Publications

Existing research often interprets the limited impact of candidate gender on vote choice as evidence of minimal gender bias in politics. However, this overlooks the dual role of candidate gender, as both a heuristic for substantive representation and a trigger for sexism in voter decision-making. These competing mechanisms can diminish the effects of each other, obscuring the true influence of gender bias in electoral behavior. Using conjoint experiments in South Korea, a context where gender issues are highly politicized and sexism remains widespread, we examine how candidate gender affects voter evaluations in low- and high-information environments. Our findings reveal that in low-information settings, candidate gender serves as a cue for substantive representation, leading to co-sex voting among women, while simultaneously activating hostile sexism among male voters, reducing support for female candidates. In high-information settings, explicit candidate policy positions diminish the reliance on gender cues but do not eliminate gender bias. Instead, sexism manifests through opposition to gender-equity policies rather than direct discrimination against female candidates. These results suggest that information environments shape the expression of gender bias, rather than eliminating it, offering a more nuanced understanding of the conditions under which candidate gender influences electoral preferences.

To verify whether a common superordinate identity promotes intergroup social capital, we analyze survey data from three cross-sectional surveys conducted in South Korea one week before, two days after, and six months after the 2018 North Korea–United States summit in Singapore. A comparison of responses from the first and second surveys reveals that the summit positively changed South Korean natives’ altruism toward and trust in North Korean refugees by encouraging a sense of co-ethnicity among the natives. From an additional analysis of data from the third survey, we further find that the positive effects of the summit persisted even after six months.

We investigate whether the effect of rising inequality on electoral outcomes is conditional on local wealth within a country. In general, rising inequality increases support for left-wing parties among both poor and relatively well-off voters, but it also discourages turnout among those who are economically disadvantaged. As a result, left-wing parties’ electoral advantages become more salient in affluent localities, while they diminish in less affluent ones with larger shares of poor voters. To test these claims, we develop a unique measure of local inequality using actual transaction prices for residential housing in South Korea. Our analysis of aggregate data across four national legislative elections between 2008 and 2020 suggests that the effect of rising economic inequality varies depending on local wealth, generating political inequality where the policy preferences of voters in more affluent localities are better represented in the policy-making process.

Large language models are increasingly applied to political discourse, but their ability to detect culturally grounded gender sensitivity remains underexplored. We introduce KOGENT, a benchmark dataset of 1,222 transcripts from the Korean National Assembly, annotated for gender sensitivity across 6,024 utterances. Each utterance is labeled as high or low in gender sensitivity, based on contextual indicators of bias, discrimination, or inclusion, and tagged for the target group. KOGENT spans Korean legislative sessions from 1948 to 2024. Annotation reliability was ensured through dual coding and adjudication, yielding high intercoder agreement. When tasked with labeling utterances by gender sensitivity, GPT-4.1 achieved F1-scores of 87.5% (zero-shot) and 91.2% (18-shot), while GPT-4o reached 90.4% and 91.1%, respectively. While incorporating in-domain examples enhanced model performance, limitations in distinguishing between criticisms and reinforcements of inequality, culturally specific terminology, and extended contexts were observed for both models. Our results demonstrate KOGENT’s utility as a robust benchmark for analyzing gender sensitivity in Korean political speech and evaluating multilingual LLMs’ sociocultural alignment.

Why do some people express optimism about their future social mobility, while others have a pessimistic view? This paper examines whether and how local wealth inequality is associated with individuals’ pessimistic or optimistic expectations of their future social mobility in South Korea. It argues that people in districts with greater economic inequality will have more pessimistic views of their future upward mobility, as high local inequality raises concerns among the public that their opportunities to move up the social ladder may be receding. Using economic inequality at the local level calculated using actual real estate transaction prices in South Korea from 2011 to 2018, the empirical results confirm the negative association between local inequality and individuals’ assessments of their future social mobility, particularly among residents in less aff luent districts, those with low incomes, and those with a subjective awareness of belonging to a lower social class.

Heuristics are used to compensate for the limitations of human cognitive capacities. However, little is known about how voters decide what cue to use when multiple cues are available. Exploiting the institutional features of elections for the nonpartisan position of superintendent of education in South Korea, we demonstrate that voters may choose an “ecologically rational” heuristic in a given context, taking into account the trade-off between the cognitive costs and the accuracy of inference associated with different cues. Our analysis shows that the three cues—ballot order, partisan color, and ideology —are commonly used in the absence of party labels, and their relative importance varies with the electoral context. In areas with large schoolage populations and in races without an incumbent, contexts where the demand for information is conceivably high, the importance of the partisan color cue increases. We further present individual-level evidence suggesting that voters with higher levels of political sophistication and interest rely more on the ideology cue.

Are homeowners happier if housing prices increase? While many studies have examined homeownership and subjective well-being, the effect of changes in local housing prices on subjective well-being has been less studied. We argue that a change in local housing prices has divergent effects on the subjective well-being of homeowners versus renters, and these effects vary with the housing market cycle. Using data on the annual change in the average transaction price for residential housing in local districts of Seoul, South Korea from 2008 to 2018, we find that homeowners have higher subjective well-being than renters, and the gap between homeowners and renters grows as housing prices increase, mainly because renters become less happy while homeowners’ happiness varies with overall housing market conditions in their city. Homeowners’ subjective well-being in­ creases only when housing prices go up in their own neighborhood while the citywide housing market remains stable. During an escalating housing market, increasing prices make both homeowners and renters less happy.

Research on the effect of candidate sex on vote choice has tended to find that, even when voters state preferences for candidates of their own sex, party identification tends to win out on election day. However, not all elections present a clear partisan choice for voters: primary elections in the United States, intra-party candidate selection in Australia, and municipal elections in a range of jurisdictions either pit intra-party candidates against each other or provide only weak partisan cues. In this paper, we use a conjoint experiment to directly compare the effects of candidate sex and partisan affiliation on voters’ preferences: in one context where partisan affiliation is constant (e.g. a primary contest) and a second where partisan affiliation varies (e.g. a general election). From a probability-based sample of Australian voters, we find left-identifying female respondents tend to prefer female candidates regardless of the candidate’s partisan affiliation and electoral context. By contrast, right-identifying male voters prefer male over female candidates in intra-party contests between right-affiliated candidates, suggesting that conservative men are the least supportive of female candidates. As conservative men dominate Australia’s current governing parties, we argue the preferences of this demographic inhibits the advancement of female politicians.

How does the descriptive representation of ethnic minorities affect how voters feel about the responsiveness of government? While there are many theoretical arguments that descriptive representation increases perceptions of legitimacy, the empirical evidence of this link is limited. We use survey data from the Australian Election Study and a separate conjoint experiment to evaluate whether the presence of ethnic minority candidates changes voters’ perceptions of government responsiveness. We find ethnic minority Australians do not appear to have higher levels of external efficacy when voting for an ethnic minority candidate. By contrast, white-Anglo respondents have lower levels of external efficacy when voting for a non-Anglo candidate. The results inform the continuing debate on how group consciousness affects political behavior.

How does local economic inequality affect the native-immigrant gap in immigration attitudes? Existing studies do not distinguish between native and immigrant citizens, which is problematic because immigrants represent an increasing share of the population and voting public. Immigrant citizens, as legal residents, receive the same legal and social protections as native citizens. However, as an out-group, they are less likely to be attached to the national and cultural identity of a host country. This paper uses the Australian Election Study to show that immigrant citizens prioritise cultural or psychological considerations in forming immigration attitudes. As local economic inequality rises, immigrant citizens’ support for immigration strengthens regardless of their country of origin, reason for migration and length of stay in Australia.

The effects of verbal accents on intergroup attitudes are well documented. This study aims to enrich our understanding by exploring how those effects vary according to the speaker’s gender and the political context. We conducted two online survey experiments in which South Korean citizens were randomly exposed to speakers exhibiting one of four accent conditions – South Korean male and female accents and North Korean male and female accents – a week before and two days after the 2018 Singapore summit between North Korea and the United States, in order to test hypotheses based on literatures from political science, social psychology and evolutionary biology. The results indicate that only exposure to a North Korean male accent, not a North Korean female accent, strengthened stereotypes about North Koreans among South Koreans prior to the summit. Further, this negative effect disappeared immediately after the summit.

This paper examines how economic inequality at the local level affects individuals’ subjective well-being (SWB) through social comparison in Seoul, South Korea. We implement a multi-level analysis combining asset inequality, calculated using the actual transaction prices of apartments, and public opinion surveys conducted by Seoul Metropolitan City between 2008 and 2016. Our analysis shows that inequality negatively affects SWB among respondents whose family income is lower than the median (the envy effect), but drives up SWB among the other half (the pride effect). Further analysis on the effect of inequality on subjective class awareness corroborates a social comparison mechanism: the haves embrace an upper-class awareness as local inequality increases, while the have-nots embrace a lower class awareness. These findings suggest that, despite concerns about economic inequality and its negative consequences, calling for policy reform to reduce inequality may be unpersuasive to the better off members of society, especially if doing so increases individual financial burdens.

Studies suggest that rainfall on an election day benefits the Republican Party in the United States and conservative parties in Western Europe. A common explanation is that marginal voters, whose turnout decisions can be affected by weather conditions, are more likely to be supporters of liberal parties. This paper shows that bad weather on an election day instead benefits the liberal parties in South Korea, where election days are designated as special holidays. Young voters, who tend to prefer liberal parties, often plan other activities whose feasibility is contingent on good weather conditions; as a result, they are more likely to turn out, and thus to provide additional electoral support to liberal parties, when weather is bad.

This paper examines how South Korea’s 2016 corruption scandal and subsequent presidential impeachment affected voter decisions in that country’s 19th presidential election. Both aggregate and individual-level analyses indicate that the landslide defeat of the country’s conservative party does not portend a fundamental shift in voter-party alignment. At the aggregate level, we find that the normal vote share of the conservative party in the 19th presidential election was similar to that of the 18th. At the individual level, an analysis of the 2017 Korean Election Panel Studies data demonstrates that regional and generational cleavages are influential factors in vote switching by majorparty supporters. Aside from their attitudes regarding the impeachment, those who changed their votes from conservative parties generally hold similar issue preferences to those who did not change. Therefore, once the political salience of the impeachment issue wanes, it is questionable whether the current disunity will end.

We study the poor’s psychological motivations to oppose income redistribution, relying on social identity and system justification theories. We find that national identification reduces differences between the poor and the rich in terms of attitudes toward income redistribution and self-esteem, by discouraging the poor from supporting redistribution but encouraging them to acquire greater self-esteem. Next, fatalism reduces and increases differences between the poor and the rich in terms of attitudes toward income redistribution and self-esteem, respectively. Yet, a closer look reveals that the responding patterns of the poor and the rich to fatalism are consistent only with the mechanism behind the prediction that concerns self-esteem. That is, fatalism increases support for income redistribution in both groups, whereas it reduces self-esteem only among the poor.

In this paper, we investigate the effect of incumbency in three different electoral settings in South Korea using a regression discontinuity (RD) design. We claim that incumbency provides advantages in local elections, but its positive effects diminish in national legislative elections. Local politicians have an opportunity to develop close ties with their constituents, insulating themselves from national politics; but, due to the centralized nature of national politics, members of the National Assembly do not. Consistent with our expectation, incumbency effects are generally positive in local elections, but insignificant or even negative in national legislative elections. We also find that politicians who previously held local elected office enjoy an incumbency advantage in national legislative elections.

Despite strong theoretical claims that politicians should target swing voters with distributive benefits, empirical evidence in the United States is inconclusive. This paper addresses the puzzle by focusing on two factors overlooked in previous work. First, I show that, owing to the bimodal distribution of partisanship among the U.S. public, swing voters can be targeted efficiently through the allocation of federal resources to areas where the opposition is strong. Secondly, I hypothesize that presidents limit swing-voter targeting to times when they are actually up for reelection; thus, the opposition county advantage appears only in first presidential terms. An analysis of the geographic distribution of federal project grants awarded between 1986 and 2009 supports the theory. Presidents target swing voters within competitive states; they do so only in the years when they seek reelection, however, and they channel benefits to counties where the opposition party maintains a stronghold.

Existing studies—most importantly, Gomez, Hansford, and Krause—provide empirical support for an idea often embraced by popular media: The vote share of the Republican Party (as the percentage of total votes) increases when it rains, because the magnitude of decrease in turnout is larger among Democratic vis-à-vis Republican supporters. Considering the compositional nature of aggregated data, we show that the alleged Republican advantage derives in part from an increase in the number of votes for the Republican Party. Based on the extensive literature of psychology and related fields, we provide a possible interpretation of this counter-intuitive empirical finding. Methodologically, our evidence suggests that researchers must be alert when using rainfall as an instrument to estimate the causal effects of voter turnout on electoral outcome.

We examine (1) how trust in foreigners and trust in political institutions affect attitudes toward immigration and (2) the moderating effect of economic development on those impacts, analyzing data from the fifth wave of the World Values Survey. We find that natives who trust foreigners are more tolerant toward immigration and that economic development positively moderates the impact of trust in foreigners on the attitudes. Meanwhile, we find only mixed evidence for the impact of trust in political institutions and the moderating role of economic development in the impact. We conclude by discussing the implications of the findings.

In this paper, we examine the extent to which wartime violence against civilians during the Korean War affects people’s current attitudes toward South Korea and other involved countries. Using a difference-in-differences (DID) approach that compares the cohorts born before and after the war, we find that direct exposure to wartime violence induces negative perceptions regarding perpetrator countries. As many of the civilian massacres were committed by the South Korean armed forces, prewar cohorts living in violence-ridden areas during the war demonstrate significantly less pride in South Korea today. In contrast, postwar cohorts from those violent areas, who were exposed to intensive anti-communist campaigns and were incentivized to differentiate themselves from the victims, show significantly greater pride in South Korea, and greater hospitality toward the United States than toward North Korea, compared to prewar cohorts in the same areas and to the same cohorts born in non-violent areas.

How does wartime violence affect public attitudes toward the government in the long run? In this paper, we examine whether violence against civilians during the Korean War continues to influence people’s attitudes toward the South Korean government more than half a century later. We find that wartime violence has clear long-term attitudinal effects. Using a difference-in-differences analysis that compares the cohorts born before and after the war, the findings indicate that people who experienced violence in their childhood (0–5 years) are less supportive of the South Korean government, especially the administration and the military, compared with those born in the same areas during the 5 years after the war. We argue that the gap between pre- and post-war cohorts is generated by the long-lasting trauma of wartime violence and the social stigma imposed on violence victims after the war.

Regional bloc voting in South Korea has been ascribed to voters’ psychological attachments to birthplace. This article seeks to expand the existing discussion of regionalism by showing that economic conditions in voters’ places of residence affect vote choices at the individual level and produce clustering of votes at the aggregate level in South Korea. While the idea of residence-based regionalism has previously been suggested, empirical scrutiny of the idea has been limited. Exploiting a Bayesian multilevel strategy, this article provides evidence that short-term economic changes at the province level affected voters’ choices in the 2007 presidential election in South Korea, independent of the long-term political affiliation between regional parties and their constituents. The positive association between local economic conditions and vote choices remains significant, controlling for perceptions of national economic conditions and other individual level covariates such as age and political attitudes.

Despite strong theoretical claims that politicians should target distributive benefits to swing voters and competitive districts, the empirical evidence is mixed. This paper resolves the inconsistencies by focusing on the time-varying incentives of an incumbent government. To the extent that election-motivated behavior entails directing government resources to marginal voters and constituencies, this behavior can be expected to peak in the period just prior to an election. An analysis of subsidy allocation in South Korea provides evidentiary support for this claim. In general, more subsidies are allotted to incumbents’ core municipalities; however, before legislative elections, municipalities with close legislative races receive greater share of subsidies.

Does direct experience in election administration change perceptions of election fraud? Experiential learning and procedural justice theories predict that direct experience will reduce fraud perceptions, while motivated reasoning theory predicts that prior beliefs will resist change. This study analyzes pre- and post-election survey data from 16,960 individuals who performed or observed election management tasks in the 2025 South Korean presidential election. First, fraud perceptions declined after direct experience, with approximately a 9 percentage point drop even among losing party supporters. Second, partisan differences persisted but were smaller than those in general public surveys. Third, among initial fraud believers, losing party supporters maintained their beliefs at much higher rates compared to winning party supporters, confirming losers’ resistance. Fourth, list experiments revealed that a substantial number of losing party supporters who claimed to believe in election fraud may not actually hold such beliefs, suggesting partisan cheerleading.

This study analyzes how South Korean citizens’ attitudes toward democracy have changed amid growing concerns over democratic backsliding following the imposition of emergency martial law in 2024 and the subsequent presidential impeachment. Analyzing rolling cross-sectional survey data from 2003 to 2025 and panel survey data collected before and after the 2025 presidential election, the findings show that diffuse support for democracy has steadily increased over the past two decades, reaching 76% in 2022, and remained stable even during the martial law and impeachment crisis. Meanwhile, satisfaction with democracy varied with electoral winner-loser status, but this does not represent a fundamental shift in attitudes toward the democratic regime itself. These results suggest that the recent crisis of Korean democracy is not “backsliding from below” driven by the erosion of citizens’ democratic values, but rather “backsliding from above” driven by political elites.

Democratic theory argues that elections function as a core mechanism that institutionalizes uncertainty in the competition for power, thereby peacefully expressing and managing conflict while enabling winners and losers to continue competing. However, given that deepening political conflict in contemporary Korean society is being intensified by inter-party competition for electoral victory and backlash from losers, it is questionable whether elections properly fulfill their function of institutionalizing conflict. In this context, this study analyzes the impact of elections on affective and ideological polarization among Korean voters using panel survey data collected before and after the 2025 South Korean presidential election. The analysis reveals that while affective polarization decreased on average, this stemmed not from the “generosity” of the winning camp but from a decline in in-group favorability among the losing camp. Hostile attitudes toward out-groups showed no significant change. In contrast, voter-perceived ideological polarization between parties decreased. Furthermore, a cognitive convergence (projection) effect was observed among the majority of voters, with individuals adjusting their self-placement closer to the winning party (Democratic Party of Korea). These findings simultaneously demonstrate both the potential and limitations of elections as a mechanism for reducing political polarization in South Korea.

According to economic voting theory, improved economic conditions increase the incumbent’s chances of victory. However, in the 2024 U.S. presidential election, despite improved macroeconomic indicators, public perceptions of the economy remained negative, and Democratic candidate Harris lost to Republican candidate Trump. To explain this discrepancy, we analyze how economic changes and racial composition at the county level affected Democratic vote share. Our findings reveal that the impact of per capita income growth on Democratic support varied significantly by racial demographics. In predominantly white counties, income growth corresponded with increased Democratic support, but this effect weakened as white proportion decreased. In counties with non-white populations above a certain threshold, income growth actually led to decreased Democratic support. These results suggest that when racial composition intensifies intergroup tensions, the connection between economic conditions and voting behavior may weaken, potentially undermining democratic accountability mechanisms based on economic voting.

Our study analyzes the impact of ‘perceived ideological polarization,’ an aspect that has received little attention in previous literature on affective polarization in South Korea. We conducted both group-level and individual-level analyses to explore this phenomenon. Our group-level analysis shows that about 35% of the polarization perceived by partisan voters stems from misperception. Specifically, these voters tend to view the issue positions of supporters of both the opposing party and their own party as more extreme than they actually are. At the individual level, our analysis reveals an interesting finding. It is not the actual polarization (measured by the distance between the respondent’s and the opposing party supporters’ actual positions) that promotes affective polarization. Rather, it is the perceived polarization (measured by the distance between the respondent and the perceived position of opposing party voters) that drives this effect. These results suggest that a significant portion of affective polarization could be reduced simply by accurately perceiving the ideological positions of in-group and out-group party supporters.

In South Korea’s Education Superintendent elections, where political parties do not nominate candidates, the order or symbol of candidates listed on the ballot can influence voters’ choices, which has been a persistent issue. However, existing studies that demonstrate order effects have limitations because they rely on aggregate-level data, such as electoral districts, which do not fully address the problem of omitted variables and fail to confirm order effects at the individual level. To overcome these limitations, this study uses a survey experiment to analyze whether order effects exist in Education Superintendent elections, what form they take if they do, and which voters are most affected by them. The findings can be summarized in three main points. First, simply being listed earlier on the ballot does not necessarily increase a candidate’s vote share. Second, the choice made in partisan elections held simultaneously with the Education Superintendent elections tends to favor the candidates listed in same positions. This suggests that the perceived advantage of being listed in a earlier position in Education Superintendent elections is not due to the order itself but rather because candidates from the two major parties are usually given top positions in partisan elections, influencing voting choices in Education Superintendent elections. Lastly, the impact of choices made in partisan elections on the Education Superintendent candidate choice is primarily observed among voters with low levels of political knowledge.

This paper analyzes the election day weather effects in South Korean local elections. Analyzing the election day precipitation and temperature data estimated at the town level, along with the results of the 7th and 8th provincial council proportional representation elections, we found that for every 1sd (3.1mm) increase in election day precipitation, the vote share of progressive parties increases by 1.8%, while that of conservative parties decreases by about 2.7%. Unlike in national assembly elections, despite the introduction of the early voting system, election day weather still influences election results. Furthermore, compared to the precipitation effects observed in national assembly election results before the introduction of the early voting system, the effect of election day weather in local elections appears to be larger.

Despite concerns about the increasing affective polarization among Korean voters, there has been a lack of empirical evidence to support this claim. This study examines whether affective polarization as a ‘process’ is occurring among Korean voters, exploiting voter opinion polls from six National Assembly elections conducted since 2000. The analysis at both the aggregate and individual levels failed to find evidence that affective polarization is intensifying in Korea. At the aggregate level, both the partisan based polarization index and the Wagner polarization index which includes non-partisans, peaked in 2004, followed by a decreasing trend until 2016. There was an increase in emotional polarization between 2016 and 2020, but further analysis in 2022 revealed that this increase is not continuing. At the individual level, we reconfirmed the significant correlation between affective polarization and factors like strong partisan identity, ideological intensity, and perceptions of ideological polarization within parties; however, this study did not find evidence that these variables are exacerbating affective polarization.

Despite the widespread interest in the effects of weather on elections and the effects of the early voting system, how the early voting system affects the election-day weather effects remains as lacuna. This paper argues that the system mitigates the effects of election-day weather on election, by providing voters with the opportunity to avoid the weather-related costs of voting. An analysis of election-day precipitation and temperature data at the regional level and 18th-21st Korean legislative election results corroborates empirical evidence. The results show that election-day rain increased the vote share of liberal parties and decreased that of conservative parties when the early voting was not available, whereas these effects disappeared as the early voting system was introduced. Relatedly, the election-day rain increased the turnout of those aged 20-24 and decreased turnout of those over 50, whereas these effects vanished with the early voting system. The findings suggest that voters who are sensitive to the direct costs and opportunity costs of voting are taking advantage of the system, based on their anticipation of election-day weather.

How do candidates’ appearance affect election outcomes? Under bounded rationality, voters use a variety of heuristics in their political decision-making. Previous studies have found mixed evidence on whether candidates’ appearances, particularly their smiles, serve as heuristics and influence election outcomes. In this regard, this study investigates the effect of smiles on vote shares using election posters of candidates running in South Korean local elections. The candidate’s smile, as measured by Microsoft Azure, increased the vote share by 1.5 to 2.6%p. The size and statistical significance of the smile effect varied depending on the electoral competitiveness and number of candidates. Overall, this study shows that candidates’ smiles help them win votes in specific electoral contexts.

Underrepresentation of women in politics is a widely recognized problem in Korea, yet empirical studies on the effect of women’s descriptive representation on female political participation are nonexistent. This research analyzes the effect of the number of female candidates and electees during the 2nd~7th Korean local elections on female turnout, revealing that female turnout increased when more female candidates ran for office and more female candidates were elected in the previous election. According to an analysis based on age, the effect of increased female representation on female turnout is most profound amongst female voters in their late 20s to 40s but not as substantial for women in their early 20s or 50s and beyond. Based on results by elections, increased female turnout due to greater representation initially appeared in the 5th election, when mandatory gender quota laws were implemented. Moreover, the effect of the number of elected females on female turnout is larger than that of the number of female candidates. These results show that an increase in female candidates and electees can lead to an increase in women’s political participation.

How does the candidate’s sex affect voters’ choice? Previous studies in South Korea have shown that the candidate’s sex does not affect voters’ choice. However, the analysis of the observational data such as aggregated election outcomes or survey data is limited in examining the causal effect of a candidate’s sex on voters’ choice. The effect of candidate sex may be offset from each other when voters engage in the cosex voting, where male voters vote for male candidates and female voters vote for female candidates. The survey data analysis, on the other hand, suffers from the selection issue given that only a small number of qualified female candidates run for elections. In this sense, this study examines the causal effect of the candidate’s sex on voters’ choices through a conjoint experiment. The results suggest that cosex voting occurs among college students, which continued even after reflecting on the effects of political parties.

The immigration level in Korea has been increasing amidst worsening individual attitudes toward immigrants. Does this mean that the increase in immigration levels leads to a negative public attitude toward immigrants? Intergroup Contact Theory supposes that a larger migrant population leads to positive attitudes: frequent contact alleviates intergroup prejudice and improves relations. This paper suggests that there can be both ‘good contact’ or ‘bad contact’ depending on local contextual characteristics, such as inequality. This paper reviews the effect of local economic inequality and local immigrant stock on public attitudes by matching the Seoul Survey dataset(2010-2017) with originally generated local economic inequality levels. With a multi-level time-series data of twenty-five local level districts, six-year spans, and 274,514 individual observations, we find that local districts with larger immigrant populations and lower inequality levels are more likely to exhibit positive attitudes, and the positive effect of the immigrant population on public attitudes diminishes as the inequality level increases. These findings suggest that the deteriorating net level of public attitudes toward immigrants is attributable to economic inequality rather than the size of the immigrant population.

This paper presents the results of ‘The Survey on the 21st National Assembly members’ conducted by the Korean Association of Party Studies (KAPS) and the Hankyoreh following the 21st Korean National Assembly Election. Since the 16th Korean National Assembly Election, the KAPS has surveyed the candidates and/or the elected regarding their views on major policy issues and perceived self-ideology, which has contributed to our understanding of overtime changes in ideology of political parties in South Korea and their members. This year’s survey includes 21 questions on the four major policy issue areas including foreign policy, economy, social issues and cultural issues as well as their perceived ideology. Among the 300 elected, 197 participated in the survey. The results suggest that the Justice Party is most liberal, the United Future Party is most conservative, and the Democratic Party is in the middle on average in terms of issue preference and perceived ideology. Compared to the preceding National Assembly, the partisan gap continues to appear salient in foreign policy, economy, and the cultural issues. In contrast, the gap narrows down in the social issues because the members of the Democratic Party embrace more conservative preference. It is noteworthy to examine whether this shift leads to cooperative decision making on social policies between liberal and conservative parties in the upcoming National Assembly. The composite policy preference index of individual assembly members, on the other hand, shows significant difference among members of different parties. Political parties in South Korea has evolved from a group of people from the same region into a group of people with distinctive policy preferences.

This paper analyzes the election-day weather effects on turnout and electoral outcomes in South Korea. Existing studies on the importance of election-day weather from developed western democracies generally consider poor weather an additional cost in the voting calculus, which reduces turnout. In this paper, I claim that we need to take the opportunity cost of voting into account in the country like South Korea, where an election-day is designated as a special holiday. Voters who plan leisure activities on the election-day are more likely to turn out in bad weather than in good weather. Given that young people often have stronger preference to leisure activities and affinities with liberal parties, moreover, bad weather in an election-day results in more favorable electoral outcomes to liberal parties than conservative parties. An analysis of three legislative elections since 2004 shows that rainfall increases the number of votes for liberal parties, but decreases the number of votes for conservative parties. However, rainfall does not produce any significant change in turnout because the changes in votes between liberal parties and conservatives parties cancel out each other.

Can generative AI stand in for human respondents when judging faces? We answer by decomposing human–AI agreement across a hierarchy of judgments—from perceptual labels that can be read off a face, to social traits that must be inferred, to electoral outcomes that can only be predicted—using 2,601 South Korean local-election candidate images. We compare 16,890 evaluations from 1,689 human respondents with image-level ratings from two current vision-language models, Claude Sonnet 4.6 and GPT-5.5, each evaluating the images in survey-like ten-image sessions that mirror the human protocol. Agreement is high for gender typicality (r ≈ 0.73), attenuates for competence, attractiveness, and trustworthiness (r ≈ 0.18–0.38), and degenerates for electoral prediction, where neither humans nor AI exceed the base rate (r ≈ 0.09 with actual outcomes). This gradient survives a three-way reliability adjustment and is not accounted for by facial-geometry controls, which explain human and AI trait ratings about equally poorly. Both models rate candidates more positively than humans, most so for candidates humans rate low, although this bias is smaller in the newer model; and AI ties competence and trustworthiness to electability far more tightly than humans do. The findings bound algorithmic fidelity in the visual domain: AI most faithfully reproduces what humans read, less what they infer, and least what they predict.

As claims of election fraud spread across democracies, official rebuttals routinely fail, filtered by partisan priors whether they come from media, officials, or experts. We study a different kind of evidence, experiential information, what a citizen learns by watching for herself. A correction can be dismissed as coming from a biased source. Direct observation cannot, because the citizen is her own source. A random lottery assigned citizens to observe vote counting in South Korea’s 2025 election, letting us identify the causal effect, free of selection bias. Confidence in the election commission rises, with the largest gains among initial skeptics. But the gain stays specific to what they witnessed, not spreading to other institutions, unobserved domains, or democratic attitudes. This points to learning rather than empowerment or civic identity, which we measure directly and find too small to account for the gain. Observation answers the specific procedural doubts that fuel fraud claims, where electoral confidence is most contested.

Although youth is broadly valued in contemporary society, the political domain rewards experience and maturity, leaving the electoral role of youthful appearance ambiguous. This study examines how candidates’ babyfaced appearance influences voters’ trait evaluations, electability perceptions, vote choice, and actual electoral outcomes. Unlike prior research assuming a uniform directional effect of babyfacedness, we argue that each stage of voter judgment is a distinct cognitive task: electability is a social inference about how others will respond, whereas vote choice is a personal evaluation, and babyfacedness may operate in opposite directions across the two. Analyzing 2,601 candidate images from the 6th, 7th, and 8th South Korean basic-level local council elections (2014, 2018, 2022) evaluated by 1,689 respondents, we find that babyfacedness depresses trait evaluations of competence, trustworthiness, and attractiveness yet raises electability perceptions, while reducing vote choice and actual electoral outcomes. Heterogeneity by candidate age reveals contrasting patterns: the positive effect on electability remains stable across candidate age, whereas the negative effects on trait evaluations and vote choice are amplified for younger candidates. Heterogeneity by candidate gender shows that the outcome-stage penalty is concentrated among female candidates and absent among males, even though no comparable gender heterogeneity appears at the individual-judgment stages. We reframe babyfacedness as a cue whose political meaning depends jointly on the judgment task at hand and on the candidate’s category position, offering a unified account of mixed prior findings.

Existing research shows that voters in established democracies often fail to penalize politicians who violate democratic norms, as many are willing to trade democratic principles for partisan interests in polarized environments. Yet such partisan double standards remain underexplored in East Asia, where countries share authoritarian legacies, face ongoing risks of democratic backsliding, and operate under the geopolitical shadow of powerful authoritarian neighbors. Using a large-scale vignette experiment conducted in Japan, South Korea, Taiwan, and the United States (N = 10,194), this study examines whether voters sanction politicians who engage in democratic norm violations. We focus on reactions to three types of undemocratic statements: election denial, protest suppression, and attacks on the judiciary. The results show that voters in East Asian democracies impose stronger penalties than their American counterparts, though the degree and pattern of accountability differ across cases. Election denial consistently generates the strongest sanctions, while Japan and Taiwan show robust reactions across all scenarios. South Korea displays comparatively modest responses to protest suppression, and the United States shows the weakest effects overall. These findings highlight both the resilience and the heterogeneity of democratic accountability in East Asia and underscore the need to understand regional variation when assessing the prospects for democratic stability.

When women’s share in legislatures rises, how do the men who remain a majority respond? Using approximately two million standing-committee speeches from the 16th through 21st Korean National Assembly (2000–2024) classified by large language models, we build a male legislator committee half-term panel (N = 5,040 rows for 1,531 male legislators across 199 committee half-term cells) and decompose male gender-related speech along two orthogonal axes: the stage of representation (extensive participation vs.\ intensive sensitivity) and the within-cell aggregator (binary max vs.\ continuous mean), yielding four dependent variables. The main specification is a linear probability model with a 42-committee random intercept and half-term fixed effects. Committee-level women’s share significantly raises male gender speech on all four margins—both whether men engage gender topics and how sensitively they speak. The share of past-half-term critical actors who are now in the committee independently raises the participation extensive margin, with weaker positive effects on the sensitivity margins; the joint structure of CMT and CA is additive rather than multiplicative. Thresholds differ structurally: a binned LPM with 5-pp bins locates the first qualitative jump at the 20% boundary, while critical actors saturate early at 5–10%. The Korean case functions as a stringent test of theoretical portability: spillover dominates specialization and group threat despite a below-OECD-average women’s share, consistent with the Korean institutional context—biennial committee reassignment and strong party discipline—which prevents women from being long-term locked into gender-agenda committees and thus suppresses the specialization channel

This study examines how international naming and shaming influences domestic attitudes toward refugees in South Korea, a country with historically low refugee acceptance rates. Through a survey experiment involving 2,901 adults, we examined whether criticism from the U.S. Department of State and Amnesty International influences public perceptions of refugee policies. Our findings show that international criticism significantly increases both negative evaluations of current refugee policies as well as support for more inclusive approaches. We also found that shame serves as an important mediator in these effects. Notably, the impact of naming and shaming varied depending on the source of criticism. Criticism from Amnesty International elicited a stronger sense of shame than that from the U.S. State Department, suggesting that international NGOs may be more effective messengers owing to their perceived normative legitimacy. These results advance our understanding of international norm diffusion by highlighting how external criticism can influence domestic attitudes.

What does it mean to be a “good citizen” when partisan hostility threatens democratic functioning? This study examines whether correcting distorted beliefs about political opponents can reduce affective polarization and strengthen support for core democratic norms. We focus on two types of misperceptions: (1) inaccurate beliefs about the ideological positions of opposing party supporters and (2) inaccurate meta-perceptions—how people think opposing party supporters view their own group. In a three-wave panel survey experiment conducted in South Korea, respondents received individualized, data-based feedback comparing their beliefs to contemporaneous survey benchmarks. Across both correction conditions, out-party favorability increased sharply in the immediate pre–post window in Wave 2 (roughly a dozen points on a 0–100 scale) and affective polarization declined, but these shifts did not differ systematically by the direction or type of corrective message. In contrast, democratic-norm attitudes (electoral legitimacy, rejection of political violence, pluralism, and minority-rights protection) showed little short-term movement. In the follow-up wave (Wave 3), most affective changes attenuated substantially, suggesting limited durability; however, meta-perception feedback showed evidence consistent with a delayed buffering pattern for democratic norms relative to the accurate-perception group. By examining affective and democratic outcomes and assessing whether any effects persist over time in a non-U.S. setting, the study contributes to debates on belief-accuracy interventions and democratic resilience.

Voters form rapid impressions from candidate faces, and these impressions carry electoral consequences. This study examines whether gender typicality—the degree to which a face displays features prototypical of one’s biological sex—affects competence evaluations and electoral outcomes in South Korea. Drawing on two surveys of higher-level elections (5,576 candidates) and lower-level local council elections (2,602 candidates) linked to official election returns, I find that gender typicality enhances women’s perceived competence in both studies, contradicting the linear masculine premium of Sex Role Theory. In high-information elections, the femininity–competence relationship is curvilinear (extreme femininity penalized) for female candidates evaluated by male respondents; in low-information elections it is monotonically positive. Yet only in low-information elections does gender typicality shape actual vote share, where an inverted-U pattern survives party controls. The effect does not vary by candidate ideology, favoring Gender Typicality Theory over Role Congruity Theory among the theories predicting a curvilinear relationship. The information environment is a boundary condition for the electoral relevance of facial gender typicality.