundaries presents a major
election at stake. Compared to the ballot-and-envelope sys- challenge for identifying the SES characteristics of precincts.
tem, this option makes it simple to cast a split ticket, while it To address this, we built an original data set using ArcGIS
assures that ballots for each of the political parties at stake are to recreate precinct boundaries based on the location of vot-
always available at every polling station. ing centers. We map this information onto existing census
tract boundaries to identify precinct-level demographics. As
Threats to random assignment the shape of the precincts is unknown, we use two differ-
Despite the partial implementation of e-voting, the treat- ent GIS techniques to approximate the precinct boundaries,
ment assignment criteria are not random. Indeed, treatments allowing us to examine the sensitivity of our results to the dif-
in quasi-experiments do not involve an explicit random as- ferent procedures (see appendix A for details; appendices A–F
signment, which is the primary feature of our research design are available online). The results are remarkably robust to the
that distinguishes it from a natural experiment. We identified two GIS techniques; thus, we report results from the Thiessen
the threats to random assignment, and we use this informa- polygon technique in the text and results from the nearest
tion to match precincts from the treatment group and the neighbor technique in appendices D and E.
control group to improve our ability to make inferences.
To begin with, bureaucrats and politicians may have had Matching methods
incentives to pilot e-voting in some precincts rather than Matching methods attempt to construct a data set where
others. Elite interviews with the bureaucrats who organized background conditions in the sample are equal across the
the reform report that partial implementation was chosen treatment and control groups, operating as a pre-processing
as a means to assure a successful execution and to avoid any step before the parametric analysis. This approach allows us
resistance to the reform (Alvarez et al. 2013; ONPE 2014; to account for the potential confounding effects between
Pomares, Levin, and Alvarez 2014a). Thus they initiated the treatment and the outcome, and it reduces the proba-
e-voting in places with more sophisticated voters, that is, in bility that the results are model dependent (King and Zeng
precincts where socioeconomic status (SES), education, and 2006). We employ two alternative matching methods to
technological sophistication are higher (Alvarez et al. 2013; reconstruct the balance in the sample: Propensity Scores (PS)
Pomares and Zárate 2014). Moreover, given that the in- matching without replacement (Gelman and Hill 2007; Ho
cumbent governor spearheaded the decision to implement et al. 2007) and Coarsened Exact (CE) matching (Iacus, King,
e-voting and to do so incrementally, it is possible that some and Porro 2011, 2012).
political calculations guided the treatment assignment. The Quantile-quantile plots (Q-Q plots) used to assess the
incumbent’s performance in the district of Salta in the pre- post-matching balance improvement indicate that the bal-
vious election was 5% below the province average, and his ance improved substantially for two of the three background
support was distributed relatively evenly across the precincts confounders (i.e., poverty and education; see appendix C for
in the district of Salta. Q-Q plots and summary statistics). Balance does not improve
To address bureaucratic concerns, we match precincts for the opposition vote share, as it was already well balanced
on the share of families that live below the poverty line and prior to matching. The CE matching produces the larger bal-
on the share of the population with elementary education ance improvement. Of the 54 precincts in the district of Salta,
or less (i.e., the first seven years of formal education) as a the treatment and control subsamples comprise 14 and 19 ob-
proxy of voters’ sophistication. To address the potential servations, respectively, for the Thiessen polygon technique,
political motivations, we match precincts on the percentage and 14 and 20, respectively, for the nearest neighbor tech-
of votes the runner up (i.e., the opposition candidate) won nique. In both samples, the balance for poverty and educa-
in the 2007 gubernatorial election—the election prior to the tion improves by at least 75%. The PS matching produces the
assignment of e-voting.5

tronic devices were somewhat more likely in precincts with more educated
voters and with less support for the runner up governor in the previous
5. The power of these variables was modeled to predict the treatment gubernatorial election. These findings also suggest how suitable the pool of
assignment, and the results suggest that there is some association: elec- variables used to match the data was. See table B1 in appendix B

This content downloaded from 132.239.001.230 on February 02, 2017 09:21:58 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
000 / Ballot Structure and Split Ticket Voting Tiffany D. Barnes, Carolina Tchintian, and Santiago Alles

poorest post-matching balance in the data. The treatment of split ballots was higher in those precincts using e-voting
and control subsamples comprise 25 and 9 observations, re- than in those precincts using the ballot-and-envelope system.
spectively, for the Thiessen polygon technique, and 25 and In fact, in the vast majority of e-voting precincts, the share
10, respectively, for the nearest neighbor technique, and none of split ballots is greater than in paper ballot precincts. Out
of the variables improved by more than 55%. Typically there of 14 matched precincts using e-voting, 13 have the high-
is a trade-off between sample size and post-matching balance est share of split ballots in the 2011 sample. Whereas there
(King, Lucas, and Nielsen 2014), but in the case of our data, is a clear jump from 6 to 10 points in the share of split bal-
CE matching produces better balance and a sample size that lots between the control and treatment precincts in 2011,
compares favorably to the one produced from PS match- there is no clear pattern between levels of ballot splitting and
ing. For this reason, and because the results are robust to both voting systems in 2007 (no implementation) or 2015 (full
matching procedures, we report results using CE matching implementation) elections.
in the text and those from the PS matching in Appendices D
and E.
Vote splitting at the precinct-party level
Our second dependent variable calculates the share of split
MEASURING SPLIT TICKET VOTING tickets at the precinct-party level for each party fielding a
Our theory posits that reforms to the ballot structure intro- gubernatorial candidate. Here the unit of observation is the
duced by e-voting should increase the overall level of ballot precinct-party. We calculate yij to reflect ballot splitting
splitting (hypothesis 1). Yet, we should not expect to see uni- for every political party j that fielded a gubernatorial can-
form increases in ballot splitting across all political parties didate in precinct i. We measure the difference between the
(hypothesis 2). In this section, we explain how we measure number of votes obtained by the gubernatorial candidate,
each of our two dependent variables to test our precinct- and Gij, and the number of votes his legislative lists won, Dij, di-
party-level hypotheses. vided by the total number of votes obtained by each party, Vij.
This gives us our precinct-party level dependent variable:
Vote splitting at the precinct level Splitij p (Gij  Dij =V ij ). Positive (negative) values indicate
Our first dependent variable measures the overall share of that the gubernatorial candidate won more (less) votes than
split tickets within the precinct. Thus, the unit of analy- the legislative list.
sis is the precinct. To calculate the share of split ballots When gubernatorial candidates are featured on multi-
within the precinct, yi, we take the absolute value of the ple ballots with different legislative lists, voters can technically
difference in the gubernatorial candidate’s votes, Gij, com- vote for the gubernatorial candidate and a legislative list from
pared to the legislative ticket’s votes, Dij, for each party j a different party without physically splitting their ballot. In
competing in each precinct i.6 We sum across all parties in our analysis, we define split ticket voting as the cases in which
precinct i, divide by two, and weigh the value as a share of voters using the ballot-and-envelope system physically split
the total valid votes, Vi, in the precinct.7 This gives us the their ballot (and not those cases where voters can technically
share of split ticket votes in a given precinct: Spliti p vote for two different parties without splitting their ballot).


^jGij 2 Dij j=2 =V i . When more than one list is attached to the gubernatorial
Figure 2 illustrates the trends in ballot splitting across ticket, we calculate the number of votes the legislative lists won
precincts and over time. It represents each precinct in our by summing the votes for all of the lists affiliated with a sin-
matched sample on the x-axis and the share of split ballots gle gubernatorial candidate. The total number of votes is de-
(i.e., the dependent variable) on the y-axis for each year in fined by the maximum number of votes obtained by the party
our analysis. The dark bars indicate precincts using e-voting, in any of the two categories.
and the light bars denote those using paper ballots. We can
see immediately from figure 2 that in 2011 the overall share
Potential data limitations
Scholars studying ballot splitting routinely face a host of
6. When the gubernatorial candidate runs with multiple legislative methodological challenges (Burden and Helmke 2009). Al-
lists, we take the difference between the votes obtained by the guberna- though we are able to overcome some very difficult research
torial candidate and the sum of votes obtained by the different lists at-
tached to his ticket.
design problems using a quasi-experimental approach, data
7. Transforming the dependent variable into a ratio makes it compa- availability still poses potential limitations for our analyses.
rable across electoral precincts regardless of the number of registered voters. One limitation of our precinct-party level variable is that the

This content downloaded from 132.239.001.230 on February 02, 2017 09:21:58 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 Figure 2. Share of split ballots, by precinct and type of ballot: A, 2007 election, no implementation; B, 2011 election, partial implementation; C, 2015 election, total implementation

This content downloaded from 132.239.001.230 on February 02, 2017 09:21:58 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
000 / Ballot Structure and Split Ticket Voting Tiffany D. Barnes, Carolina Tchintian, and Santiago Alles

Electoral Tribunal only reports the overall number of votes been a problem in Salta.8 In theory, ballot splitting may be
each gubernatorial candidate won and the overall number lower under the ballot-and-envelope system if smaller parties
of votes each of the legislative lists won. The tribunal does cannot consistently supply ballots. As the number of parties
not distinguish between the votes gubernatorial candidates competing in each contest decreases, the probability that a
won from each of the different legislative lists. We cannot, voter randomly choosing between parties for each contest at
therefore, empirically distinguish among the votes guber- stake splits his/her ticket also decreases (Ames et al. 2009).
natorial candidates earned when competing with different Consequently, if ballot supply is a problem, the Australian
legislative tickets. An empirical implication of this data ballot will be associated with a higher level of overall voting
limitation is that it provides a conservative estimate of how and ballot splitting for smaller parties than the ballot-and-
frequently voters support a leading gubernatorial candidate envelope system, simply by assuring that all options are avail-
without supporting the candidate’s party. As such, it may able at the voting booth. Although ballot availability may
decrease the overall level of ballot splitting we observe at the have been a problem in some precincts, additional analyses
precinct level and in particular among parties that field of precincts using paper ballots in our matched sample indi-
gubernatorial candidates with multiple supporting legisla- cate that this was not problematic for our results (see ap-
tive party lists at the precinct-party level. Nonetheless, be- pendix F).
cause the same parties compete with multiple lists attached
to their gubernatorial ticket across all of the precincts in our PRECINCT-LEVEL ANALYSES AND RESULTS:
analysis, and this does not vary across the treatment and DIFFERENCE-IN-DIFFERENCE TESTS
control groups, it does not limit our ability to observe the We use a standard difference-in-difference approach to ex-
causal impact of the ballot reform on split ticket voting. amine the causal relationship between e-voting and ballot
Importantly, our results are robust to the exclusion of par- splitting using the matched sample (hypothesis 1). The top
ties that run with multiple list attached to the gubernatorial panel of table 1 shows the results from the difference-in-
ticket (see appendix F). difference analysis comparing the 2007 election (no e-voting)
A second limitation of our data is that in precincts using to the 2011 election (partial implementation). Before the im-
the ballot-and-envelope system, it is clear to voters which plementation of e-voting, the average share of split ballots
legislative lists are associated with which gubernatorial can- between the executive and legislative election was 8.52 in the
didate. By contrast, in the e-voting system, this same infor- control group in 2007, and 9.26 in the treatment group. The
mation is not presented on the ballot if the voter chooses to difference is not statistically significant. In 2011, the difference
view the ballot by elected position. If voters in precincts with is substantially larger. Whereas the share of vote splitting was
e-voting do not have information regarding which legislative only 5.23 in precincts with paper ballots, the share of split bal-
lists support which gubernatorial candidates, this could in- lots increased almost twofold in districts with e-voting to 10.6.
crease ballot splitting in precincts with e-voting due to a lack The difference-in-difference approach identifies the causal
of information. For instance, voters may want to vote for one effect of e-voting, which corresponds to a statistically signif-
of the parties that supports their favored gubernatorial can- icant increase of 4.64. The difference-in-difference compar-
didate, but they may erroneously choose a legislative list that ison between 2007 and 2011 provides systematic support for
does not support their favored candidate. That said, this sce- our hypothesis, as it demonstrates that precincts using e-voting
nario would only increase ballot splitting for the political parties with the option to view the Australian ballot structure had
that run with multiple legislative lists (in 2011 the FREJUREVI higher levels of ballot splitting than precincts using the ballot-
and Frente Olmedo), thus increasing ballot splitting for parties and-envelope system.
with viable gubernatorial candidates and reducing the prob- The bottom panel in table 1 provides further evidence in
ability that we find support for hypothesis 2. By contrast, parties support of hypothesis 1. Here we compare the partial im-
that run with only one legislative list attached to their gu- plementation of e-voting in 2011 to the full implementation
bernatorial candidate include the party name, the logo, and the in 2015. Whereas the partial implementation resulted in a
registration number on both the gubernatorial and legislative two-fold increase in ballot splitting between the treatment
portions of the ballot, providing a cue for those voters who group (precincts with e-voting) as compared to the control
want to vote straight ticket. group (precincts with paper ballots) in 2011, once e-voting
Finally, as noted, under the ballot-and-envelope system
parties are responsible for supplying their own ballots (Po- 8. See, e.g., reports from La Nación, “En Salta se estrena el voto electró-
mares et al. 2014b). This sometimes presents a challenge for nico,” April 5, 2011, http://www.lanacion.com.ar/1363033 (accessed May 7,
smaller parties; in fact, news reports suggest that this has 2016).

This content downloaded from 132.239.001.230 on February 02, 2017 09:21:58 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 Volume 79 Number 2 April 2017 / 000

Table 1. Difference-in-Difference Analyses

Control Group Treated Group

(Paper Ballots in 2011) (Electronic Ballots in 2011) Difference

No Implementation (2007) versus Partial Implementation (2011)

2007 elections (no implementation) 8.5198 9.2592 2.7394

2011 elections (partial Implementation) 5.2254 10.6010 25.3756***
Difference 3.2944*** 21.3418 4.6362***

Partial Implementation (2011) versus Fulll Implementation (2015)

2011 elections (partial implementation) 5.2254 10.6010 25.3756***

2015 elections (full implementation) 7.8052 7.4684 .3368
Difference 22.5798*** 3.1326*** 25.7124***

Note. Cells with plain text indicate the average level of vote splitting (the dependent variable) for each group, the bolded cells
indicate the difference between two groups, and the bolded and italicized cells indicate the difference-in-difference. Results
presented in this table were calculated based on the Thiessen Polygon Geographic Information Systems (GIS) technique and the
Coarsened Exact (CE) matching procedure. Results based on the alternative GIS technique and matching procedure appear in
appendix D.
* p ! .05.
** p ! .01.
*** p ! .001.

is implemented in all precincts (2015), the share of split bal- voting is slightly lower among treated groups in 2015 than
lots is statistically indistinguishable between the treatment in 2011.
and control groups. Moreover, the difference-in-difference
results show a statistically significant increase of 5.71, offer- PRECINCT-PARTY LEVEL ANALYSES AND RESULTS
ing further evidence that the new ballot structure introduced It is clear from the difference-in-difference analyses that the
by e-voting increased ballot splitting. new ballot structure is associated with significant increases
Still, it is important to note that even though e-voting in the average level of ballot splitting. Nonetheless, we an-
is associated with an increase in split ticket voting within ticipate that the new ballot structure is more likely to result
the exact same election, the overall levels of split ticket vot- in an increased level of ballot splitting for some political par-
ing decreased between the 2007 (no implementation) and ties than for others (hypothesis 2). In this section, we use our
2015 (full implementation) elections. To understand the dif- precinct-party level measure of vote splitting to investigate
ference in the overall level of split ticket voting between which parties experienced an uptick in split tickets after the
different elections, it is necessary to consider the larger po- reform.
litical context. In the case of Salta, split ticket voting is likely Using the matched data, we estimated a linear model
higher in 2007 because the gubernatorial election was more with clustered standard errors (on precincts) to account for
competitive and a larger number of political parties fielded the correlation between observations in the same precincts.
gubernatorial candidates that year. Indeed, electoral compe- We include a dummy variable to distinguish between the
tition and multipartism are both associated with increases treatment and control groups. It is coded 1 for the treatment
in split ticket voting (Ames et al. 2009; Burden and Kim- group (i.e., e-voting in 2011) and 0 for the control group (i.e.,
ball 1998). Whereas the governor only won the 2007 elec- paper ballots in 2011). We include a dummy variable for each
tion by 1.07% of the vote, the winner of the gubernatorial of the party lists fielding a gubernatorial candidate in the elec-
elections in both 2011 and 2015 secured at least 20% more tion and an interaction between the party dummy variables
votes than the runner-up. Further, 10 political parties fielded and the e-voting variable to analyze which party ballots were
gubernatorial candidates in 2007, compared to only eight most affected by the implementation of e-voting. We analyze
in 2011 and five in 2015. The larger number of parties field- three models, one for each election year. The results for these
ing gubernatorial candidates also explains why split ticket analyses are reported in table 2.

This content downloaded from 132.239.001.230 on February 02, 2017 09:21:58 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
000 / Ballot Structure and Split Ticket Voting Tiffany D. Barnes, Carolina Tchintian, and Santiago Alles

To determine which parties saw an uptick in ballot for splitting one’s ticket to vote for Urtubey but not his
splitting after the ballot reform, we calculate the marginal legislative list is clear. Thus, we argue that voters should be
effect of e-voting on ballot splitting for each party in our more willing to incur the cost of split ticket voting regard-
analyses. Recall that all coefficients must be interpreted rel- less of the ballot structure, and lowering cost of split ticket
ative to the baseline category and cannot be interpreted voting via ballot reform should not, therefore, substantially
based only on the size and direction of the coefficient. The increase the level of split ticket voting for the FREJUREVI.
marginal effects are calculated from the least-squares re- Technically, Alfredo Olmedo was the runner-up, so we
gression models, and the confidence area of the estimations may anticipate the same logic would apply to his election.
are based on 50,000 simulated independent draws (Gelman That said it is not clear how “viable” Olmedo was in the
et al. 2015; Gelman and Hill 2007). election. He earned almost 35% less votes than the top gu-
Figure 3 reports the marginal effect of e-voting on ballot bernatorial candidate and immediately prior to the election
splitting by political party. A value of zero (denoted by the was polling third. Pollsters predicted he would capture only
horizontal reference line) indicates that ballot splitting was 17.8% of the vote compared to Walter Wayar, who was
no more likely in precincts with e-voting than in precincts predicted to win 17.9% of the vote.9 Given that Olmedo and
with a paper ballot for a given political party. Positive values Wayar were both trailing far behind the front-runner, the
show that e-voting is associated with an increase in the share electoral payoffs for voting for either of these candidates
of split ballots in favor of the gubernatorial candidate—that is unclear. Our argument implies that when the electoral
is, the executive ticket won more votes than the party’s leg- payoffs of split ticket voting are unclear, lowering the cost
islative ticket. Hence, a marginal effect of .25 represents a of ballot splitting (by introducing a ballot structure that fa-
25% increase in ballot splitting in favor of the gubernatorial cilitates split ticket voting) will lead to an increase in ballot
ticket in e-voting precincts, with respect to the average pa- splitting among those voters who prefer to split their ticket.
per ballot precinct. By contrast, negative values indicate an As a matter of fact, there was a statistically significant
increase in split ballots in favor of the legislative list. increase in the share of split tickets for both Olmedo and
The left panel in figure 3 plots the marginal effect of the Wayar, and the size of the effect was substantial. Wayar
treatment group by party in 2007 when all precincts used the won 24.5% more votes than his party’s legislative list
ballot-and-envelope system, and the panel on the right plots among voters using the paper ballot. That is, for every 100
the results from 2015 when all precincts used e-voting. As legislative votes his party won, he obtained about 125 votes
with the previous analysis, these panels show that there are (a party coefficient of 0.393 and 20.147 for the equation
no significant differences between the treatment and control intercept). Yet our model estimated that the share of split
groups when all voters use the same ballot structure. The votes was 23.7% larger in precincts using e-voting than in
marginal effect of e-voting is close to zero for each political precincts with paper ballots (a coefficient of 0.340 for the
party, and in all cases, the confidence intervals cross zero. party interaction term and 20.103 for the intercept on e-
These results provide further evidence that the differences voting). Thus, for every 100 legislative votes his party won,
observed in the 2011 analysis are not simply a product of he obtained about 148 votes in precincts with e-voting (i.e.,
underling differences in the treatment and control groups. 24.5% 1 23.7%). Put differently, in districts with e-voting,
Next, turning to the 2011 election when e-voting was Wayar won almost 50% more votes than his party list.
partially implemented, the center panel in figure 3 shows Olmedo, by contrast, won fewer votes than did his tan-
that precincts using e-voting have a significantly higher share dem legislative lists. This does not come as a surprise given
of split ballots than do those using paper ballots. As anti- that Olmedo was embroiled in a scandal that broke just
cipated this pattern varies across political parties, lending before the election and received substantial news cover-
some support for hypothesis 2, that an increase in ballot age.10 In paper ballot precincts, he won only 91% of the votes
splitting will be most likely among smaller political parties
with more uncertainty surrounding their electoral fates.
To begin with, in 2011, there was only one clear front- 9. The polls were reported in Infobe, “Salta: las últimas encuestas otor-
runner. Juan Manuel Urtubey from the FREJUREVI won gaban una cómoda victoria para Urtubey,” April 10, 2011, http://www.info
60% of the popular vote and was polling above 50% prior bae.com/c574779 (accessed May 7, 2016).
to the election. The closest competitor was only poling at 10. The story was widely covered by national publications (i.e., Ámbito
Financiero or La Razón) as well as leading provincial newspapers (i.e.,
18%. According to the strategic voting literature, for vot- Los Andes). See, e.g., Ámbito Financiero: “Olmedo justifica trabajo en negro
ers who prefer Urtubey to other potentially electable can- en su campo por los planes sociales,” March 17, 2011, http://www.ambito.com
didates but do not truly prefer his party, the electoral payoff /573245 (accessed May 7, 2016).

This content downloaded from 132.239.001.230 on February 02, 2017 09:21:58 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
Table 2. The Effect of Ballot Structure on the Share of Split Ballots, by Party

2007 Elections: No Implementation 2011 Elections: Partial Implementation 2015 Elections: Full Implementation

Electronic device (ED) .0096 Electronic device (ED) 2.1025** Electronic device (ED) 2.0481
(.0169) (.038 (.0338)
PH 2.2766*** PO .0741*** Frente Popular 2.036
(.0323) (.0158) (.0413)
MST 2.3333*** Wayar .3927*** PO 2.2218***
(.0339) (.0232) (.0245)
Mov. Pop. Salteño 2.2365*** CC-ARI .0693* UCR-UNEN-PS .3289***
(.0569) (.0288) (.0236)
PO 2.4124*** MIJD .0899 FREJUREVI 2.1449***
(.0212) (.0536) (.0268)
Nueva Generación 2.2462*** SUR-PS .3091*** Romero 1 Olmedo .0834**
(.0481) (.0186) (.0261)
Propuesta Salteña 2.2786*** FREJUREVI .2266***
(.0238) (.0197)
FPV-PRS .0723*** Olmedo .0615***
(.0061) (.0175)
Enc. Pop. Amplio 2.2791***
(.0301)
UCR 2.3217***
(.0357)
ED # PH 2.0265 ED # PO .0295 ED # Frente Popular .0698
(.0580) (.0448) (.0528)
ED # MST 2.0757 ED # Wayar .3397*** ED # PO .0398
(.0745) (.0506) (.0325)
ED # Mov. P. Salteño .0223 ED # CC-ARI .072 ED # UCR-UNEN-PS .0686
(.1247) (.0637) (.0353)
ED # PO 2.0009 ED # MIJD .0998 ED # FREJUREVI .0668
(.0450) (.1093) (.0343)
ED # Nva. Generación .0373 ED # SUR-PS .3413*** ED # Rom. 1 Olmedo .0162
(.0551) (.0525) (.0390)
ED # P. Salteña 2.0184 ED # FREJUREVI .1186**
(.0324) (.0377)
ED # FPV-PRS .0132 ED # Olmedo 2.0546
(.0231) (.0381)
ED # E. P. Amplio 2.0689
(.0551)
ED # UCR 2.0095
(.0467)
Intercept .1203*** Intercept 2.1474*** Intercept .0292
(.0122) (.0161) (.0277)
Observations 329 Observations 264 Observations 198
Adjusted R2 .5267 R2 .7518 R2 .8720

Note. Coefficients from ordinary least squares regression with clustered standard errors by precinct (in parentheses). Results in table 2 are based on the
Thiessen Polygon Geographic Information Systems (GIS) technique and the Coarsened Exact (CE) matching procedure. Results based on the alternative GIS
technique and matching procedure appear in appendix E.
* p ! .05.
** p ! .01.
*** p ! .001.

This content downloaded from 132.239.001.230 on February 02, 2017 09:21:58 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
000 / Ballot Structure and Split Ticket Voting Tiffany D. Barnes, Carolina Tchintian, and Santiago Alles

Figure 3. Marginal effect of e-voting on the share of split ballots, by party. This figure plots the marginal effect of belonging to the treatment group (i.e.,
assignment to electronic voting in 2011) on ballot splitting for each party competing in the 2007 (no implementation), 2011 (partial implementation), and
2015 (full implementation) elections relative to the control group (i.e., assignment to paper ballots in the 2011 election). Values different from zero indicate
that there was more ballot splitting in precincts with electronic voting than in precincts with paper ballots. Positive values indicate that the governor’s party
received more votes than the legislative party, and negative values indicate the legislative party received more votes than the governor’s ticket. Marginal
effects are based on results presented in table 2.

his legislative lists won: for every 100 legislative votes, he Together the results from the 2011 analysis indicate not
only won 91 votes (a party coefficient of 0.062 and 0.147 for only that ballot splitting is more likely under e-voting than
the equation intercept). In precincts with e-voting, the split under the traditional paper ballot but also that the substan-
punished Olmedo by an extra 15.7% (a baseline for the tive results are larger for parties with unclear electoral fates.
party interaction term of 20.055 and 20.103 for the inter- Although this analysis can provide some insights into how
cept on e-voting). Thus, for every 76 votes the gubernatorial parties are differentially affected by changes in the ballot
candidate won, his legislative lists won 100 in districts with structure, as we can only examine partial implementation of
e-voting. ballot reform for one legislative election, we are limited in
Additionally, we observe an uptick in split tickets for our ability to offer more generalizable conclusions. Future
other small parties with uncertain electoral fates: the SUR-PS, research should further consider the differential effects of
the PO, and the UCR. Only the two smallest parties (MIJD ballot reforms across different political parties.
and CC-ARI), both winning less than 1% of the overall vote In all, the findings from the 2011 analysis of partial im-
share, did not see an uptick in split tickets. Perhaps lowering plementation of e-voting demonstrate that ballot structure
the cost of split ticket voting did not translate into an increase has a significant effect on ballot splitting. Voters in precincts
in split tickets for the MIJD and the CC-ARI because there with the option to view the Australian ballot structure are
was very little uncertainty surrounding the electoral payoffs. significantly more likely to split their ballot—particularly
Both parties were extremely unlikely to secure enough votes among political parties fielding nonviable gubernatorial can-
to win a seat in the legislature. didates or those with a lower probability of winning a legis-

This content downloaded from 132.239.001.230 on February 02, 2017 09:21:58 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 Volume 79 Number 2 April 2017 / 000

lative seat, where the electoral payoff for split ticket voting only a small share of votes in legislative elections, the im-
is unclear. Analyses of the 2007 and 2015 elections indicate plementation of e-voting stands to alter electoral outcomes
that the significant differences observed between the treat- by facilitating ballot splitting in favor of the legislative list
ment and control group in 2011 are not a product of under- when the party list curries more favor with voters than does
lying differences between the two groups. Instead, there is the gubernatorial candidate. The overall substantive effects
strong evidence to support our argument that e-voting facil- of e-voting were curtailed in 2011, as it was only partially
itates ballot splitting. implemented in the province, but under full implementa-
tion of e-voting, the level of ballot splitting may be much
ELECTORAL IMPLICATIONS OF BALLOT SPLITTING larger as parties and voters learn to take advantage of the
The effect of e-voting is sizable and statistically significant, ease of ballot splitting on the electronic devises.
but the question remains, are the effects of e-voting on bal- Indeed, ballot splitting in the 2015 election resulted in a
lot splitting large enough to have electoral implications? To victory for the Partido Obrero (PO). In the district of Salta,
evaluate the electoral implications of e-voting, we simulated the gubernatorial candidate for the PO won 8.80% of the
election results over the distribution of party votes from the valid votes. Our simulations predict that if his party list
2011 election. The simulation is based on 10,000 indepen- won the exact same share of votes in the legislative election,
dent random draws, and the probability of winning is cal- the party would have less than a 50% chance of winning a
culated according to the set of electoral rules governing the legislative seat. The PO’s legislative list, however, won al-
state legislative election in the district of Salta: a district mag- most 2 percentage points more votes than did the guber-
nitude of 9, with a 5% vote threshold for winning a seat, natorial candidate (10.78% of the valid votes). According
and vote to seat allocation calculated based on the D’Hondt to our simulations, parties winning more than 10% of the
formula. valid votes in the legislative election are virtually guaran-
Figure 4 plots the simulation results. The results show teed a seat. And, consistent with our expectations, the PO
that for minor parties, small increases in the vote share won one seat in the district of Salta in 2015. One seat is not
obtained by the legislative ticket can dramatically increase trivial, as there are only nine seats up for grabs in the dis-
their probability of winning a seat in the legislature. Spe- trict. It is far less likely that the PO would have won enough
cifically, the predicted probability of winning one seat in votes to secure a seat if voters had cast their ballots using
office is only 7.8% when a party obtains 8.50% of the leg- the ballot-and-envelope system.
islative votes. The probability of winning a seat increases More generally, the PO’s case illustrates that the shift
substantially to 30.9% when the party obtains 8.75% of the away from the ballot-and-envelope systems may have im-
votes. Further, parties winning 9.00% of votes have a 70.4% portant electoral implications in settings where the effective
chance of winning a seat in office. Thus, for parties winning vote threshold to win a seat is low. In Argentina, for example,
multiple provinces with high district magnitudes—which
decreases the effective threshold—have piloted e-voting de-
vices (Calvo et al. 2008) or introduced paper Australian bal-
lots (Pomares and Zárate 2014). Our findings indicate that
these ballot reforms may manufacture opportunities for small
parties to win seats by facilitating ballot splitting.
The implications of this research go beyond Argentina.
The effect of a ballot reform on split ticket voting ultimately
depends on the difference between the layout of the old
ballot and the new ballot and the broader electoral rules
and dynamics. Where e-voting is introduced, the electronic
device defines the structure of the new ballot (Calvo et al.
2008; Herrnson et al. 2009). The screen may display the
full ballot at once (similar to an Australian paper ballot)
Figure 4. Predicted probability of winning a first seat. We use 10,000 in- or show the ballot by elected position, both making it easier
dependent random draws to estimate the predicted probability of a party
for voters to split their ballot. By contrast, if voters view the
winning a first seat in the legislative office (y-axis) as the percentage the
party wins increases (x-axis). The predicted probabilities are calculated
ballot organized by different political parties, ballot splitting
based on the electoral rule in the district of Salta, i.e., a 9 seat electoral may be more burdensome—albeit still easier than physi-
district, 5% threshold, and D’Hondt seat allocation formula. cally cutting different ballot strips. Finally, if the ballot

This content downloaded from 132.239.001.230 on February 02, 2017 09:21:58 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
000 / Ballot Structure and Split Ticket Voting Tiffany D. Barnes, Carolina Tchintian, and Santiago Alles

prompts voters to vote a straight ticket on the first screen, tive representation, future research should examine whether
it may discourage ballot splitting similar to the ballot-and- similar results hold for other contexts and other types of
envelope system. Nonetheless, the overall impact of ballot reforms. In particular, in contexts with large district mag-
reform depends on the broader electoral context. nitudes, the electoral effects of ballot splitting may be larger
given that the vote share necessary to win a seat decreases
CONCLUSIONS as the district magnitude increases. Moreover, our framework
Our study provides a systematic analysis of how a change is applicable to other types of reforms, such as replacing the
of the ballot structure introduced by the implementation ballot-and-envelope system with a single paper ballot (e.g.,
of e-voting can facilitate ballot splitting. We find support for Australian ballot), that simplify voting procedures and enable
the hypothesis that implementing an electronic ballot that citizens to more easily express their preferences at the polls.
makes ballot splitting more convenient, compared to the tra-
ditional ballot-and-envelope system, significantly increases the ACKNOWLEDGMENTS
share of split ballots. The party-level analysis shows that al- We would like to thank Rick Wilson, Robert Stein, Royce
though the average level of ballot splitting increases in pre- Carroll, Jean Aroom, Mark P. Jones, Mirya Holman, Jesse
cincts with electronic voting, the magnitude of this effect is Johnson, Emily Beaulieu, Facundo G. Galvan, Justin Wede-
larger for parties with no viable executive candidates or where king, editor Jeffery Jenkins, and three anonymous reviewers
the electoral payoff for split ticket voting was less clear for for very useful feedback. Moreover, this article benefited
the voter. from comments by participants of panels at the annual meet-
These findings have important implications for political ing of the Midwest Political Science Association (Chicago,
representation and the policy-making process alike. To be- April 3–6, 2014) and the Sociedad Argentina de Análisis
gin with, in presidential systems, voters often feel compelled Político [Argentine Association of Political Analysis] Na-
to vote for viable executive candidates; thus, executive coat- tional Conference (Mendoza, August 12–15, 2015), par-
tails reduce the number of legislative parties. Yet research ticipants of the Comparative Politics Workshop at Rice Uni-
has demonstrated that voters sometimes choose to split their versity, and participants of the Evidence in Governance and
vote, resulting in more diverse legislative representation. Politics (EGAP) poster session held at Rice University. We
Still, some voters who want to split their ballot often refrain also thank Teresa Ovejero at the Salta Electoral Tribunal for
from doing so because the process is unnecessarily compli- sharing data with us.
cated and laborious. This is particularly relevant when the
benefit from splitting the vote is uncertain. Simplified voting
REFERENCES
procedures, however, can provide voters an additional in- ACE, The Electoral Knowledge Network. 2011. “Senegal: Adoption of
centive to cast a split ticket vote by lowering the cost and the Single Ballot for 2012 Presidential Election,” November 9. http://ace
risk associated with this process. Consequently, the findings project.org/electoral-advice/archive/questions/replies/33586788 (accessed
from our research imply that ballot structures can inadver- May 3, 2016).
Áleman, Eduardo, and Ernesto Calvo. 2010. “Unified Government, Bill
tently affect the composition of legislative representation. Approval, and the Legislative Weight of the President.” Comparative
At the same time, voting procedures that increase the Political Studies 43 (4): 511–34.
probability of ballot splitting, and that therefore augment Alesina, Alberto, and Howard Rosenthal. 1995. Partisan Politics, Divided
the number of legislative parties in office, have consequences Government, and the Economy. New York: Cambridge University Press.
Alvarez, R. Michael, and Thad E. Hall. 2008. Electronic Elections: The
for substantive representation. When the effective number
Perils and Promises of Digital Democracy. Princeton, NJ: Princeton
of parties is large, the probability that any single party holds University Press.
a majority of seats is smaller, making it more difficult for Alvarez, R. Michael, Thad E. Hall, and Morgan H. Llewellyn. 2008. “Are
legislators to advance policy. Although this situation may Americans Confident Their Ballots Are Counted?” Journal of Politics
70 (3): 754–66.
increase the probability of legislative gridlock (Mainwaring
Alvarez, R. Michael, Gabriel Katz, Ricardo Llamosa, and Hugo Martínez.
1993), it is an opportunity to improve substantive repre- 2009. “Assessing Voters’ Attitudes towards Electronic Voting in Latin
sentation, as parties are required to cultivate a broad legis- America: Evidence from Colombia’s 2007 E-Voting Pilot.” In Peter
lative coalition to reach a policy outcome (Áleman and Calvo Y. A. Ryan and Berry Schoenmakers, eds., E-Voting and Identity. Berlin:
2010; Calvo 2014), increasing the number of views that are Springer, 75–91.
Alvarez, R. Michael, Gabriel Katz, and Julia Pomares. 2011. “The Impact
articulated in the policy-making process (Barnes 2016).
of New Technologies on Voter Confidence in Latin America: Evidence
Given the potential consequences of seemingly minor from E-Voting Experiments in Argentina and Colombia.” Journal of
changes in voting procedures on descriptive and substan- Information Technology and Politics 8:199–217.

This content downloaded from 132.239.001.230 on February 02, 2017 09:21:58 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 Volume 79 Number 2 April 2017 / 000

Alvarez, R. Michael, Ines Levin, Julia Pomares, and Marcelo Leiras. 2013. Ho, Daniel, Kosuke Imai, Gary King, and Elizabeth Stuart. 2007. “Match-
“Voting Made Safe and Easy: The Impact of E-voting on Citizen Per- ing as Nonparametric Preprocessing for Reducing Model Dependence
ceptions.” Political Science Research and Methods 1 (1): 117–37. in Parametric Causal Inference.” Political Analysis 15 (3): 199–236.
Ames, Barry, Andy Baker, and Lucio R. Renno. 2009. “Split-Ticket Voting Iacus, Stefano M., Gary King, and Giuseppe Porro. 2011. “Multivariate
as the Rule: Voters and Permanent Divided Government in Brazil.” Matching Methods That Are Monotonic Imbalance Bounding.” Journal
Electoral Studies 28 (1): 8–20. of the American Statistical Association 106 (493): 345–61.
Barnes, Tiffany D. 2016. Gendering Legislative Behavior: Institutional Iacus, Stefano M., Gary King, and Giuseppe Porro. 2012. “Causal Infer-
Constraints and Collaboration. New York: Cambridge University Press. ence without Balance Checking: Coarsened Exact Matching.” Political
Barnes, Tiffany D., and Gabriela Rangel. 2014. “Election Law Reform in Analysis 20 (1): 1–24.
Chile: The Implementation of Automatic Registration and Voluntary Jones, Mark P. 1997. “Federalism and the Number of Parties in Argentine
Voting.” Election Law Journal 13 (4): 570–82. Constitutional Elections.” Journal of Politics 59 (2): 538–49.
Beaulieu, Emily. 2015. “Electronic Voting and Perceptions of Election Katz, Gabriel, R. Michael Alvarez, Ernesto Calvo, Marcelo Escolar, and
Fraud and Fairness.” Journal of Experimental Political Science 3 (1): 1–14. Julia Pomares. 2011. “Assessing the Impact of Alternative Voting Tech-
Beck, Paul, Lawrence Baum, Aage R. Clausen, and Charles E. Smith, Jr. nologies on Multi-Party Elections: Design Features, Heuristic Processing
1992. “Patterns and Sources of Ticket Splitting in Subpresidential and Voter Choice.” Political Behavior 33 (2): 247–70.
Voting.” American Political Science Review 86 (4): 916–28. Kersting, Norbert, and Harald Baldersheim. 2004. Electronic Voting and
Brunell, Thomas, and Bernard Grofman. 2009. “Testing Sincere versus Democracy: A Comparative Analysis. Basingstoke. London: Palgrave
Strategic Split-Ticket Voting at the Aggregate Level: Evidence from Split Macmillan.
House–President Outcomes, 1900–2004.” Electoral Studies 28:62–69. Kimball, David C., and Martha Kropf. 2005. “Ballot Design and Unre-
Burden, Barry, and Gretchen Helmke. 2009. “The Comparative Study of corded Votes on Paper-Based Ballots.” Public Opinion Quarterly 69 (4):
Split-Ticket Voting.” Electoral Studies 28:1–7. 508–29.
Burden, Barry C., and David C. Kimball. 1998. “A New Approach to the Study King, Gary, Christopher Lucas, and Richard Nielsen. 2014. “The Balance-
of Ticket Splitting.” American Political Science Review 92 (3): 533–44. Sample Size Frontier in Matching Methods for Causal Inference.”
Burden, Barry C., and David C. Kimball. 2002. Why Americans Split Their http://j.mp/1dRDMrE (accessed November 24, 2014).
Tickets. Ann Arbor: University of Michigan Press. King, Gary, and Langche Zeng. 2006. “The Dangers of Extreme Counter-
Calvo, Ernesto. 2014. Legislative Success in Fragmented Congresses in Argen- factuals.” Political Analysis 14 (2): 131–59.
tina: Plurality Cartels, Minority Presidents, and Lawmaking. Cambridge: Mainwaring, Scott. 1993. “Presidentialism, Multipartism, and Democracy.
Cambridge University Press. The Difficult Combination.” Comparative Political Studies 26 (2): 198–
Calvo, Ernesto, Marcelo Escolar, and Julia Pomares. 2008. “Ballot Design 228.
and Split Ticket Voting in Multiparty Systems: Experimental Evidence Massicotte, Louis, Andre Blais, and Antoine Yoshinaka. 2004. Establishing
on Information Effects and Vote Choice.” Electoral Studies 28 (2): 218–31. the Rules of the Game: Election Laws in Democracies. Toronto: Uni-
Campbell, Angus, and Warren E. Miller. 1957. “The Motivational Basis of versity of Toronto Press.
Straight and Split Ticket Voting.” American Political Science Review 51 McAllister, Ian. 2015. “Voting Behavior and Political Institutions.” Routledge
(2): 293–312. Handbook of Comparative Political Institutions. London: Routledge.
Campbell, Angus, Philip Converse, Warren E. Miller, and Donald E. Stokes. Moser, Robert G., and Ethan Scheiner. 2009. “Strategic Voting in Estab-
1960. The American Voter. Chicago: University of Chicago Press. lished and New Democracies: Ticket Splitting in Mixed-member Elec-
Card, David, and Enrico Moretti. 2007. “Does Voting Technology Affect toral Systems.” Electoral Studies 28:51–61.
Election Outcomes? Touch-Screen Voting and the 2004 Presidential Nichols, Stephen M., and Gregory A. Strizek. 1995. “Electronic Voting
Election.” Review of Economics and Statistics 89:660–73. Machines and Ballot Roll-off.” American Politics Research 23 (3): 300–
Cox, Gary W. 1997. Making Votes Count: Strategic Coordination in the 318.
World’s Electoral Systems. Cambridge: Cambridge University Press. ONPE (Oficina Nacional de Procesos Electores). 2014. “Buenas prácticas
Darcy, Robert, and Schneider, Anne. 1989. “Confusing Ballots, Roll-off, en torno al voto electrónico en América. Reflexiones y lecciones desde
and the Black Vote.” Western Political Quarterly 43:347–64. los estándares electorales internationales [Best practices on electronic
Delwit, Pascal, Erol Kulachi, and Jean-Benoit Pilet. 2005. “Electronic voting in America: Thoughts and lessons using international electoral
Voting in Belgium: A Legitimised Choice?” Politics 25 (3): 153–64. standards].” Documento de Trabajo 35. ONPE, Lima, Peru.
Fujiwara, Thomas. 2015. “Voting Technology, Political Responsiveness, Pomares, Julia, Marcelo Leiras, Maria Page, Carolina Tchintian, and
and Infant Health: Evidence from Brazil.” Econometrica 83 (2): 423–64. Anastasia Peralta Ramos. 2011. “Cambios en la forma de votar: La ex-
Gelman, Andrew, and Jennifer Hill. 2007. Data Analysis Using Regression periencia del voto electrónico en Salta.” Documento de Políticas Públicas/
and Multilevel/Hierarchical Models. New York: Cambridge University Recomendación No. 94, CIPPEC (Centro de Implementacíon de Polí-
Press. ticas Públicas para la Equidad y el Crecimiento) [Center for Public Policy
Gelman, Andrew, U-Sung Su, Masanao Yajima, Jennifer Hill, Maria Grazia Implementation toward Equity and Growth], Buenos Aires.
Pittau, Jouni Kerman, Tian Zheng, and Vicent Dorie. 2015. “Data Analy- Pomares, Julia R., Ines Levin, and Michael Alvarez. 2014a. “Do Voters and
sis Using Regression and Multilevel/Hierarchical Models.” https://cran Poll Workers Differ in Their Attitudes toward E-voting? Evidence from
.r-project.org/web/packages/arm/arm.pdf (accessed May 7, 2016). the First E-election in Salta, Argentina.” Journal of Election Technology
Helmke, Gretchen. 2009. “Ticket Splitting as Electoral Insurance: The Mex- and Systems 2 (2): 1–10.
ico 2000 Elections.” Electoral Studies 28 (1): 70–78. Pomares, Julia R., Ines Levin, Michael Alvarez, Guillermo L. Mirau, and
Herrnson, Paul S., Richard G. Niemi, Michael J. Hanmer, Benjamin B. Teresa Ovejero. 2014b. “From Piloting to Roll-out: Voting Experience
Bederson, Frederick G. Conrad, and Michael W. Traugott. 2009. Voting and Trust in the First Full E-election in Argentina.” Presented at the
Technology: The Not-So-Simple Act of Casting a Ballot. Washington, 6th International Conference on Electronic Voting: Verifying the Vote,
DC: Brookings Institution Press. Institute of Electrical and Electronics Engineers (IEEE), October 29–31.

This content downloaded from 132.239.001.230 on February 02, 2017 09:21:58 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
000 / Ballot Structure and Split Ticket Voting Tiffany D. Barnes, Carolina Tchintian, and Santiago Alles

Pomares, Julia, and Soledad Zárate. 2014. “Cambios en la forma de votar: Rusk, Jerrold G. 1970. “The Effect of the Australian Ballot Reform on Split
la primera elección provincial completa de un sistema electrónico de Ticket Voting: 1876–1908.” American Political Science Review 64 (4): 1220–38.
votación. Salta, 2013” [Changes in the voting procedure: The first pro- Shugart, Matthew Soberg. 1995. “The Electoral Cycle and Institutional
vincial election using electronic voting, Salta, 2013]. Documento de Pol- Sources of Divided Presidential Government.” American Political Sci-
íticas Públicas/Recomendación no. 130. CIPPEC, Buenos Aires. ence Review 89 (2): 327–43.
Prasad, Hari K., J. Alex Halderman, Rop Gonggrijp, Scott Wolchok, Eric Shugart, Matthew Soberg, and John M. Carey. 1992. Presidents and As-
Wustrow, Arun Kankipati, Sai Krishna Sakhamuri, and Vasavya Ya- semblies: Constitutional Design and Electoral Dynamics. Cambridge:
gati. 2010. “Security Analysis of India’s Electronic Voting Machines.” Cambridge University Press.
In Proceedings of the 17th ACM Conference on Computer and Com- Thompson Jiménez, José. 2009. “La experiencia reciente del voto elec-
munications Security, 2010. Chicago: Association for Computing Ma- trónico en América Latina: Avances y perspectivas” [The recent elec-
chinery. tronic voting experience in Latin America: Progresses and perspectives].
Reynolds, John F., and McCormick, Richard L. 1986. “Outlawing ‘Treach- Revista de Derecho Electoral 7:5–23.
ery’: Split Tickets and Ballot Laws in New York and New Jersey, 1880– Tula, María Inés. 2005. Voto electrónico: Entre votos y máquinas: Las nuevas
1910.” Journal of American History 72 (4): 835–58. tecnologías en los procesos electores [Electronic vote: Between votes and
Rezende, Pedro. 2003. “Electronic Voting Systems: Is Brazil Ahead of Its machines; the new technology in the electoral processes]. Buenos Aires:
Time?” Presented at the First Workshop on Voter-Verifiable Election Ariel.
Systems, Denver, CO. Walker, Jack L. 1966. “Ballot Forms and Voter Fatigue: An Analysis of the
Rich, Timothy. 2014. “Split-Ticket Voting in South Korea’s 2012 National Office Block and Party Column Ballots.” Midwest Journal of Political
Assembly Election.” Asian Politics an Policy 6 (3): 455–69. Science 10 (4): 448–63.

This content downloaded from 132.239.001.230 on February 02, 2017 09:21:58 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).