The Contingent Effects of Candidate Sex on Voter Choice

Yoshikuni Ono and Barry C. Burden

Although women constitute a majority of the American electorate and vote at higher rates than men, women hold only about 20% of seats in Congress and no woman has been elected president. The reasons for the underrepresentation of women are many: differences between men and women in resources, political networks, tolerance for political campaigns, and self-confidence. A nagging question that has been obscured by these factors is whether voters are biased female candidates running for national office. Our study addresses this question directly using experiments to determine if voters prefer a male candidate over a female candidate when other “upstream” differences between the candidates are neutralized.

Real world data make it difficult to determine if voters are biased for or against candidates of different sexes. This is because voters often infer other characteristics of candidates from visible traits such as sex. For example, absent other information, a voter might assume that a candidate who is female is more liberal and shows weaker leadership skills compared to a male counterpart. This is a simple application of broad stereotypes about the differences between men and women. Relying on these inferences, such a voter might disfavor the female candidate not because of sex per se but because of a desire for a candidate who is less liberal or has more leadership skill. If these inferences could be disrupted with additional information, does voters still show biases based on sex?

To distinguish between these kind of stereotype-based decisions and more fundamental biases, we conducted “conjoint” experiments. These experimental designs are often using in marketing studies to determine how consumers respond to different combinations of product attributes. In a similar fashion, we present respondents with opposing candidates who possess different combinations of attributes. In a nationally representative survey fielded in early 2016, we showed respondents pairs of hypothetical candidates for Congress or President. Each candidate had 13 different attributes – things such as age, race, personal style, issue positions, of course sex – randomly selected. But running the experiment many times we could estimate the effect of a candidate’s sex on voting patterns.

We find that the effects of candidate sex are highly contingent in at least two important ways. First, respondents were slightly biased against female candidates running for President (about 2.4% on average) but not against women running for Congress. We theorize that this difference between offices could be due a lack of experience with women in the White House compared to the more regular exposure to female legislators (including a former Speaker of the House). At least some voters appear reluctant to imagine how a woman would behave in an office only held by men. The bias against female presidential candidate might also reflect some residual hesitation about women serving as Commander-in-Chief due to the national security responsibilities that have traditionally been associated with men.

Second, there is a difference in how women fare in primary elections versus general elections. Our experiment provides evidence on this because half of the candidate pairs are from the same party while the other half are from different parties. In the general election setting, self-described Democrats or Republicans show little bias. These individuals instead rely on partisanship as the main decision heuristic as a powerful guide to which candidate is best. In contrast, independents display the largest bias against female candidates; this reflects the absence of party as a form of “insurance” about what a politician would do in office. The situation is different in primary settings, where Republicans and independents both display some bias against female candidates. Our study thus provides some of the first experimental evidence that equivalent male and female candidates get different responses even from the same voters depending on whether other informational cues are available. Candidate sex does not have one effect but many depending on the type of election, characteristic of the voters, and real-world experiences with women in office.


Negative Descriptive Social Norms and Political Action: People Aren’t Acting, So You Should

Hans Hassell And Emily Wyler

You get an email in your inbox: “No one is doing anything about this issue: we need you.” It’s a seemingly desperate call-to-action, a last-ditch plea to join a cause. But do you feel inclined to pitch in, to help out where it seems your fellow citizens don’t bother?

As social animals, we seek normative information when trying to understand how to direct our energies and act in socially-acceptable ways. Political calls-to-action often take advantage of this predilection, to motivate people to act based on a description of what others are doing… or not doing.

It’s that latter flavor of social normative information–the framing that indicates what others are not doing–that we were curious about. Political campaigns and activist organizations sometimes request action from a message recipient by describing the undesirable action (or inaction) of a group and pushing the individual to deviate from that norm by joining the cause. This type of normative framing is contentious, however; previous social behavior research on eliciting political behaviors claims it is an ineffective tactic. The argument is that the only way to truly inspire action is to formulate messaging that highlights a highly popular, successful campaign that could benefit from you joining the herd!

In our article, we challenge the literature’s dismissal of negative descriptive social norms in favor of positive descriptive norms. In fact, mediating factors such as the context, audience, and the underlying motivations to participate may be more important to consider when deciding how to frame call-to-actions in the political sphere.

With an online survey and a separate field experiment, we presented individuals with group-level messaging about community issues that either highlighted a positive descriptive social norm (“lots of people are working on this”) or a negative descriptive social norm (“no one’s doing anything about this”). We saw that those that received the negative descriptive norm messaging were more likely to indicate willingness to sign a petition or write to a policymaker in response to these call-to-action pleas. Respondents also reported higher feelings of anger after reading about what others were failing to do for an issue—and this difference was most notable among those not already active in the political sphere. Someone that has a lower propensity to engage in political activity—perhaps a non-activist—may feel more anger when introduced to the unfamiliar idea that others just don’t care about certain issues, which improves the saliency of the call-to-action in their mind and instigates action.

While behaviors measured by similar research, such as voting, hold a more intrinsic level of incentive, the actions requested in these message campaigns are not as self-serving.  Using negative descriptive social norms may be helpful when trying to drive purposive behaviors, which indicate more of an ability on the part of the individual to influence change for the “greater good” of the community.

In an age that may seem to foster a culture of apathy and disengagement, finding ways to inspire increased levels of participation of individuals in the political process and policy outcome arena is undoubtedly invaluable. Refraining from writing off the potential efficacy of negative social descriptive norms opens doors to a more diverse toolbox that can be accessed by leaders and advocates in efforts to motivate action. Our messaging may need to humbly highlight the failures of the group, rather than just the successes, in order to inspire anger about inaction and drive those that may not be as aware of their own efficacy in creating change, into deviating from the norm and taking action.


Is Sexism for White People? Gender Stereotypes, Race, and the 2016 Presidential Election

Ana Bracic, Mackenzie Israel-Trummel, and Allyson Shortle

“Where women run they win.” This reliable refrain from the gender and politics literature failed to match the results of the 2016 presidential election, and many scholars scrambled to explain Donald Trump’s surprising victory. The Democratic nominee Hillary Clinton is a seasoned public servant and was predicted to win in most polls. Meanwhile, her Republican male opponent had no office-holding experience, and was the focus of a recent scandal involving a leaked tape where Trump bragged about how his celebrity status allowed him to “grab them [women] by the pussy.” His unconventional candidacy made the 2016 electoral returns nothing short of remarkable.

Much of the popular attention to gender in elections focuses on the so-called gender gap between men and women. Since 1980, women have voted for Democratic presidential candidates at higher rates than men. However, this focus on gender identity—men versus women—obscures what we consider a more important gender effect. Namely, how voters think about gender may be more relevant to vote choice than the gender with which voters identify. In the 2016 election, Trump regularly discussed gender in ways that encouraged voters to draw on their beliefs about gender when casting their ballots.

In our article, we argue that gender beliefs, or sexism, worked to shut out the most qualified candidate in 2016. Further, White voters, as members of a dominant group incentivized to support social hierarchies, were the most likely group to apply their sexist attitudes toward their voting behavior. In order to test our claims, we used an original exit poll survey data of over 1,300 voters conducted at 12 precincts in Oklahoma City on Election Day and a post-election national survey of over 10,000 White and Black Americans. Oklahoma City is one of the ten metropolitan areas that most closely match national demographics—age, educational attainment, race, and ethnicity—in the country.  We took advantage of that and selected our precincts based on the racial demography of the neighborhoods.  We conducted the poll in four types of precincts—predominantly Black, predominantly Latino, predominantly White, and racially mixed.  We fielded the election exit survey with the help of 54 undergraduate and 7 graduate students, all from the University of Oklahoma.

We find that while some members of all racial groups endorse sexist stereotypes about women’s lack of fitness for office, White voters are especially likely to apply such beliefs to their support for Trump. Moreover, we find that sexism is not only linked to vote choice, but to how much voters favor Trump. By contrast, there is not a significant association between sexism and favoring Clinton, which supports our contention that sexism was primarily activated in the campaign, not by Clinton’s gender identity, but by Trump’s rhetoric about women.

Interestingly, in the Oklahoma City analysis, we find that White women’s sexism was significantly linked to their vote, whereas there was no relationship for White men (see Figure 1). Among Whites who express the lowest levels of sexism there is no significant difference in vote choice for men and women, although men are more likely to support Trump than women. However, among voters who express the highest levels of sexism, White women are significantly more likely to vote for Trump than White men. Meanwhile, in the national survey, we find that both White voters and Black voters, across gender, link sexism to the vote, although the effect of sexism on vote choice is much larger for Whites than Blacks.

We conclude that beliefs about gender influenced the 2016 presidential election, and that White voters—including women—who believe that women are less competent than men in the political realm, employed sexism in their voting behavior. Men of all races may have held higher overall levels of sexism, but it was White men and women who responded most strongly to sexist campaign rhetoric, to the benefit of the candidate who chose to politicize sexism.






Left–Right Categorization and Perceptions of Party Ideologies

Federico Vegetti and Daniela Širinić

In Western democracies, left and right are among the most common terms to describe everything that is political. The heuristic power of these categories is so great that they have been able to remain in style since the French Revolution, adapting to the different issues and rhetorical styles that have characterized politics in many places and times. In fact, left and right are so adaptable, that not only do they act as good summaries for issues, values or worldviews, but also as “team flags” for political groups. This is no news for scholars of political behavior, especially for those focused on American politics – where ideology can be both symbolic and operational and it overlaps with partisanship (which in turn can be both expressive and instrumental). However, this “double nature” of ideological labels has two implications that have been overlooked so far. First, as the issues that are included in the left and right containers varies across contexts, the meaning of the labels (policy issue aggregators or political group identifiers) may vary too. Second, while issues can offer nuanced positions, group identities are by definition categorical. And when social categories enter the equation, parties and candidates cease to be treated equal, like shops on a Downsian street, and start being evaluated differently depending on whether they are in-group or out-group.

In our article published on Political Behavior we put together these two insights, looking at how the contextual variation in meaning of left and right labels relates to the way citizens perceive the ideological positions of political parties. First, we propose a method to assess the relative importance of political group memberships, vis-à-vis issue preferences and social structure, for citizens’ left-right self placements. Second, we test whether in places where group memberships are more prominent, people’s perceptions of parties are affected by a cognitive bias typical of categorical perception. This implies accentuating the perceived similarity between the self and ideological in-group parties, and the difference between the self and ideological out-group parties. We test our expectations on a sample of 24 European countries using data from the European Election Study of 2009 and the Chapel Hill Expert Survey of 2010. We find that, the greater is the relative importance of partisan group membership for left-right positions, (1) the better able citizens are to place parties in the correct left-right category, and (2) the further away from themselves they perceive ideological out-group parties, compared with experts’ assessments. We find no conclusive evidence for accentuated within-group similarity.

Our study suggests an important link between ideological and group polarization. Whereas left-right differences are usually interpreted as differences in policy content, we provide evidence that such differences may be accentuated in citizens’ perception due to a group-identitarian understanding of left and right labels. As a consequence, the observed degrees of left-right polarization in cross-country studies may as well reflect the amount of out-groupness between parties and citizens belonging to opposite ideological camps. We maintain that the mechanism governing this link is categorization, a basic cognitive device that people use in many situations arising in everyday life. While categories help individuals organize reality, they may help them organize the political reality as well. However, this may come at the expense of an accurate perception of political similarities and differences.


The Winner Takes it All: Revisiting the Effect of Direct Democracy on Political Support

Sofie Marien and Anna Kern

Declining levels of turnout and engagement in political parties as well as widespread political distrust have raised concerns about a democratic legitimacy deficit in advanced democracies. To address citizens’ novel participation preferences and to foster political support, proposals to give citizens more voice in the political decision-making process have been launched increasingly. Citizen involvement is expected to foster political support because citizens value voice and influence in political decision-making processes. However, despite these strong theoretical expectations, empirical studies on the effect of citizen involvement on political support remain scarce and the findings are inconsistent.

In our recent article in Political Behavior we aim to enhance our understanding of how and why citizen involvement affects political support. We gathered panel data in two comparable Belgian neighborhoods (with and without local referendum) in 2015. The data was collected using postal surveys in the month before the local referendum and in the three months following the referendum. This research design enables us to conduct a stronger causal test than previous cross-sectional studies with higher levels of ecological validity than experiments. We find an increase in political support in the aftermath of the local referendum. However, the driving force behind this increase is having voted for the outcome that has received the majority of the vote. As, in a direct democratic process, winners are by definition the majority, this explains the overall increase in support. Hence, the increase is not the result of satisfaction with being involved but merely with getting what one wanted as an outcome. Moreover, outcome favorability also shapes the process evaluations as winners think the referendum was fairer than losers.

Remarkably, however, and despite the contested nature of the issue, we also find that losers retained their political support. This is particularly noticeable, because a recent study in Sweden, for instance, showed that when trying to make a decision on a contested issue using a representative decision-making process, decision losers became less supporting of the political system and this negative effect proved to be remarkably stable over time (Esaiasson et al. in press). In contrast, in our study of a direct democratic decision-making process, losers’ support levels did not decline. In sum, involving citizens is not sufficient to increase their political support. Citizens deeply care about the outcomes of decision-making processes. Yet involving citizens does seem to be a powerful way to make contested decisions while keeping the support of decision winners and losers.

Phenotypic Preference In Mexican Migrants: Evidence From A Random Household Survey.

Rosario Aguilar, D. Alex Hughes, and Micah Gell-Redman

Our study explores phenotypic prejudice, which means prejudice related to aspects of appearance that denote a person’s race. We examine this topic within a group that has received little scholarly attention in this regard – Mexican migrants to the United States. Overall, our findings suggest that phenotypic prejudice does matter, and that the way prejudice works may change over time.

Ideally, voters would rely on candidates and parties’ platforms, proposals, and performance when making their electoral choice. But research on voting behavior shows that, depending on the context, non-political factors such as gender, class, and race also affect voters’ decision on the ballot. In the case of Latin America, and Mexico in particular, many scholars have hypothesized that more inclusive racial ideologies could make race less important for electoral choice. Others have begun to question this idea as evidence of racial appearance overlapping with other dimensions like socioeconomic status or political representation has emerged.

Our work contributes to this nascent research agenda on the political consequences of phenotypic prejudice in places where most people identify as members of the same racial group, like in Mexico. At the same time, we contribute to the research of the effect of new contexts on people’s attitudes as we explore the effect of tenure in the U.S. on migrants’ political expression of phenotypic prejudice.

Our results are based on an experiment that manipulated the appearance of a candidate supposedly competing for a governorship in the Mexican state of Nayarit. In total there were four conditions: the control that did not include a picture of the candidate and three conditions in which the candidate’s racial appearance was either white, indigenous, or mixed (mestizo).

The experiment was part of a random household survey of San Diego County, in which door-to-door canvassers contacted respondents who were born in Mexico and had resided in the United States for at least two months. This methodology allows us to generate a random sample, an advantage relative to many other experiments that use convenience samples. At the same time, we focus on a hard-to-reach population often neglected in academic research on vote choice.

Our findings show that migrants express a stronger preference for the indigenous-looking candidate, and they tend to think of the white and mestizo candidates as more ideologically conservative than the control. We interpret this as evidence that appearance may be signaling shared socio-economic interests which link voters to candidates. Our final result focused on migrants’ tenure in the U.S. Specifically, we find that migrants who have lived in the U.S. longer show greater support for the white candidate.

This last finding is particularly interesting, but also challenging methodologically. The ideal approach would be to conduct a panel survey, as cohorts of migrants might vary on relevant political, social, and economical factors. Data limitations force us to use a single cross section, and compare those with greater tenure in the U.S. to those who have been in the U.S. for a shorter time. The relevant theoretical question remains – how could tenure in the U.S. affect migrants’ phenotypic prejudice? Our cross-sectional evidence suggests that, as they spend longer in the United States, migrants may come to adopt a racial ideology that privileges whiteness.

This study contributes to the study of race and politics from a comparative perspective, as well as the study of the effects of migration on citizens’ political behavior. Our results underscore the need to incorporate non-political factors, such as race, into the study of electoral behavior even in societies whose racial ideology tends to be more inclusive.

Electoral discrimination of immigrant-origin candidates

Lea Portmann and Nenad Stojanović

Most of us believe we know what discrimination is and why it is wrong. After all, the term “discrimination” is omnipresent in today’s public discourse. And yet, perhaps surprisingly, defining discrimination is a very complex matter. While political philosophers, legal theorists, sociologists, political scientists and scholars from other disciplines have recently produced new insights, no consensus definition has yet been reached.

And even if we were to agree on what discrimination is, another difficult challenge would remain: how can we empirically measure it? In our article we focus on what we call “electoral discrimination.” It describes the phenomenon of significant electoral penalties incurred by candidates from minority groups running for political office, when majority voters prefer candidates who share their own identity traits. In the US context, numerous studies have explored the supposition that white voters tend to support white candidates over black and Hispanic ones. In the Swiss context, the Electoral Discrimination thesis holds that immigrant-origin candidates, bearing non-Swiss names, face potential discrimination in elections.

It is particularly inviting to explore Electoral Discrimination in Switzerland, because its electoral system (free-list proportional representation) allows voters to cross off candidates from their ballots. In other words, Swiss voters can allocate not only positive but also negative preference votes to individual candidates. For our study, we collected data from real ballots cast in six municipal elections in the Canton of Zurich. We used an open-access database to classify the names of all candidates into two categories: “Swiss” and “non-Swiss”. We then examined whether, all else being equal, candidates with “non- Swiss” names were at a disadvantage.

Our results provide evidence that immigrant-origin candidates did incur a significant electoral penalty in the 2014 municipal elections in Zurich. That is, they received more negative preference votes compared to similar candidates with typically Swiss names. The effect was stronger for candidates who were running on lists of right and center-right parties than for those on lists of left parties. Surprisingly with respect to a previous study, we did not find that candidates with “Western” non-Swiss names (e.g. those from England, Spain or Scandinavia) fared better than candidates with names from the former Yugoslavia or Turkey.

The novelty of our study is that our unique dataset allowed us to explore the phenomenon of Electoral Discrimination in a real-world environment. Compared with prior studies on the same topic – which have typically relied on aggregated electoral data, experiments or surveys – our analysis mitigates some important methodological concerns that have plagued this field of research. We hope that our results and our method will mark a step change in the study of Electoral Discrimination.