[EAI Opinion Review] Analysis of Swing Voters' Ballot Movement and Issue Influence

1. Panel Data and Short-Term Voter Fluctuations

Factors influencing election outcomes can be categorized into long-term factors that do not change, such as party identification, and short-term factors, such as the rise and fall of issues that momentarily dominate the election landscape. Typically, it is impossible to analyze the impact of short-term issues on elections using election survey data, which usually takes the form of cross-sectional data, because it cannot capture short-term voter changes.

For example, the question of how much the controversial remarks by candidate Kim Yong-min, which were reported by the media to have significantly influenced the election results, actually affected voters' choices is not an easy one to answer. This is because only if a person who originally supported a Democratic United Party candidate switched their support to another party after encountering that issue can it be said that the issue influenced their individual choice. Standard one-off opinion polls are merely single cross-sectional data points, so they cannot directly reveal respondent changes; they can only allow respondents to bear the burden of analysis through questions like, "What do you think was the most important issue in this general election?"

In this regard, the 2012 General Election Panel Survey has the advantage of capturing short-term voter changes. This data allows us to examine voter changes during the period through the first survey conducted in late March/early April and the second survey collected immediately after the election, and to identify the drivers of those changes. The following will analyze in detail the "changes" in voters during this general election period and the issues that influenced them.

2. The Size of the Undecided Bloc and Ballot Movement

Voters were revealed to have made their final decisions late in this general election. Approximately 39% of respondents had not decided which candidate to vote for until 2-3 days before the election, and over 60% of voters had not finalized their choice until a week before the election (see Figure 1). In fact, a significant number of voters make their final decision on ballot casting from 6 days before the election, when media predictions are prohibited, until election day, which explains why media election forecasts encountered so many difficulties.

[Figure 1] Timing of Vote Decision: 1479 Voters

Who are these swing voters, and why did they change their minds at the last moment? First, among those who participated in both surveys, 30.1% (448 people) of the 1,487 respondents who stated they would vote in the first survey were supporters of the Saenuri Party, while 29.3% (436 people) were supporters of the Democratic United Party, and approximately 29.9% (445 people) responded that they had no candidate to support or had not yet decided.

[Table 1] Change in Supported Candidate (1st - 2nd Survey)

The majority of respondents who expressed support for candidates from the two major parties maintained their support for those parties' candidates in the general election. First, the proportion of Saenuri Party supporters in the first survey who voted for a Saenuri Party candidate in the general election (85.9%) was very high, whereas the proportion for the Democratic United Party was relatively low (77.1%). This is due to the defection of Democratic United Party supporters, who defected not only to the Unified Progressive Party following the opposition alliance but also to some extent to the Saenuri Party. Second, those who had no supported candidate or had not yet decided in the first survey were generally absorbed by the Saenuri Party and the Democratic United Party candidates in a 1:1 ratio. In other words, neither party was able to attract the votes of undecided voters at the last moment, which explains why the election remained a close contest until the end.

3. How Much Impact Did the Controversial Remarks Have?: The Impact of Short-Term Issues

When voters were asked about the most important issues they considered when voting in this parliamentary election, the responses were as shown in Figure 2. Voters responded that regional development/pledges (18.0%), the controversial remarks by Kim Yong-min (17.2%), and civilian surveillance (14.5%) were the main issues considered in their voting decisions. This indicates that the controversy surrounding the remarks by candidate Kim Yong-min had a considerable impact on the election, which is consistent with reports in various media outlets.

[Figure 2] Most Influential Issue in the 4.11 General Election Vote Decision (2nd Survey)

However, strictly speaking, an issue influencing an election means that the issue influenced the behavior of voters. In other words, the influence of an issue can only be claimed if it can be proven that the issue caused voters to switch their support. Therefore, this can only be considered in the relationship between support in the first survey (late March/early April) and voting in the second survey (immediately after the election).

Table 2 classifies respondents into those who maintained Saenuri Party support between the first and second surveys (a), those who defected from Saenuri Party support to Democratic Party support (b), those who maintained Democratic Party support (c) and those who defected from Democratic Party support to Saenuri Party support (d), those who switched to the Saenuri Party (e) and the Democratic Party (f) from the undecided group, and the abstention group (g). It then presents the results of their responses regarding the most influential issues in the general election decision. This allows us to understand how the main issues considered differ according to the type of voter support change.

[Table 2] Issues Influencing General Election Decisions by Support Change Type (1st - 2nd Survey)

The results can be summarized as follows. First, the most important issues distinguishing between "defection" from and "maintenance" of Saenuri Party support were regional development pledges and civilian surveillance. The issue that decisively caused Saenuri Party supporters to defect was civilian surveillance, with over 30% of defectors citing it as the most important issue, contrasting sharply with the "maintainers" who selected it as a key issue by only 2% of respondents. Furthermore, those who considered regional development projects important tended to have solidified their support for Saenuri Party candidates, whereas those who did not consider regional development projects were highly likely to withdraw their support for Saenuri Party candidates.

Second, the issues distinguishing between "defection" from and "maintenance" of the Democratic United Party were also regional development pledges and civilian surveillance, with the controversial remarks by Kim Yong-min also showing considerable influence. "Defectors" from the Democratic United Party cited regional development pledges as the most important issue (38.3%), a stark contrast to the 13.4% among "maintainers." Similarly, the low interest in the civilian surveillance issue among defectors (8.9%) contrasted with the high interest among "maintainers" (28.4%).

The controversial remarks by Kim Yong-min were an issue that fueled significant defection among Democratic Party supporters, as predicted. 32.0% of defectors cited it as the most important issue for their voting consideration, a clear distinction from the relatively low interest of "maintainers" (13.7%). However, it is difficult to say that this was the sole issue that decisively shook the election landscape, and it cannot be said to have had a greater influence on the election than regional development pledges or the civilian surveillance issue, as shown above.

Third, if we assume that those who stated they had not yet decided on a party to support in the first survey were also influenced by short-term issues when they voted, then the third comparison pair in Table 2 also shows interesting content. What factors led these undecided voters to the Saenuri Party and the Democratic United Party? Respondents absorbed by the Democratic Party showed relatively high interest in the civilian surveillance issue (18.8%), whereas those who ultimately decided to vote for the Saenuri Party showed relative interest in the economic growth issue (18.3%). While both groups cited the controversial remarks by Kim Yong-min as the most important issue, the fact that the direction of this change was exactly opposite means that the effect should be considered offset.

Fourth, finally, the issues considered important by respondents who did not participate in the vote and abstained did not differ significantly from the overall average. However, regarding the major issues mentioned above, such as civilian surveillance or the controversial remarks by Kim Yong-min, they showed a lower level of importance compared to the national average.

4. Who Are the Defectors and Maintainers (Those Who Maintained Support)?

The following briefly describes the demographic characteristics of each group examined above. First, in terms of age, Saenuri Party "maintainers" had the highest average age at 55, while the average age of defectors was about 7 years lower, showing a clear age gap. Democratic United Party supporters had an average age of 42-43, a significant gap from the above groups. Furthermore, the party "defector" group consistently had a lower average age than the "maintainer" group, regardless of party. Abstainers were the youngest, with an average age of 36.

[Table 3] Socioeconomic Background Variables by Support Change Type

As can be predicted, Saenuri Party "maintainers" were the most ideologically conservative (6.3). The Democratic Party "maintainer" group and the Democratic Party "add-on" (undecided -> Democratic) group were the most progressive (4.6), indicating that not deciding on a candidate does not necessarily mean ideological neutrality. It is also noteworthy that the ideology of the abstention group (5.0) was to the left of the average ideological score (5.3).

Another characteristic of the "add-on" group, who had not decided on a candidate to support in the first survey, is that the proportion of women was higher than in other groups. This suggests that female respondents tend to postpone their decision on candidate support until later. The high proportion of women in the abstention group also indicates that women's voter turnout is generally lower.

The educational level of the Saenuri Party "maintainer" group was relatively low (28% with college degrees or higher). The group with the highest educational level was the Democratic United Party "add-on" group, who were undecided in the first survey but ultimately voted for a Democratic Party candidate in the election (48.4% with college degrees or higher).

The aspect related to religion is very interesting. Among those who switched their party support from the Democratic Party to the Saenuri Party, the proportion of Protestants was the highest at 34%. Conversely, among those who switched their party support from the Saenuri Party to the Democratic Party, the proportion of Protestants was the lowest at 9%. This appears to be indirect evidence that religion had some influence on the election choice, given that the controversial remarks by Kim Yong-min contained religious content. Notably, it seems that this not only acted in the common-sense direction during the election—favoring the Saenuri Party—but also had a reverse effect—defection from the Saenuri Party to the Democratic Party—centered around non-Protestants. ■