Intergenerational transmission of party affiliation within political families

Linuz Aggeborn, Uppsala University

Pär Nyman, Uppsala University

The accumulation of power within families has recently received increased attention from the academic community. A common objection against the Bushes, Cuomos, Le Pens, and Kennedys has been that they accumulate political power and consequently hinder candidates with new ideas. But is it true that children walk the same path as their parents?

In this paper, we investigate the intergenerational transmission of political affiliation using Swedish registry data. Our data cover everyone who has ran for local or national office between 1982 and 2014, but the analysis only uses those cases where both a child and its parent have been politicians. We have restricted the sample to candidates for the eight parties currently represented in the national parliament.

Note to figure: Illustration of a part of the main results, based on Table 1 in the published paper.

The probability that two randomly selected candidates belong to the same party is approximately 18 percent (it would be 12.5 percent if all parties were of equal size). It is therefore remarkable that around 76 percent of the children of a politician (n=14,275) choose to run for the same party as their parent. Given that Sweden is an egalitarian country with high income mobility, as well as a consensus-oriented democracy where the ideological distance between parties is relatively small, we would expect even larger effects in most other contexts.

To better understand this transmission we conduct a regression analysis where we control for the number of years that the child has lived together with the parent. This analysis shows that among children who never resided with the parent, only 31 percent run for the same party, highlighting socialization as the primary driver behind the strong intergenerational correlation.

Norms and Political Payoffs in Supreme Court Recusals

Udi Sommer, Tel Aviv University

Quan Li, Wuhan University

Jonathan Parent, Le Moyne College

Political developments in the United States over the past several years, and indeed in a number of democracies around the world, have focused attention on the importance of norms in the maintenance of democratic institutions and procedures. Increasingly it seems that norms have given way to political expediency and interests, particularly within elected branches of government. That said, some institutions would seem to be more norm-dependent than others. The United States Supreme Court, for instance, relies heavily on norm adherence for a number of its functions as well as, arguably, its very legitimacy. Concepts such as stare decisis and the practice of judicial review itself are based primarily on norm-compliance rather than formal rules or laws. Our paper examines the extent to which judicial recusals among members of the High Court are governed by norms rather than political considerations.

Unlike lower court judges, members of the Supreme Court are not bound by statute to recuse themselves from cases in which they may have a conflict of interest or other reason for abstaining from judgment. In this sense, the decision to recuse is norm-dependent. In order to test the degree to which norms, rather than political considerations, lead Justices to recuse themselves from particular cases, we compare the patterns of discretionary recusals to health-related recusals over several decades.

We chose these two types of recusals due to the unique features of each and what they are able to tell us about the decision-making processes of individual Justices. Health recusals are likely to be decisions over which the Justices have little control. We would therefore expect to see these randomly distributed throughout any particular Supreme Court term and for any individual Justice. Discretionary recusals, by contrast, represent instances where the Justice will be required to make a conscious choice about whether or not to participate. Our integration of the binary entropy function from information theory into a formal model of individual justice’s recusal decision illustrates that if political considerations indeed play a determinative role in Justices’ decisions regarding whether or not to take part in a case, we would expect the likelihood of discretionary recusals to decline when a minimum winning coalition (MWC) is expected as well as in cases which are highly salient. In other words, we expect a politically-motivated Justice to be less likely to recuse herself when the case is decided by a 5-4 majority than when the Court rules unanimously (or, for that matter, by an 8-1 or 7-2 vote).

Our empirical results demonstrate that neither a MWC nor the salience of the case had any influence on health-related recusals. However, both of these variables were significant when considering discretionary recusals. This finding not only confirms our theoretical expectations but indicates that, at least as far as recusals are concerned, Supreme Court Justices seem to be motivated by political considerations as well.

While a deep body of work exists in the law and politics literature that considers the various influences on Supreme Court decision making at the stages of certiorari (whether or not to accept a case for review) and merits (the ultimate vote on a case), relatively little attention has been paid either to the role of norms on judicial choices or the decision to recuse itself. This oversight is unfortunate as we believe much insight regarding judicial behavior can be gleaned from these types of considerations. While the notion that policy preferences are influential in determining the votes of Supreme Court Justices is not new in the scholarship on law and courts, understanding how political considerations, and political norms, influence the gamut of aspects of High Court activity is still an incomplete task.

Perhaps more importantly, though, our study demonstrates that even within an institution as heavily norm-dependent as the U.S. Supreme Court, politics still seems to play an outsized, if not dominant role in shaping behavior. The tension between norms and political interests seems to influence facets of the Court’s activity even at a very early stage. This should be of interest to anyone who values the role of an independent judiciary and the rule of law as a check on government authority and the separation of powers. And more generally, this is of interest to those concerned with the importance of norms in the maintenance of democratic institutions and procedures.

Policy Preferences and Policy Legitimacy After Referendums: Evidence from the Brexit Negotiations

Sara B. Hobolt, LSE

James Tilley, University of Oxford

Thomas J. Leeper, Core Data Science at Facebook and LSE

Referendums are increasingly common, yet the simple binary nature of most referendums also means that complex policy problems are often reduced to stark either/or choices. This is a particular problem in sovereignty referendums, which account for nearly half of referendums around the world, as such referendums are often open-ended, offering no clear blueprint for the complex future relationship between countries. Governments are thus given a difficult challenge: how to translate the outcome of a dichotomous vote choice into policy which involves difficult political trade-offs. This was particularly apparent in Britain after the 2016 EU membership referendum.

The narrow vote in favor of Brexit gave the British government a democratic mandate to exit, but it was a lot less obvious what that mandate meant in terms of the final political settlement. The post-Brexit landscape could have looked very close to the status quo of EU membership or could have involved a much more dramatic shift away from the previous legal, political and trading relationship with the EU.

While we know a lot about why people chose to vote Leave or Remain, we know far less about attitudes towards policy changes after the vote. Our article therefore set out to answer two important questions. What did people want after the referendum: what were their policy preferences? And what were people willing to accept after the referendum: which outcomes did they think were legitimate?

To answer these two questions, we ran a conjoint experiment on a representative sample of British citizens in late April 2017 (a month after Britain initiated the process to exit the EU, but before any actual negotiations began), which asked people to choose between different bundles of future policies. These covered eight different areas carefully selected to reflect the full breadth of the negotiations at that stage, and included rates of trade tariffs, immigration rules, budgetary contributions, the border between Ireland and Northern Ireland, rights of EU nationals in the UK and so on. Rather than asking people directly about each separate policy feature, we allowed people’s choices to reveal the acceptability of different policies. This meant that people had to engage directly with the difficult trade-offs involved in the negotiations. Crucially, we also measured both preferences towards the outcome of the Brexit process and views of which outcomes were legitimate. We then tested how people’s views were shaped by their own vote in the referendum.

Our results show that during the initial negotiation period there was agreement among both ‘Remainers’ and ‘Leavers’ in some areas of future policy. Most voters shared preferences for a settlement that ensured relatively frictionless trade between the UK and the EU and avoided any large one-off payments to settle Britain’s outstanding fiscal obligations to the EU.  Nonetheless support for many aspects of any negotiated settlement varied substantially according to how people voted in the referendum. Deep-seated divisions remained over the issues of immigration, legal sovereignty and EU citizens’ rights. In line with the research on vote choices in the referendum, we find that Leave voters were much more concerned with ensuring an outcome that guaranteed immigration control and greater legal sovereignty from the EU than those who voted Remain, who in turn cared more about guaranteeing EU citizens’ right to stay in the UK.

Crucially, we also find that neither side considered any of the available policy outcomes as adequately respecting the vote. When facing actual policy bundles rather than simply the Leave/ Remain choice, levels of perceived policy legitimacy were low among both Leavers and Remainers. It is not just that no policy bundle was seen as legitimate by both Leavers and Remainers, but that there was no policy bundle that could individually satisfy either group. Indeed, it was the winners of the referendum, Leave voters, who were particularly unlikely to perceive bundled policy outcomes as legitimate. This suggests that the reality of policy making following some referendums may not live up to people’s expectations, even for those on the winning side. It also means that while sovereignty referendum votes may often be seen as legitimate, the eventual policy outcome may not have the same legitimacy.

Taken together, this suggests that while referendums are generally popular with voters, there is a danger that both winners and losers may end up less than satisfied with the eventual outcome. This challenge is much greater in open-ended and high-stakes sovereignty referendums, such as those on EU membership and secession, since there is often considerable uncertainty about the policy consequences. This contrasts with narrowly focused single-issue referendums which are likely to offer much greater clarity on the policies stemming from the vote. It is thus vital that future research should do more to explore voters’ preferences towards the policy outcomes emanating from different types of direct democracy and how these shape attitudes towards democracy itself.

Who Cares if you Vote? Partisan Pressure and Social Norms of Voting

Edward Fieldhouse, University of Manchester

David Cutts, University of Birmingham

Jack Bailey, University of Manchester

Explaining why people vote is a fundamental puzzle in political science. We understand a great deal about when people are more or less likely to vote, and also about which kind of people are most likely to vote. Yet, when it comes to understanding why voters vote at all, narrow instrumental explanations are often unable to solve the so-called ‘paradox of voting’ – that is, why a utility maximising individual would vote if he or she has almost no chance of determining the outcome.

One of the most enduring answers to this puzzle is that people consider it their duty to vote, either for the good of democracy as a whole or for the good of a particular political party or candidate. While there is plenty of research about the importance of civic duty in encouraging people to vote, there little evidence about why such a duty exists. In particular, very little research asks either where this pressure comes from or whether people with whom voters share a political allegiance produce more pressure than others. In our new article in Political Behavior we use political discussion network data from the British Election Study to examine how this “partisan pressure” determines civic duty and voter turnout, and how normative expectations (measured by social approval of voting) vary depending on whether citizens have opposing or shared partisanship (Figure 1).

From a rational choice perspective, we might expect a norm of voting to be restricted to supporters of the same party, since encouraging extra votes for any other party or candidate would reduce their chance of winning. Conversely, a partisan voter might be more likely to pressure a friend or relative to vote if they share a party preference as an increase in turnout among fellow partisans provides expected benefits like improving the party’s chance of winning, but also expressive benefits too (e.g. “cheerleading”). The expectation that social pressure will be higher among pairs of fellow party supporters is consistent with group models of voter turnout which suggest that social pressure to vote comes about because of the promise of collective goals which can be achieved though high levels of group participation. A perception of group interests means that those belonging to the group are more likely to face social pressures from fellow citizens or group leaders to follow a group norm of voting.

Figure 1.  Relationship between Normative Expectations (NE), Civic Duty (D) and Vote (V)

Following these arguments, we consider three ways in which dyadic partisanship (i.e. whether the partisanship of two discussants matches or clashes) can affect turnout, as illustrated in the right-hand panel in figure 1. First, we expect likeminded citizens – who share support of a party – to be more likely to apply social pressure to vote.  We refer to this as partisan pressure. As argued above, we expect that where political interests are opposing, social pressure to vote should be weaker or absent. Second, we expect dyadic partisanship also to moderate the relationship between social pressure and voting: that is, people may be more inclined to take heed of social pressure from a discussant with whom they share a political identity. Third, discussants who share a political identity may more strongly shape the belief in the duty to vote than those who do not.

Our findings in respect to partisan pressure are quite clear. In short, people are more likely to perceive social pressure to vote from fellow partisans than from discussants who support an opposing party (see figure 2). The differences are quite substantial and are comparable in size to the differences we see by social relationship (for example, whether they are married, related, friends or colleagues, etc.). This partisan pressure also has a strong effect on the civic duty of discussants and, despite this, approval of voting also has a strong direct effect on turnout after controlling for the role of duty. 

Figure 2. Predicted Probabilities of Approval for Dyadic Relationship Types and Dyadic Partisanship Types during the 2014 European Elections with 95% Credible Intervals

Our findings concerning whether shared partisanship moderates the size of the effect of social pressure on duty and turnout are more mixed. The effect of partisan pressure on duty does not vary by dyadic partisanship and, although in our primary dataset we find the effect of social pressure on turnout is greater for shared partisans, we do not find this consistently across alternative data sources. This may reflect either the unique electoral context of our primary dataset, peculiarities of the supplementary samples, or both.

So why does partisan pressure matter? We argue that because normative beliefs of voting are critical to understanding turnout, it is important to understand more about where those beliefs come from and when they are most effective. Our research makes a clear link between party identity, social pressure, and voter turnout. It might, therefore, give us important clues to explain the high level of turnout seen in the 2020 U.S. Presidential election.

Partisanship as a Social Identity: Implications for Polarization

Emily A. West, University of Pittsburgh

Shanto Iyengar, Stanford University

We live in hyper-polarized times.  An extensive body of survey evidence documents that over time, but especially post-1990, partisans (defined as people who identify with one of the two major political parties) have increasingly adopted a hostile posture toward their opponents.  Partisans have also become averse to social interactions with their political opponents. The most vivid evidence of increased social distance across the party divide concerns dating and marriage.  In a longitudinal analysis, scholars found that spousal agreement on partisanship among recently married couples increased dramatically between 1965 and 2018.  Similarly, a recent Pew study demonstrates that discordant partisanship decreases people’s likelihood of being friends even when they are not seeking a romantic relationship.

What accounts for increased animus across the party divide?  Political scientists take for granted that the classic symptoms of affective polarization — in-group favoritism and out-group animus — are inevitable outcomes stemming from partisans’ sense of social identity.  It follows that a potential path toward reduced polarization is to weaken the connection between party affiliation and the sense of self.  In this paper, we provide evidence that the salience of partisan’ social identity can in fact be weakened – either by naturally occurring changes in the political environment, or by exposure to experimental manipulations that boost individuating as opposed to group-based elements of personal identity.  Yet no matter the mechanism involved, we find that the strength of partisanship as a social identity does not impact the level of affective polarization. 

Using a two-wave panel survey capturing natural variation in the salience of politics, we find that partisanship becomes a more important ingredient of partisans’ self-concept and that it contributes more to the sense of self in the midst of a national political campaign (mid-November 2018) than during a period characterized by attention to family relationships (late December).   We further show that partisans can be detached from their Democratic or Republican identity through a self-affirmation treatment that focuses their attention on personal traits and values.  Despite these significant effects of both the natural and experimental treatments on partisan identity salience, we find only limited evidence that when partisan social identity is not as salient, partisans become any less inclined to express in-party favoritism and out-party hostility. (We use standard indicators of partisan affect including feeling thermometers and trait ratings.)  Put differently, detaching partisans from their social identity does not reliably make them any less likely to elevate their in-group or denigrate their opponents. Thus, our study suggests that affective polarization is likely attributable to more than the classic in-group versus out-group distinction. 

In closing, we consider a variety of explanations for this surprising result.  First, both our treatments (changes in the electoral context and exposure to a self-affirmation manipulation) exerted significant but only modest effects on partisan social identity.  Future research might pursue stronger treatments to confirm that despite their weakened social identity, partisans remain sufficiently motivated to favor co-partisans and denigrate opposing partisans.  Second, it is possible that our inability to trace variation in affective polarization to changes in sense of partisan identity is due not to weak treatments, but rather to the possibility that in-party favoritism and out-party animus have been overlearned and are now operating implicitly.  To address this possibility, future research will need to pair treatments similar to those used here with implicit measures of partisan affect.  Finally, it is possible that partisan animus stems not only from the dynamics of group identity, but also from non-identity related processes, such as ideological disagreement or observational learning from elite cues. Testing these and other possible alternative pathways to out-group animus would bolster the evidence we provide here, suggesting that partisan animus is driven by something other than attachment to partisan social identity.

Shades of Privilege: The Relationship Between Skin Color and Political Attitudes Among White Americans

Nicole Yadon, Ohio State University

Mara C. Ostfeld, University of Michigan

Despite the growth in attention to the ways that skin color affects important social, economic, and political outcomes within racial identity groups, this body of work has almost entirely excluded those classified as non-Hispanic White.[1]In some ways, this is not surprising given Whites’ privileged position in the American racial hierarchy. Whites have historically tended to focus on the racial dynamics and nuances of other groups, with typically reflecting on their own racial identity and experiences. That is, their size and power have reinforced a notion of Whites as the status quo against which non-White racial identity groups are claimed to be distinctive. As stated by Sears and Savalei (2006, p. 901), “Whites’ whiteness is usually likely to be no more noteworthy to them than is breathing the air around them.”

But, as with every racial group, there is variation in skin color within the White racial identity group as well.  Discussions of Whites’ skin color is neither as prevalent nor explicit as it tends to be among many other groups – such as African Americans, Asian Americans, and Latinos (Wilder 2010; Hunter 2005; Rondilla and Spickard 2007; Godreau 2008; Telles 2014). However, there is a long history of Whites not only recognizing differences in the skin color among Whites, but assigning status and value to other Whites based on their skin color (Glenn 2008; Peiss 1998; Nowatzki 2006; Orsi 1992). This has included making assumptions about the status of other Whites based on their skin color, as well as imposing skin color identities on individuals based on their perceived status. This skin color-status association has been sufficiently clear that Whites have historically sought to emphasize their Whiteness and distance themselves from non-Whites when their social status felt threatened (Nowatzki 2006; Orsi 1992). 

More importantly, variation in skin color is associated with significant life experiences and opportunities for White Americans today. Whites with darker skin do, in fact, face more adversity across a range of contexts—including education and arrest rates—than Whites with lighter skin (Branigan et al. 2013, 2017; King and Johnson 2016). Some White employees have even filed claims of color-based discrimination against their White employers (Hill v. Textron Automotive Interiors, Inc., 160 F. Supp. 2d 179, D.N.H. 2001). Thus, to the extent that access to opportunities affects political views, we should expect skin color to be an important predictor of political attitudes among White Americans. 

Further, there is reason to believe that the significance of skin color among Whites may be growing in importance and relevance. There is both a growing population of individuals (e.g. Latinos, Arab Americans, Asian Americans) who don’t fit neatly into either of the oldest American racial categories—Black or White—as well as a growing population of individuals who identify with more than one racial identity group (Hochschild et al 2012; Davenport 2016; Masuoka 2011; Williams 2006). Each of these demographic changes has contributed to a blurring of the White-Black categorical boundaries that have historically been at the root of American racial divisions. In turn, the boundaries of categorical Whiteness—and the privilege it entails—are becoming less clear. Yet it is unlikely that the heightened salience and import of one’s White racial identity would fall equally across Whites. Importantly, the ones who have the most to lose from the blurring of boundaries between racial categories, and a greater emphasis on skin color, are darker-skinnedWhites. Unlike lighter-skinned Whites, darker-skinned White individuals will no longer be at the top of the American racial hierarchy should skin color gain in prominence as a form of social stratification.

Our theory and results highlight that skin color is an important factor in both White socioeconomic outcomes and political attitudes, and one that operates in distinctive ways relative to members of other racial identity groups. Specifically, we argue that Whites with darker skin will be more motivated to protect racial group boundaries and interests, assert their White identity, and consequently have more conservative attitudes on political issues that heighten their advantage over non-Whites, relative to those with lighter skin. Our findings offer evidence consistent with this theory. As is the case among other racial identity groups, Whites with comparatively dark skin, on average, have lower levels of education as well as more negative criminal justice outcomes than lighter-skinned Whites. Yet unlike other racial identity groups, Whites with darker skin tend to be more conservative on racialized political issues that privilege them over non-Whites, than those with lighter skin. Importantly, however, this conservatism does not appear to be driven by the anti-Black bias that has historically shaped White racial attitudes. Instead, this conservatism appears to be particularly oriented toward the groups that have at times been able to permeate and blur the boundaries of the White racial identity through a type of “honorary White” status (Bonilla-Silva 2004; Jardina 2019; Haney-Lopez 1997). Taken together, our work sheds new light on the meaning and inner workings of Whiteness in contemporary American politics.


[1] For purposes of brevity, we use “White” to refer to individuals who identify as White and do not identify with any other racial group, or as Hispanic, Latino, Middle Eastern or North African.

Eye of the Beholder: Partisanship, Identity, and Politics of Sexual Harassment

Stephen C. Craig, University of Florida

Paulina S. Cossette, Acadia Editing Services

Two current members of the U.S. Supreme Court took their seats despite allegations of sexual harassment (Clarence Thomas) and sexual assault (Brett Kavanaugh) leveled against them during their confirmation hearings. In each instance, the Senate vote was close and split mainly along party lines: Republicans for and Democrats against. Polls showed that a similar division existed among party supporters in the electorate. There are, however, differences among rank-and-file partisans that help shape their views on the issues raised by these two controversial appointments to the nation’s highest court.

Using data from a 2018 online survey of registered voters nationwide, our study examines the factors associated with citizens’ attitudes about the role of women in politics, sexual harassment, and the #MeToo movement. We are particularly interested in determining whether a strong sense of partisan identity adds significantly to our understanding of people’s attitudes on these matters. In addition, our experimental design allows us to determine whether shared partisanship overrides other factors when an elected official from one’s own party is accused of sexual misconduct. The survey was constructed as follows:

After completing an extensive background questionnaire, respondents read short biographies of two (fictional) candidates for a U.S. House seat and were asked to indicate a vote preference based solely on that information. The party affiliation and gender of each candidate varied randomly, though all races pitted Republican vs. Democrat and male vs. female. Not surprisingly, most partisans indicated support for their party’s nominee, while Independents split more or less down the middle. Respondents then read a news story describing harassment allegations (solicitation for sex, inappropriate touching, talking about his/her sex life, remarks about the accuser’s personal appearance, and the use of vulgar or abusive language) that had been lodged against the incumbent by two former staffers, a campaign worker, and a lobbyist, all of the opposite gender. After answering the vote preference question a second time, respondents read another story in which the incumbent accounted for his/her behavior in one of three ways (denial, apology, counterframe) and indicated not only (a) which candidate they would support but also (b) whether the alleged wrongdoer should resign from office.

Charges of sexual harassment had little effect among opposing partisans because very few of those individuals planned to vote for our hypothetical incumbent (at T1) in the first place. Support for an accused candidate fell precipitously, however, among both co-partisans and, to a lesser extent, Independents. While the dropoff among co-partisans was evident for all incumbents regardless of party, its magnitude was greater for Democrats (roughly a 55-point decline for men and women alike) than for Republicans (about 40 points for candidates of either sex). In other words, rank-and-file Democrats were more likely than their Republican counterparts to hold co-partisans accountable for their alleged transgressions.

Consistent with prior research, we hypothesized that the loss of support would be less among an accused candidate’s co-partisans who possessed a strong sense of partisan identity. As shown in the figure below, this was indeed the case. For those with the highest score on our identity index, the probability of expressing a preference for the incumbent after learning of the harassment charges was about 52% for Democrats and 76% for Republicans; these totals are significantly higher (p < .001) than the 16% and 24% registered by co-partisans having the lowest identity score. Further, this figure indicates that each additional increment on the identity index was associated (though not always significantly) with greater support at T2 for the incumbent.

Note: Results represent the probability of a post-allegation vote for the incumbent estimated at the values shown for shared partisanship and strength of respondent’s partisan identity as obtained using the margins command in Stata 14.1. Solid lines represent estimates for respondents who expressed a vote preference for the incumbent at the baseline; dashed lines represent those who preferred the challenger at the baseline. All other variables in the logit model were held at their actual values in the dataset.

How did any of this change when the incumbent addressed the charges against him/her? While the largest post-response gains in vote support (from T2 to T3) occurred, as one would expect, among co-partisans on both sides of the aisle, these gains were fairly modest in most cases (ranging from zero to 19.1 points). Once again, however, identity played an important role; that is, a stronger sense of identity was positively and significantly associated with greater post-response support for both Democratic and Republican incumbents – at least among respondents who had planned to vote for that candidate from the start. This pattern is evident in the figure presented below.

Note: Results represent the probability of a post-response vote for the incumbent estimated at the values shown for shared partisanship and strength of respondent’s partisan identity as obtained using the margins command in Stata 14.1. Solid lines represent estimates for respondents who expressed a vote preference for the incumbent at the baseline; dashed lines represent those who preferred the challenger at the baseline. All other variables in the logit model were held at their actual values in the dataset.

One result not shown here is that, unlike what we found at T2 (post-allegation, pre-response), respondents overall were more receptive to the “image-repair” strategies of women relative to those employed by men; that is, after hearing an incumbent’s response to allegations of sexual harassment, they were more willing to give women the benefit of the doubt.

When Trust Matters: The Case of Gun Control

John Barry Ryan, Stony Brook University

Talbot M. Andrews, Stony Brook University

Tracy Goodwin, Girl Scout Research Institute

Yanna Krupnikov, Stony Brook University

In our forthcoming paper in Political Behavior, we look at how a lack of trust in government undermines support for gun control via fears of a slippery slope. Trust in government – a concept distinct from people who want a smaller government – serves to bolster public policy support among those that pay the ideological costs of a policy. Pivotal to the effect of trust is the broader political context, which can vary over time. In a context of deep partisan divisions, for individuals who do not trust the government, even small ideological costs can signal the beginning of a process that leads to much larger ideological costs down the line – a process akin to a “slippery slope.” 

In political science, slippery slope arguments best describe cases where there is a continuum of policy options. The concern of the slippery slope is that the policy continuum will act like a row of dominos, and you must oppose any change from the status quo because it will start the process of the dominoes falling towards even less desirable policy outcomes. Once the first policy is adopted, then norms change which makes it easier to pass more extreme policies until an undesirable end state is reached.

As policies change, when exactly an individual begins to fear the slippery slope depends on the individual’s preferred policy. For example, in the case of gun control, a strong conservative who does not trust the government would oppose even small changes in the liberal direction, toward more restrictive gun policies, even if in principle they approve of slightly more gun control. This is because a small change in the liberal direction might be the first step to a more moderate position, which is unacceptably far from the conservative’s preferred policy. A moderate who doesn’t trust the government would support more restrictive gun policies up to a point. Eventually, however, the slippery slope fear kicks in as they come to believe adopting more restrictive policies will lead to an outright ban.

Our data analysis comes in two parts. First, we use the American National Election Study from 2000 to 2016 to show that trust in government has an effect on gun control support in recent years as elite opinion has sorted and polarized. As shown in the figure below, in 2000 trust had no effect on gun control policy support regardless of the respondent’s ideology. By 2016, however, conservatives who trusted the government were more likely to support gun control than conservatives who did not trust the government. This demonstrates the first point that the effect of trust in government on policy attitudes is conditional on the political context.

However, since the ANES only asks one general gun control question, it cannot provide any evidence in support of the proposed slippery slope mechanism. For this reason, we conduct another analysis using The American Panel Survey (TAPS). This dataset includes questions about five different gun control policies, allowing us to construct a continuum of support for gun control. We use a scaling technique developed by Mokken (1971) to create a scale in which each policy is more restrictive than the one before – from requiring background checks for all gun buyers to banning the possession of handguns.

The results illustrated in the figure below show that, for conservatives, trust in government increases support for the least restrictive gun control policies. This is where we would expect to see the fear of the slippery slope mattering. Conservatives oppose the more restrictive policies regardless of how much they trust the government – worrying about a slippery slope is no longer relevant because we are now at unacceptable point of restrictive policies. As a result, trust in government does not affect opinion on those policies.

For less conservative respondents, trust only has an effect on the most restrictive gun control policies. Trust does not matter for the least restrictive policies because support is so high that there is a ceiling effect. It is only as we get further down the restrictiveness scale that support drops among those who do not trust the government. These are the cases where the respondents who do not trust the government fear where these more restrictive policies might lead.

While we illustrate the influence of the slippery slope fear on gun control attitudes, slippery slope arguments are made in a variety of policy domains including gay marriage, abortion, euthanasia / death with dignity, and gene editing. Slippery slope arguments are often dismissed as rhetorical devices with dubious logic, but our paper suggests that voters may internalize the possibility that policies may increase in restrictiveness. As trust in government declines and elite polarization spreads to more issues, then the fear of the slippery slope may become an even more important predictor of public policy attitudes.

How Moral Motives Link Party Stereotypes

Scott Clifford, University of Houston

The Democratic party has long been perceived as “owning” issues such as welfare, Social Security, poverty, and protecting the environment. That is, much of the public believes that the Democratic party is more committed to these issues and more competent at solving these problems. Similarly, the Republican party has long been perceived as owning issues such as terrorism, national security and crime. Ownership of these issues affords each party a number of advantages in elections, particularly when their owned issues are salient.

More recent work suggests that the parties also “own” certain character traits. Democrats tend to be seen as more compassionate than Republicans, while Republicans tend to be seen as tougher leaders. While the theories of issue ownership and trait ownership have been influential in understanding voter decision-making and electoral outcomes, we know rather little about how these issue and trait stereotypes relate to one another.

In my paper, I argue that each party’s owned issues and owned traits are linked together in the minds of voters. This is because each party’s owned issues represent a particular type of moral problem. Thus, to solve these problems, voters look for politicians who are dispositionally motivated to do so. For Democrats, their owned issues, such as poverty and healthcare, share common moral themes of harm, suffering, and fairness. In this case, politicians who are compassionate and fair-minded should be seen as the most motivated and most capable at solving these problems. For Republicans, their owned issues, such as national security and crime, share common moral themes of defending the interests of Americans and protecting law and order. Thus, to solve these types of problems, voters look for politicians who are patriotic and tough. 

I test these expectations across three studies, including a nationally representative survey, two within-subjects designs, and using open-ended trait-listing tasks. The results show that when people are asked to think about Democrat-owned issues, they rate compassion as being much more important than when they are asked to think about Republican-owned issues. In contrast, when asked to think about Republican-owned issues, people rate traits like patriotism and leadership as much more important than when they are asked to think about Democrat-owned issues. These patterns emerge even among the least politically knowledgeable, suggesting that the links between these issues and traits may be partially intrinsic, rather than elite-driven. 

Additionally, my results show that people use these issue-trait connections when evaluating the parties’ issue competencies. When thinking about a particular issue, people who prioritize a party’s owned traits are more likely to rate that party as more competent at handling it. For example, those who rate compassion as particularly important for handling healthcare are more likely to see the Democratic party as more competent on the issue. These findings are clear even using within-subjects analyses, ruling out the effects of partisan identity, political ideology, or other differences between people. 

These findings help to explain why the parties’ issue and trait reputations are so durable and why it is so difficult for a party to gain an advantage on their opponent’s owned issues and traits. This is because each party’s owned issues are not an arbitrary collection of topics. Rather, they share a similar underlying moral problem. Thus, even if an issue, like terrorism, goes dormant for years, the Republican party will likely retain ownership over the issue so long as they maintain an image of toughness and patriotism. Similarly, Democrats can maintain ownership over healthcare so long as they maintain an image of compassion. These trait perceptions can likely be maintained through their efforts on related issues. For example, an emphasis on crime and immigration can help Republicans maintain their image of toughness and patriotism. And an emphasis on healthcare and social welfare will help Democrats maintain their image of compassion. As a result, attempts to trespass on the opposing party’s owned issues or traits, such as Republicans’ attempts at claiming “compassionate conservatism,” will be a hard sell to voters.

Corrupted Estimates? Response Bias in Citizen Surveys on Corruption

Mattias Agerberg, University of Gothenburg

There is now broad agreement among researchers about the destructive consequences of corruption. As a consequence, measuring corruption has become a global industry, with leading actors like Transparency International spending millions of dollars on the construction of corruption indicators and the surveying of ordinary citizens’ attitudes about, and experiences of, corruption. The latter type of measure is sometimes referred to as “citizen surveys” and is commonly used in the social sciences. These measures provide individual-level data from people around the world that can be used to study important individual-level research questions on topics like clientelism, political legitimacy, and voting. However, while data availability has increased rapidly, little is known about how people form their perceptions of corruption and to what extent their reports of encounters with corruption are accurate.

A large body of research has documented reporting bias among survey respondents, where an individual’s reported perception of some phenomenon might be influenced by factors other than the actual occurrence of the phenomenon. This paper explores two potential sources of bias in individual reports on perceptions and experiences of corruption. First, I draw upon research on economic perceptions and economic voting, and related research on political bias and motivated reasoning, and argue that respondents are likely to respond in a political manner when asked how they perceive corruption in their country. I refer to this as “political bias”. As a result, we should, for instance, expect incumbent supporters to in general report a substantially more positive view of corruption in their country, compared to other groups. Second, I argue that direct questions about corruption experiences are likely to be sensitive and hence subject to “sensitivity bias”. Admitting to having been part of a corrupt transaction is arguably an act of revealing sensitive information and hence something that is likely to be under-reported. This, in turn, makes estimates of the level of corruption in society based on experiential surveys likely to be biased.

I test these conjectures with two preregistered experiments embedded in a large survey fielded in Romania to over 3000 respondents. The first experiment is designed to randomly make the political affiliations of one group of respondents more salient before answering questions about corruption perceptions. The second experiment, deploying a so called “list experiment”, is designed to minimize the likelihood of sensitivity bias among another group of respondents. 

The results show strong evidence of different types of response bias with regard to questions about corruption. First, government supporters systematically report a much more positive view of the corruption situation in Romania. A simple question order prime – asking about political affiliation before corruption perceptions – makes this effect almost twice as large, suggesting that a substantial part of the gap is the result of respondents defending and justifying their political identity. Second, the results suggest that direct questions about corruption experiences are sensitive and under-reported. For some groups, like women, the under-reporting is massive, according to the estimates. For this group the true rate of corruption victimization might be three times as high as the reported rate.

Researchers should hence be cautious in estimating models with individual-level measures of corruption perceptions and individual-level political outcomes such as incumbent support or vote intention. Relationships like these are likely to be affected by strong feedback mechanisms and reversed causality, especially in surveys asking political questions before corruption questions. This also implies that corruption perceptions among the public should be expected to be more polarized along political lines at times when political affiliations are more salient, for instance during an election year. From a broader perspective, the results show that political bias can be substantial even outside of traditionally studied topics like perceptions about unemployment and inflation.

The findings from the second experiment on sensitivity bias strongly suggest that direct questions about corruption experiences need to be treated as sensitive questions. According to the results, the direct question both fails to accurately capture the overall occurrence (which is heavily under-reported), and to capture the dynamics of bribery and which groups are most likely to be targeted. This is something that needs to be taken into account by anyone who uses similar measures.

At the same time, the findings in this paper should not be taken as a discouragement of research on corruption or of efforts to quantify the incidence of corruption. Rather, given the immense importance of the topic, it is crucial that we scrutinize the methods we use and try to be cognizant of potential sources of error and bias. I am also not saying that existing measures based on citizen surveys should not be used, but that researchers should approach such measures with caution and think about how reporting bias might influence a given estimate. Overall, this study suggests that paying more attention to issues of response bias is an important part of further advancing the field of corruption research.