[22nd General Election Research Series] The 22nd National Assembly Election and Patterns of Split Voting

1. The One-Person, Two-Vote System and Voting Choices

The Constitutional Court ruled in 2001 that electing district and proportional representatives simultaneously under a one-person, one-vote system was unconstitutional. Consequently, since the 2002 simultaneous local elections and the 17th National Assembly election in 2004, votes for district candidates and proportional representatives have been cast separately, allowing voters to cast split votes for different parties in each election. The question then arises: why do voters engage in split voting? What are the patterns of split voting in the 22nd National Assembly election? Are there differences based on political parties? Who engages in split voting?

A study by Cho Sung-dae (2020), analyzing survey results from the 21st National Assembly election, found that approximately 32.3% of voters engaged in split voting. Jung Min-seok and Lee Hyun-woo (2022), analyzing survey data from the 20th National Assembly election, reported that about 18.33% of participants engaged in split voting. In contrast, according to Choi Hyo-no (2015), who analyzed the 6th simultaneous local elections in 2014, one survey indicated that approximately 35.7% of respondents engaged in split voting, while another found that about 39.8% of participants did so. These results suggest that split voting is not an exceptional behavior and that variations exist across elections.

What, then, was the proportion of split voters in the 2024 general election? According to survey results conducted by the East Asia Institute immediately after the election[1], the proportion of split voters in the 22nd general election was approximately 41.42%.[2] This figure is relatively high compared to previous general elections. The emergence of the Cho Kuk Innovation Party, a proportional representation party, may have contributed to this high figure. When asked about their voting choices, approximately 24.39% of survey respondents indicated they voted for the Cho Kuk Innovation Party in the proportional representation election.[3] So, what are the reasons for voters engaging in split voting?

2. Why Do Voters Engage in Split Voting?

Scholars often view split voting as the result of rational and strategic choices by voters (Burden and Helmke 2009; Fiorina 1996). Here, 'rational' implies that voters vote according to their preferences. 'Strategic,' on the other hand, means that voters do not always make their most preferred choice to achieve a certain objective. If voters expect their actions to yield utility, they will not vote for candidates with no chance of winning (Duverger 1954). If a voter's most preferred candidate has a very low probability of winning, but their second most preferred candidate has a moderate chance of winning, this voter, wishing to avoid a wasted vote, may choose to vote for the second-best candidate rather than the top choice.

The behavior of voters seeking to avoid wasted votes can lead to split voting. In South Korean general elections, where voters cast one vote for district representatives and another for proportional representatives, split voting occurs if a voter chooses their most preferred party for the proportional representation election but votes for a more viable candidate in the district election to avoid a wasted vote. This explanation assumes that voters' party preferences determine their candidate choices. However, in reality, voters may evaluate individuals rather than parties when choosing a district candidate. For voters who base their district candidate selection on individual assessments, split voting can occur when their assessment of a candidate's character differs from their party preference. This can be difficult to distinguish from split voting aimed at avoiding wasted votes. For example, if a voter who most prefers a minor party selects a major party candidate in the district election based on a higher personal evaluation, while voting for their most preferred minor party in the proportional representation election, this voter could be seen as strategically split voting to avoid a wasted vote. However, since the criterion for selecting the district candidate was the individual rather than the party, this can also be interpreted as voting according to preference.

Secondly, voters may engage in split voting by considering inter-party cooperation in the legislature after the election (Gschwend 2007). Voters may vote not simply for a party or candidate to enter the legislature, but by considering the legislative activities of parties (Kedar 2009). This is because what matters to voters are the policies that will be enacted after the election. If voters consider post-election legislative decision-making, they will vote by considering which party or bloc will hold legislative initiative and whether legislative alliances or cooperation will occur among parties. In this scenario, voters may vote for a party they do not most support. For instance, a center-right voter might vote for a center-right party based on preference, but this party might form an alliance with a centrist party after the election, leading to insufficient production of right-wing policies. Considering this, the voter might vote for a right-wing party, rather than their most preferred center-right party, hoping that this party will enter the legislature and cooperate with the center-right party to realize more right-wing policies.

Such split voting is typically observed in parliamentary systems with multiple parties and significant power dispersion within the legislature. However, it can also occur in presidential systems with a one-person, two-vote system for legislative elections. For example, a voter who believes their most preferred party has already secured sufficient seats[4] may split their votes between proportional and district representatives, hoping their second-most preferred party will enter the legislature and cooperate with their top-choice party. For instance, in systems with electoral thresholds, voters might consider split voting to help a second-choice minor party enter parliament.

Thirdly, voters may engage in split voting to ensure checks and balances among parties (Fiorina 1996). Voters who lack strong partisan loyalty or have moderate policy preferences may not want a single party to hold both legislative and executive power. If a left-wing or right-wing party controls both the legislature and the executive, policies might lean heavily towards that ideology. Therefore, voters with moderate policy stances might choose to split their votes between left-wing and right-wing parties. For similar reasons, the vote share for the ruling party may decrease in parliamentary elections held in years without a presidential election (Kedar 2009). Voting behavior aimed at checks and balances is primarily observed in vertical split voting.[5]

Theoretically, split voting based on checks and balances can also occur in National Assembly elections with a one-person, two-vote system. For example, a voter might cast one vote for a district or proportional representative according to their preference, and the other vote for a competing party based on their preference. For the logic of checks and balances to be practically applied, votes would need to be split between the two major parties. However, it is rare for preferences for two fiercely competing major parties to be at a similar level, and splitting votes between the two major parties in district and proportional representation elections is also uncommon (Cho Sung-dae 2020).

Fourthly, there are cases of non-strategic split voting. Previously, it was argued that voters engage in split voting by choosing to vote for a major party candidate with a high chance of winning, rather than their second-best choice, to avoid wasted votes. Voters may also vote for a party that is not their top preference based on post-election policy considerations. Unlike these strategic choices, voters might vote for their most preferred minor party in the proportional representation election and for their second-best choice, which is also a minor party, in the district election. This voting behavior can be considered non-strategic in the sense that it involves voting for a minor party candidate in a district election with a low chance of winning (Park Chan-wook 2004). However, it can be considered rational in terms of voting according to preference.

Finally, voters' split voting can occur in ways that are not rational. This applies to cases where voters split their votes but do not vote according to their preferences. If voters choose candidates based on criteria other than party preference, voting behavior that does not align with party preferences may be observed. As mentioned earlier, split voting that does not align with party preferences can occur when voting is based on an evaluation of individuals rather than party preferences. Furthermore, split voting can arise from conflicting attitudes. When voters have unclear or conflicting party preferences, splitting their votes may lead to voting behavior that does not align with their preferences.

3. The 22nd National Assembly Election and Types of Split Voting

Based on the reasons why rational voters engage in split voting, existing research categorizes split voting into types such as preventing wasted votes, ensuring alliances, checks and balances, non-strategic voting, and overlapping types (Park Chan-wook 2004; Cho Sung-dae 2020). This study also categorizes the types of split voting in the 22nd National Assembly election according to the classification methods of existing research. As introduced earlier, the 'preventing wasted votes' type refers to cases where a voter who most prefers a minor party votes for that minor party in the proportional representation election and for a preferred candidate among those with a high chance of winning in the district election, or votes for one of the major party candidates when preferences for the major parties are equal.[6] For example, if a voter who most prefers the Reform Party votes for the Democratic Party or the People Power Party in the district election based on preference, it is classified as the 'preventing wasted votes' type.

The 'ensuring alliances' type refers to cases where a voter votes in the district election according to their highest party preference and then votes for their second-most preferred party in the proportional representation election. For example, if the most preferred party is the Democratic Party and the second-most preferred party is the Cho Kuk Innovation Party, the voter votes for the Democratic Party candidate in the district election based on their top preference and for the Cho Kuk Innovation Party, their second-most preferred party, in the proportional representation election.

The 'checks and balances' type refers to voting behavior intended to create checks and balances between major parties. In other words, it involves splitting votes between major parties. For example, a voter whose most preferred party is either the Democratic Party or the People Power Party votes for the Future Korea Party or the Democratic Alliance in the proportional representation election, contrary to their party preference. Cases where votes are split between two parties with equal preference levels are also included in this type.

The 'non-strategic' type involves voting for the most preferred minor party in the proportional representation election and, if the second-most preferred party is also a minor party, voting for that minor party in the district election. Alternatively, voting for a minor party candidate in the district election based on preference, despite a low chance of winning, can be considered non-strategic. The higher the 'preventing wasted votes' sentiment and the lower the overall chance of minor party candidates winning, the lower the proportion of this type is likely to be.

There are also overlapping types of split voting. For example, if a voter has the same preference for a major party and a minor party, voting for the major party in the district election and the minor party in the proportional representation election could be for the purpose of preventing wasted votes, but it could also be for ensuring alliances if the preferences are equal. A second overlapping type occurs when the most preferred party is a minor party, and the second-most preferred options include both a minor party and a major party. In this case, voting for the most preferred minor party in the proportional representation election and choosing a major party for the district election could be classified as preventing wasted votes, while voting for a minor party could be considered non-strategic. If there are three or more most preferred parties, various overlapping types are possible. For instance, if the most preferred parties include two major parties and one minor party, classification into preventing wasted votes, ensuring alliances, and checks and balances types is all possible. Finally, cases where split voting occurs when preferences for multiple minor parties and major parties are all equal are also included in the overlapping types.

As mentioned, split voting based on rationality assumes that voters vote according to their preferences. However, in reality, non-preferential split voting behavior also exists. For example, in the case of split voting to prevent wasted votes, voters should vote for their most preferred minor party in the proportional representation election and choose a candidate from a major party in the district election according to their preference. However, there are instances where voters do not vote in accordance with their stated party preference order. Split votes that do not fall into the aforementioned categories and do not align with party preferences are included in the 'other' type. <Table 1> provides information on the types of split voting.

<Table 1> Types of Split Voting (%)

According to <Table 1>, excluding the 'other' category, the most frequent type of split voting in this election was 'preventing wasted votes,' accounting for approximately 34.4% of split votes and 22.43% of all votes. Cho Sung-dae (2020), analyzing survey results after the 21st National Assembly election, reported that the 'preventing wasted votes' type accounted for about 13.3%, while the most frequent type was 'ensuring alliances' at approximately 41.7%. In contrast, the 'ensuring alliances' type in this election was 22.99%, making it the second highest proportion after excluding overlapping types. As mentioned earlier, the proportion of the 'checks and balances' type was low at 4.6%, and the 'non-strategic' type was the lowest at 2.3%.

The high proportion of the 'preventing wasted votes' type in the 22nd National Assembly election is likely related to the participation of several proportional representation parties, including the Cho Kuk Innovation Party. Since proportional representation parties do not field district candidates, their supporters must engage in split voting. For example, cases where respondents who most preferred the Cho Kuk Innovation Party voted for it in the proportional representation election and for the Democratic Party in the district election numbered 105, accounting for approximately 19.96% of all split votes.[7] Additionally, cases classified as 'ensuring alliances,' where voters who most preferred the Democratic Party chose a Democratic Party candidate in the district election and their second-most preferred party, the Cho Kuk Innovation Party, in the proportional representation election, numbered 55.[8] Excluding overlapping and other types, the fact that the Cho Kuk Innovation Party accounts for over half of the 'preventing wasted votes' and 'ensuring alliances' types, which represent the highest proportions, suggests its significant impact on split voting in this election.

4. Patterns of Split Voting

The theoretical reasons for voters engaging in split voting were examined, followed by an analysis of the proportions of each type. What, then, are the patterns of split voting observed in the 22nd National Assembly election? To understand these patterns, we first examine district and proportional representation votes by party, as shown in <Table 2>.[9] Comparing consistent voting and split voting patterns by party, the People Power Party shows a higher proportion of consistent voting than the Democratic Party. Based on district voters, approximately 46.08% of Democratic Party voters and about 79.27% of People Power Party voters voted consistently. This could be due to Democratic Party supporters voting for the Cho Kuk Innovation Party following its emergence. Although the number of observations is small, making it difficult to assign significance, among minor parties, the Reform Party shows a high consistent voting rate (approximately 60.98%), while the New Future shows the lowest (approximately 47.37%).

<Table 2> District and Proportional Representation Votes (% relative to District Votes) (% relative to Proportional Representation Votes)

Conversely, based on proportional representation votes, 91.86% of voters who voted for the Democratic Alliance also voted for the Democratic Party in the district election, and 92.81% of voters who voted for the Future Korea Party also voted for the People Power Party. Among minor parties, the Reform Party had the highest consistent voting rate at approximately 32.05%, while the New Future had the lowest at approximately 25%.

<Table 2> shows that approximately 41.83% of voters who voted for the Democratic Party in the district election also voted for the Cho Kuk Innovation Party in the proportional representation election. Similarly, about 81.27% of voters who voted for the Cho Kuk Innovation Party in the proportional representation election chose a Democratic Party candidate in the district election. Excluding respondents who voted for the Democratic Party in the district election, the number of respondents who voted for the Cho Kuk Innovation Party in the proportional representation election is not very large. The appeal to 'vote for the Democratic Party in the district and the Cho Kuk Innovation Party in the proportional representation' appears to have been effective during the election period.

Split voting patterns can also be examined through party support. <Table 3> provides information on party support and split voting.[10] According to this table, while approximately 54.17% of voters who identified as supporters of the Democratic Party voted consistently, about 84.31% of respondents who identified as supporters of the People Power Party voted consistently. The lower consistent voting rate among Democratic Party supporters is likely due to votes cast for the Cho Kuk Innovation Party. In contrast, the high split voting rate among supporters of other minor parties can be inferred as a phenomenon resulting from voting for major parties in district elections to avoid wasted votes. Furthermore, since minor parties sometimes do not field district candidates, split voting rates tend to be high. The split voting rate among voters who stated they do not support any particular party is approximately 52.43%, showing no significant difference.

<Table 3> Support by Party and Split Voting (%)

To confirm these inferences, it is necessary to examine the split voting behavior according to party support in more detail. <Table 4> categorizes split voting types by party support.[12] As previously confirmed, most voters who voted for the Cho Kuk Innovation Party in the proportional representation election also voted for the Democratic Party candidate in the district election. This type of split voting is classified as 'preventing wasted votes,' and indeed, split voting among Cho Kuk Innovation Party supporters was predominantly categorized as such. According to this table, 44.14% of voters who identified as supporters of the Cho Kuk Innovation Party engaged in split voting to prevent wasted votes.[13]

<Table 4> Types of Split Voting by Supported Party (% by Party) (% by Type)

In contrast, among those who identified as supporters of the Democratic Party, excluding consistent voters, it is highly likely they supported the Cho Kuk Innovation Party. Support for the Cho Kuk Innovation Party by Democratic Party supporters can be viewed as split voting to ensure alliances. In fact, according to this table, approximately 13.16% of Democratic Party supporters were classified under the 'ensuring alliances' type. This figure is the highest after consistent voting and overlapping types. Compared to the 54.17% of Democratic Party supporters who voted consistently, the 84.31% of People Power Party supporters who voted consistently stands out. When considering all consistent voters, People Power Party supporters account for the largest proportion at approximately 52.58%. In contrast, Democratic Party supporters account for about 33.56% of consistent voters.

Similar to the split voting patterns based on party support, patterns based on ideology are also observed. <Table 5> shows the differences in split voting based on ideology. This table reveals a tendency for consistent voting rates to increase and split voting rates to decrease among respondents who identify as conservative. In this election, supporters of the People Power Party and conservatives exhibit relatively higher consistent voting rates compared to supporters of the Democratic Party or progressive voters.

<Table 5> Ideology and Split-Ticket Voting (%)

The variation in split-ticket voting based on ideology or partisanship observed in the 22nd National Assembly election can be considered an exception rather than the norm. Existing research suggests that the patterns of split-ticket voting are more likely to vary based on the intensity of ideology or partisanship rather than their presence alone (Jeong Min-seok and Lee Hyun-woo 2022; Jo Seong-dae 2020; Choi Hyo-no 2015). To confirm this assertion, it is necessary to examine in more detail who engages in split-ticket voting.

5. Variables and Data: Who Votes a Split Ticket?

If split-ticket voting is understood as a strategic choice, then voters who engage in it can be seen as having both the motivation and the ability to do so. For instance, voters with strong preferences for a particular party may have less motivation for split-ticket voting compared to those without such strong preferences. Indeed, existing research indicates that the stronger the existing preferences, such as partisanship or ideology, the lower the likelihood of split-ticket voting (Jeong Min-seok and Lee Hyun-woo 2022).

This study measures the impact of partisanship and ideology intensity on split-ticket voting, using them as independent variables. The survey analyzed in this study asked two questions regarding party support. First, it is based on the question, 'Do you have a party you support?' Those who answered 'yes' to this question can be considered voters with relatively strong partisanship. If they answered 'no' to the first question but 'yes' to the second question, 'Is there any party you prefer even slightly?', they are considered to have relatively weak partisanship. Those who answered 'no' to both questions are classified as independents. Ideology is measured by voters' self-assessment. Survey participants chose their ideology on a scale from very liberal (0) to very conservative (10). The intensity of ideology is set with 5 as the midpoint, with intensity increasing as one moves towards liberal (0) or conservative (10).

If split-ticket voting is strategic, voters will consider the policies or ideologies of political parties when voting. For example, voters who perceive little difference between the policies or ideologies of parties are more likely to vote a split ticket. Specifically, voters who feel a large ideological distance between themselves and the parties may attempt to increase their voting utility by voting a straight ticket. This study measures the ideological distance between voters and each party and identifies the difference between the ideological distance to the closest party and the second closest party as variables. The smaller this difference, the greater the likelihood that voters will engage in split-ticket voting.

We have examined the impact of partisanship and ideology, which can be classified as motivations, on split-ticket voting. In addition to motivation, the ability to make strategic choices is also important for split-ticket voting. Existing research measures the ability for split-ticket voting using political knowledge or educational attainment (Jeong Min-seok and Lee Hyun-woo 2022; Jo Seong-dae 2020). This study follows existing research by measuring educational attainment and[14]political knowledge as independent variables. Political knowledge is measured by the number of correct answers survey participants gave to political questions.[15]These variables are expected to have a positive impact on split-ticket voting.

While split-ticket voting may be a strategic choice for voters, it can also occur when preferences for parties are ambiguous. Existing research suggests that voters are more likely to engage in split-ticket voting when their feelings toward parties are ambiguous or ambivalent (Jeong Min-seok and Lee Hyun-woo 2022). This study adopts an approach that simultaneously considers the intensity and difference of feelings, as proposed by existing research (Thompson, Zanna and Griffin 1995), to measure the ambivalence of voters' feelings. Ambivalent feelings are calculated as the difference between the intensity of favorability toward a party and the difference in favorability. [16]The formula for this is as follows (Jeong Min-seok and Lee Hyun-woo 2022; Heo Seok-jae and Jeong Han-wool 2019). As ambivalent feelings increase, meaning the difference in preference between two parties is not pronounced, the likelihood of split-ticket voting will increase.

Theoretically, split-ticket voting generally revolves around voting for a party. For this reason, the likelihood of split-ticket voting may differ depending on the party supported. For example, supporters of major parties such as the Democratic Party and the People Power Party are more likely to vote a straight ticket compared to supporters of minor parties. This study measures voters' party support and uses it as an independent variable.[17]Additionally, to reflect the characteristics of this election, the response to 'the meaning of the election'[18]is included as a control variable in the model. This variable may also include an assessment of the Yoon Suk-yeol administration along with the meaning of the election. Finally, birth year and gender are included as control variables in the analysis model.

6. Regression Analysis Results

This study conducts a regression analysis with split-ticket voting as the dependent variable to systematically understand split-ticket voting.[19][Table 6] presents the results. According to this table, the intensity of party supportModel 1andModel 2shows statistical significance. Based on the results ofModel 2, which has higher efficiency among the models including this variable,[20]assuming other variables are at their average,[21]the predicted probability of split-ticket voting for voters who do not have a party to support is approximately 52.03%, while the predicted probability for voters who strongly support a party is only about 35.25%. In other words, the stronger the party support, the lower the likelihood of split-ticket voting.

<Table 6> Regression Analysis of Split-Ticket Voting

However, unlike party intensity, ideology-related variables do not show statistical significance. Even when using the difference in distance between the party closest in ideology and the second closest party as a variable, no significant ideological influence is observed, not only for the intensity of ideology. Even when excluding the ideology intensity variable, the influence of the ideological distance difference variable is not statistically significant. Even when creating interaction terms with political knowledge below and including them in the model, statistical significance is not observed. These results suggest that voters do not evaluate the ideological differences between parties to engage in straight-ticket or split-ticket voting.

If split-ticket voting is strategic, it requires not only motivation but also ability. Ability is measured by the variables of political knowledge and education. Models 1, 2, and 3show that as voters' political knowledge increases, the likelihood of split-ticket voting decreases. This is contrary to expectations. <Figure 1>, based on the results ofModel 4, which has the highest efficiency and explanatory power, shows a decrease in the predicted probability of split-ticket voting as political knowledge increases. The predicted probability of split-ticket voting for voters with the lowest political knowledge is approximately 46.72%, while for citizens with the highest political knowledge, this probability is about 36.54%. This result implies that voters with high political knowledge tend to vote a straight ticket. On the other hand, the education variable does not show statistical significance.

<Figure 1> Political Knowledge and Split-Ticket Voting[22]

Existing research suggests that the more ambiguous feelings toward a party, the greater the likelihood of split-ticket voting (Jeong Min-seok and Lee Hyun-woo 2022; Heo Seok-jae and Jeong Han-wool 2019). To measure how clear or distinct voters' feelings toward parties are, this study first examines the standard deviation of favorability ratings for each party.[23]A smaller value indicates that the difference in favorability ratings among parties is not large. That is, as this value decreases, the likelihood of split-ticket voting will increase. However, contrary to expectations, this variable did not show statistical significance.

Even if the ambiguity of feelings affects split-ticket voting, voters will not consider their feelings toward all parties when voting. That is, it is necessary to measure the ambiguity of feelings between the parties that voters consider for voting. To measure the ambiguity of feelings between the most preferred party and the second most preferred party, the variable 'ambivalent feelings,' which simultaneously considers the intensity and difference of feelings, is included inModel 2andModel 4.[24]The results of this model show that as the ambivalence of feelings toward the preferred and second-preferred parties increases, meaning the difference in feelings toward them is less pronounced, the likelihood of split-ticket voting increases. <Figure 2> shows the predicted probability of split-ticket voting according to the variation in ambivalent feelings, based onModel 4.[25]When ambivalent feelings are lowest, the predicted probability of split-ticket voting is approximately 22.58%, but it increases to 58.52% when they are highest. This result is consistent with previous claims that the more similar the favorability ratings toward two preferred parties, the greater the likelihood of split-ticket voting (Jeong Min-seok and Lee Hyun-woo 2022).

<Figure 2> Ambivalent Feelings and Split-Ticket Voting

Along with ambivalent feelings toward parties, party support can also influence split-ticket voting. Specifically, split-ticket voting behavior may differ depending on the party supported. As mentioned earlier, supporters of major parties are relatively more likely to vote a straight ticket. Conversely, supporters of minor parties are more likely to engage in split-ticket voting. Indeed, the results in this table show that voters who support the Democratic Party and the People Power Party tend to vote a straight ticket. Furthermore, supporters of the Rebuilding Korea Party, which did not field a candidate in the district, showed a high probability of split-ticket voting.

<Figure 3> is based onModel 4Based on the results, it shows the predicted probability of split-ticket voting according to support for the Democratic Party, the People Power Party, and the Cho Kuk Innovation Party.[26]The predicted probability of split-ticket voting for Democratic Party supporters is approximately 31.32%, and for People Power Party supporters, this probability is approximately 21.88%. In contrast, for voters who responded that they support the Cho Kuk Innovation Party, the predicted probability of split-ticket voting reaches 84.53%.[27]These results indicate that voters exhibited different split-ticket voting behaviors depending on the party they supported in this election.

<Figure 3> Party Support and Split-Ticket Voting

Finally, among the results in this table, the variable "Meaning of the Election" stands out. This variable was constructed based on voters' evaluations of the statement, 'This election should be one to empower the Yoon Suk-yeol administration.' In all models, the probability of split-ticket voting significantly increases as opposition to this statement grows. <Figure 4> shows the predicted probability of split-ticket voting according to voters' evaluations of the meaning of the election.

<Figure 4> Meaning of the Election and Split-Ticket Voting

Looking at the results of the fourth model, the predicted probability of split-ticket voting for voters who completely disagree with the meaning of the election is approximately 50.38%, whereas for respondents who strongly agree, the probability is only 27.71%. Voters who believe the election should empower the Yoon Suk-yeol administration likely voted consistently, while a majority of voters who disagreed with this sentiment probably engaged in split-ticket voting. This may also be related to the emergence of the Cho Kuk Innovation Party, which emphasized a referendum on the Yoon Suk-yeol administration in this election. Furthermore, even among People Power Party supporters, those who held negative views of the Yoon Suk-yeol administration might have split their votes between the district and proportional representation constituencies.

5. Conclusion

This study examines the split-ticket voting behavior observed in the 22nd National Assembly election. Analysis of the survey data reveals that the proportion of split-ticket voting in this election was approximately 41.42%, which is relatively high compared to previous National Assembly elections. This is likely related to the Cho Kuk Innovation Party achieving an impressive 24.25% of the vote, despite not fielding candidates in district constituencies. Additionally, Democratic Party supporters voting for the Cho Kuk Innovation Party can also explain the high rate of split-ticket voting.

If split-ticket voting is understood as a strategic choice, its intentions can be categorized. This paper classifies split-ticket voting into types such as preventing wasted votes, coalition formation, checks and balances, non-strategic voting, and duplicate voting, following existing research (Park Chan-wook 2004; Cho Sung-dae 2020). According to the classification results, the most frequent types of split-ticket voting observed in this election, excluding the 'other' category, were duplicate voting and preventing wasted votes, at approximately 35.28% and 34.4%, respectively. The emergence of numerous proportional representation parties and the vote share of the Cho Kuk Innovation Party can explain the high proportion of the 'preventing wasted votes' type. Furthermore, Democratic Party supporters voting for the Cho Kuk Innovation Party likely contributed significantly to the proportion of the 'coalition guarantee' type (approximately 23.32%).

So, who engaged in split-ticket voting in the 22nd National Assembly election? Firstly, as argued in previous research, voters who support a particular party are more likely to vote consistently than those who do not (Park Chan-wook 2004; Yoon Kwang-il 2014; Jeong Min-seok·Lee Hyun-woo 2022; Cho Sung-dae 2020). This phenomenon is particularly observed among supporters of the major parties, the Democratic Party and the People Power Party. Ideological intensity, ideological distance, or perceived ideological position, which could be other motivational variables, did not explain split-ticket voting. Conversely, political knowledge, which could be a variable for the ability to split-ticket vote, showed an increase in the likelihood of consistent voting as it increased.

Split-ticket voting can be a strategic choice, but it can also arise from conflicting preferences (Jeong Min-seok·Lee Hyun-woo 2022; Jeong Han-wool 2013; Heo Seok-jae·Jeong Han-wool 2019). The analysis results of this study indicate that the probability of split-ticket voting increases as the difference in preference between the most preferred party and the second most preferred party becomes less pronounced. This result implies that conflicting preferences or attitudes influence voting behavior. Furthermore, a considerable number of split-ticket votes do not follow preferences. In-depth research on this matter is needed in the future.■

References

Park Chan-wook. 2004. “Split-Ticket Voting and the Two-Vote System in the 17th General Election: A Micro-Level Analysis of Electoral System Effects.” *Korean Political Research* 13(2): 39-85.

Yoon Kwang-il. 2014. “The June 4th Local Elections and Split-Ticket Voting: Focusing on Metropolitan Mayor and Metropolitan Council Elections.” *Journal of Korean Party Studies* 13(3): 35-67.

Jeong Min-seok·Lee Hyun-woo. 2022. “The Impact of Voters' Conflicting Party Sentiments on Split-Ticket Voting: A Case Study of the 20th National Assembly General Election.” *Korean Political Research* 31(2): 151-181.

Jeong Han-wool. 2013. “The Impact of Party Attitude Conflict on Changes in Voting Behavior: Focusing on the Analysis of Panel Survey Data (KEPS) from the 18th and 19th General Elections.” *Journal of Korean Party Studies* 12(1): 243-277.

Cho Sung-dae. 2020. “A Study on Party Preference Ranking and Strategic Split-Ticket Voting under the Quasi-Proportional Representation System: A Case Study of the 21st Korean General Election.” *Korean Political Science Review* 54(5): 125-151.

Choi Hyo-no. 2015. “Analysis of the Impact of Competitive Party Preference and Candidate Factors on Split-Ticket Voting: Focusing on the 2014 Local Elections.” *Journal of Korean Party Studies* 14(2): 55-82.

Heo Seok-jae·Jeong Han-wool. 2019. “Attitude Conflict and Split-Ticket Voting: Analysis of Split-Ticket Voting in the 20th National Assembly General Election in 2016.” *Journal of Parliamentary Studies* 137-162.

Burden, Barry C., Gretchen Helmke. 2009. “The Comparative Study of Split-Ticket Voting.” Electoral Studies 28: 1-7.

Duverger, Maurice. 1954. Political Parties. New York: Wiley.

Fiorina, Morris P. 1996. Divided Government, 2nd ed. Boston: Allyn and Bacon.

Gschwend, Thomas. 2007. “Ticket-Splitting and Strategic Voting under Mixed Electoral Rules: Evidence from Germany.” European Journal of Political Research 46: 1-23.

Kedar, Orit. 2009. Voting for Policy, Not Parties: How Voters Compensate for Power Sharing. New York: Cambridge Univiersity Press.

Schoen, Harald. 1999. “Split-ticket Voting in German Federal Elections, 1953-90: An Example of Sophisticated Balloting?” Electoral Studies 18: 473-496.

Thompson, Megan M., Mark P. Zanna, and Dale W. Griffin. 1995. “Let’s not Be Indifferent about (Attitudinal) Ambivalence.” In Richard E. Petty, and Jon A. Krosnik, eds. Attitude Strength: Antecedents and Consequences, 361-386. New York: Psychology Press.

■ Author: Lee Han-soo, Professor, Department of Political Science and International Relations, Ajou University.

■ Management and Editing: Kim Sun-hee, Senior Researcher, EAI

Inquiries: 02-2277-1683 (ext. 209) shkim@eai.or.kr

[1]Immediately after the general election, an online survey was conducted by Korea Research from April 12 to 16, targeting 1,528 individuals aged 18 and above, proportionally sampled by region, gender, and age.

[2]In the district constituency election, 49 respondents indicated they voted for "Other Parties" besides the Democratic Party, People Power Party, Green Justice Party, New Future Party, Reform Party, or independents. In the proportional representation election, 57 respondents stated they voted for "Other Parties" besides the Democratic Alliance, People Power Future Party, Green Justice Party, New Future Party, Reform Party, or Cho Kuk Innovation Party. Seventeen respondents indicated they voted for "Other Parties" in both district and proportional representation elections. As it is unknown whether they voted consistently or split their tickets, they are excluded. Excluding these individuals, there are 526 split-ticket voters (approximately 41.42%) and 744 consistent voters (approximately 58.58%).

[3]Rounded value; the vote share for the Cho Kuk Innovation Party in this election was approximately 24.25%.

[4]This could refer to a situation where a significant number of seats, such as a majority, is expected to be secured, but exceeding another significant threshold like 3/5 or 2/3 is deemed impossible.

[5]This refers to cases where, for example, votes for the presidential election and the parliamentary election are cast differently when both are held concurrently. Horizontal split-ticket voting refers to splitting votes between district and proportional representation constituencies in a parliamentary election.

[6]This classification raises the question of whether voters would split their tickets if a proportional representation party, which did not field district candidates, had done so. In other words, these voters could be classified as forced split-ticket voters rather than strategic split-ticket voters (Cho Sung-dae 2020). Due to data limitations, voters who cast a vote for an "other party" in proportional representation and a leading candidate in a district constituency are classified as 'preventing wasted votes' type. In this election, candidates from minor parties who won seats include Lee Jun-seok of the Reform Party, Kim Jong-min of the New Future Party, and Yoon Jong-oh of the Progressive Party. Among independents who competed significantly but did not win, there was Choi Kyung-hwan. Voters whose top preference is a minor party and who split their vote for it are also classified as 'preventing wasted votes' type. Finally, there are rare cases where voters' top preference is a minor party, they voted for a major party in the district constituency, but did not vote according to their preference. For example, a voter whose top preference is the Green Justice Party, who prefers the People Power Party over the Democratic Party, but votes for the Democratic Party in the district constituency. This case can be seen as an act to avoid wasted votes, but it is classified as 'other split-ticket voting' because it was not a vote according to preference. On the other hand, there is also a 3% threshold for proportional representation. Voters whose top preference is the Green Justice Party or the New Future Party might cast their proportional representation vote for a second-choice party that has a higher chance of securing seats, judging that their vote for their top choice has a low probability of translating into seats. However, it is considered difficult for voters to accurately assess this possibility, so it is not considered in the classification.

[7]Additionally, there are cases where voters chose the People Power Party in their district. This analysis includes only those instances where preference for the Democratic Party of Korea precedes preference for the People Power Party.

[8]Additionally, there are cases where voters who most prefer the People Power Party chose the Cho Kuk Innovation Party. This analysis includes only those instances where the Cho Kuk Innovation Party is the second-most preferred party after the Democratic Party of Korea.

[9]Responses of "Don't know" to the questions on district or proportional representation candidates were treated as non-responses. The total in this table may differ from the total for each question due to instances where respondents answered one candidate selection question but responded "Don't know" to another. For example, while 24.39% of respondents previously stated they voted for the Cho Kuk Innovation Party, the figure here is 24.48%.

[10]The party support in this table includes weak party support.

[11]This refers to cases where individuals stated support for the Cho Kuk Innovation Party but did not vote for it.

[12]The difference from the preceding table arises because favorability toward parties is used to categorize split-voting patterns, and some respondents did not answer this question. The party support in this table includes weak party support.

[13]The lower percentage compared to the preceding table is because respondents who voted for the Cho Kuk Innovation Party showed variations in their preferences for each party.

[14]Educational attainment was surveyed on a scale of 1=no schooling to 8=graduate school (doctoral program).

[15]The questions pertained to 'Budget Size,' 'Number of National Assembly Members,' 'Voting Age,' and the name of the 'Prime Minister.' While political knowledge measurement in online surveys may be imprecise, the measurement results show a distribution close to normal: 0=7.59%, 1=19.24%, 2=32.40%, 3=26.77%, 4=14.00%.

[16]The favorability survey asked about 'liking or disliking,' with response options ranging from '100=very favorable feeling' to '0=very unfavorable feeling.'

[17]Support for each party is measured as a binary variable.

[18]This refers to the opinion during this election that 'the Yoon Suk-yeol administration should be empowered,' with response options ranging from '1=strongly disagree' to '7=strongly agree.'

[19]Probit regression analysis is used for the model.

[20]A lower AIC value indicates a more efficient model.

[21]When calculating the predicted probabilities that follow, other variables are set to their average values.

[22]In this figure and the subsequent ones, the solid line represents the effect, and the shaded area represents the 90% confidence interval.

[23]Democratic Party of Korea, People Power Party, Green Justice Party, New Future Party, Reform Party, Cho Kuk Innovation Party.

[24]Similar results are obtained when simply using the distance between the most preferred and second most preferred party as a variable. However, the variable for conflicting emotions shows higher explanatory power.

[25]Predicted probabilities were calculated with other variables set to their average values.

[26]This refers to the influence compared to other parties and unaffiliated voters, which were not included as independent variables in the model.

[27]When the Cho Kuk Innovation Party and support for other minor and miscellaneous parties were included in the model as a single minor party variable, this variable also showed statistical significance and a positive influence on the dependent variable. However, including each party individually in the model demonstrates better model efficiency.

f2e34dd1a320d063