This post was developed by Ken Kollman and Tevah Platt, based on the talk, “When People Change Their Partisanship, is it Bottom-Up or Top-Down?” that Ken Kollman presented for the Research Center for Group Dynamics Winter Seminar Series on Political Polarization (2023) at the University of Michigan Institute for Social Research. Ken Kollman is the Director of the Center for Political Studies.
Partisanship is sticky. People tend to vote like their parents and to maintain their partisan leanings over time. But to understand partisanship, we need a model that can explain why people change party loyalties when they do. This is what Ken Kollman and John E. Jackson of the University of Michigan Center for Political Studies (CPS) provide in Dynamic Partisanship: How and Why Voter Loyalties Change. The following summarizes their overarching argument.
What is partisanship?
Partisanship is a group-based, shared identity. A classic work from 1960, The American Voter, also out of ISR, describes partisan identity as a long-term, affective, psychological attachment to a political party. According to this famous “Michigan model,” the socially-informed attitudes and values we form early in life durably influence the way we identify with political parties and how we vote.
Kollman and Jackson argue that partisanship has similarities to brand loyalty. It’s relatively stable and habitual, but it’s also evaluative and cognitive. Parties compete for votes and, importantly, for voter loyalty among “consumers” who are considering and comparing candidates and party ideas. Voters “experience” parties in office and in campaigns, and evaluate parties like consumers with products. Yet voting over time for the same party can also become habitual until voters become dissatisfied with what they chose.
What drives partisanship change?
Ronald Reagan often said that he didn’t leave the Democratic party, but the Democratic party left him. The quip encapsulates what Kollman and Jackson find to be the primary answer to the question of what moves partisanship. Two other processes do influence partisan dynamics– changes in people’s political attitudes and their evaluations of the performance of politicians in office – but it’s the behaviors of parties that they find are the greatest contributors to changing partisanship.
At the micro-level, partisanship is driven by evaluations of parties and politicians who are themselves changing for strategic reasons to try to win office.
At the macro-level, party polarization is a consequence of elite-level competition for voters, mostly at a national scale– for example, in response to national policies and movements.
In the broader debates about polarization, the stake they claim is that polarization is driven by elite-level competition for power, and not by ordinary people changing their minds about their ideologies or issue positions. It’s top-down, driven by what politicians and their parties do.
How parties compete
A canonical model of party competition came out of the mid-century work of Anthony Downs, who developed a theory of party competition in ideological space. This theory drew a picture of the Democratic and Republican parties converging on the “median voter” the way that ice cream trucks would converge at the middle of a beach to attract the most customers. More complex models admit that political ideology and conflict takes place in multiple dimensions; on the ground, for example, a candidate or party that is moving right on social issues could be moving left on economic policy, perhaps testing out impacts on voters.
A case in point: the language of industrial protectionism (saving factories) was an economically leftward move of the Trump-guided GOP that effectively turned Ohio from purple to red by attracting whites in Northeastern Ohio to the Republicans. Dynamic Partisanship tracks such patterns across the US, the UK, Canada and Australia over more than a half-century, but the overarching trend is that parties are the moving gear in dynamic partisanship. Voters don’t need to be moved, but partisanship can change because voters are reacting to parties that move– and that’s the underlying dynamic.
Partisan trends in the US
This figure, from Dynamic Partisanship, plots partisanship among three groups of the U.S. electorate– northern whites, southern whites, and African Americans– from 1956 to 2016, with Democratic partisanship increasing on the y axis. There are three distinct patterns:
Northern white partisanship is the most stable, coming closest to the traditional view of party identification as an unchanging personal attribute;
The 1964 election, on the heels of the passage of the Civil Rights Act, is a critical turning point in African American partisanship, making a full-point leap and remaining consistently high on the Democratic scale from that time;
Southern white partisanship shows a strong, gradual trend shifting from moderately Democratic to weakly Republican over 61 years.
The twin phenomena of southern Black voters becoming more Democratic since the 1960s and southern whites becoming slowly more Republican over time represent two of the major tectonic shifts in American society and politics that have occurred in the last half century.
The innovation is that the model used in Dynamic Partisanship can accommodate these divergent patterns– relative stasis, abrupt changes, and gradual changes. For the details and the myriad examples, check out the book.
A major challenge confronting our society is the intensity of partisan disagreements, even about what constitutes fact. To address this challenge, fact-checking was initiated by organizations such as FactCheck.org and PolitiFact in the 2000s to assess the accuracy of political claims. For fact-checking journalism to be helpful, there should be widespread public trust in such websites. Yet about half of Americans believe fact-checkers are biased.
Two decades after its emergence, why has fact-checking failed to get traction among the public? What changes can help fact-checking sites earn greater trust?
Research shows that FactCheck.org, PolitiFact, and Washington Post Fact Checker in recent years often corrected Republicans at a greater rate than Democrats. These asymmetries indicate a specific way in which fact-checkers pursue objectivity. Unlike traditional journalism that in principle seeks objectivity by covering all sides equally, fact-checking’s coverage decisions are guided by what is consistent with the best available evidence. Thus, fact-checking sites at times disproportionally critique one party more often than the other as guided by evidence. However, this approach poses a dilemma. Their asymmetric coverage may reflect genuine imbalances in reality, but it can also undermine trust in fact-checking sites among partisans.
To examine how asymmetric coverage affects partisans’ source credibility perceptions, I conducted a survey experiment in 2020. Participants were asked to assess a news source based on a set of headlines where the majority of headlines challenged 1) Republicans, 2) Democrats, or 3) both parties equally (baseline condition). The condition where most headlines challenged Republicans was considered as uncongenial to Republicans and congenial to Democrats (vice versa for the condition in which most headlines challenged Democrats).
The results indicate that partisans perceive a source to be less credible when its coverage corrects their own party more often than the other (uncongenial asymmetry, i.e., most headlines challenge in-group), compared to when the source corrects each party at a similar rate. Contrary to popular belief that Republicans tend to be more resistant to uncongenial news and facts, Democrats reacted more negatively to uncongenial asymmetric coverage than Republicans. Even when a source challenged the opposite party more often (congenial asymmetry, i.e., most headlines challenge out-group), partisans found the source to be less credible than symmetric coverage. In fact, the results show that partisans found congenial asymmetry to be particularly less credible when a portion of coverage challenged their own party on a highly polarized topic such as immigration, which is the type of issue that fact-checking sites often cover.
This study highlights a difficult dilemma that fact-checkers face in their coverage decisions. While asymmetric coverage of political parties is at times needed in response to evidence, it can alienate partisans from both sides. We might believe that balance for its own sake should be avoided, but more symmetric coverage of political parties and further efforts to signal nonpartisanship can help build broader public trust in fact-checking sites.
Hwayong Shin is a Ph.D. candidate in Political Science at the University of Michigan. Her research investigates ways to mitigate the partisan divide and build common ground by increasing the credibility of evidence-based information and by invoking shared emotions and experiences. Shin is a Next Generation scholar at ISR’s Center for Political Studies and was awarded the 2022 Garth Taylor Fellowship in Public Opinion.
Post developed by Francy Luna Diaz and Tevah Platt, based on the work Luna Diaz presented at the 2022 Annual Meeting of the American Political Science Association (APSA), “Latinos for Trump in 2020: A Story of Heterogeneous Information Environments and Social Media.”
Latinos continue to surprise with their sustained and increased support for former President Trump. After numerous episodes during his campaign for president and time in office espousing what was seen as anti-Latino and anti-immigrant rhetoric and policies, it is puzzling to many that almost one in three Latinos voted for him during the 2020 presidential election.
Los Angeles, United States – October 22, 2016: LOS ANGELES, USA – OCTOBER 22: A pro Donald Trump rally outside a CNN building on Sunset Blvd in Hollywood, Los Angeles
What is his appeal to many Latinos? Scholars have found that many Latinos agree with his policies, and more Latinos say they are Republicans than before. Political science PhD student Francy Luna Diaz, however, wants to probe deeper. She analyzes why some Latinos are attracted to Trump and others are not. She proposes that Latinos’ information environments are crucial to understanding the wide-ranging variance of the groups’ political attitudes.
Latinos are embedded in information environments that differ from other Americans because of distinct social media use and social networks. Information environments refer to the various sources of information that people have around them, mainly social media platforms. Latinos, in fact, are more likely than other ethnic groups to rely on social media and messaging applications to share and obtain information with close friends and relatives.
Additionally, Latinos often maintain ties to Latin America and are exposed to political information about and emerging from Latin America.
These factors combined—higher use of social media and more diverse information networks—increased Latinos’ vulnerability to disinformation and misinformation in 2020 and may have influenced some Latinos to distance themselves from the Democratic party.
An example of social media misinformation: Cuban-American influencer Alex Otoala falsely claimed Democrats were going to send a caravan of Cuban immigrants to the US border to disrupt the election.
Luna Diaz analyzed ANES data from 2020 and 2016 to explore whether traditionally recognized factors such as party identification, age, education, income, trust, generation in the U.S., language, and place of birth, among others, correlated with respondents’ decision to vote for Trump. She found that in 2020, Latinos who used Facebook more frequently were significantly more likely to vote for Trump, while the same pattern was not present for non-Latinos.
Luna Diaz also looked at answers to open-ended questions in ANES, summarized in the Table below, and found that while Latinos offered similar considerations when discussing why they like each candidate or party, differences emerged when they discussed why they disliked the Democratic candidate and the Democratic party in 2020. Interestingly, the reasons offered may point to (sometimes false or misleading) news spreading online claiming that the Democratic party is leading the U.S. towards socialism and President Biden behaves inappropriately with children.
Understanding Latinos’ political behavior is crucial to evaluating the present and future of American electoral politics. Latinos’ share of the population is steadily increasing along with their political influence in close elections. On a larger scale, it is important to uncover whether the spread of potential disinformation via social media impacts the political participation of different groups. Understanding the effect of online disinformation and misinformation will only increase in importance as democracy remains under threat in the United States.
The theme of the roundtable presented on April 8, 2021, was “Can Democracy Survive?” Ken Kollman introduced the event, noting that the survival of democracy was a question that Miller and Converse worried about. Their ambition was to study survey respondents and political parties and candidates much like other scientists studied cells and atoms and planets, but they cared about the fate of democracy. Their legacy of scientific inquiry into politics and society continues at ISR and CPS. A recording of the event is available below.
Ken Kollman: Moderation and Extremism in American Political Parties
Ken Kollman examines partisanship in a forthcoming book written with John E. Jackson, Dynamic Partisanship: How and Why Voter Loyalties Change. The book, from University of Chicago Press, presents a framework that relates the changes that political parties undergo, and the partisanship of the electorate in four countries: the United States, the United Kingdom, Canada, and Australia.
Political parties are changing and adopting new issue positions, says Kollman, and the mass public pays attention to these movements. Both partisanship and voting decisions respond to the people’s evaluations of these partisan positions relative to their own interests.
Kollman makes the case that both major parties in the United States are perceived to have moved away from the center since 2008. These patterns include a continuation of the shift of the working class towards the right and the Republicans the shift of more educated voters to the left and the Democrats. These shifts have consequences for politics and for the survival of democratic processes.
Most people in the US hold their partisanship for life, but notable portions of the electorate change over the course of their lifetime. The most common reason they change is that they perceive the major parties as moving away from them or toward them on issues of fundamental importance, including economics and racial liberalism. They change much less often because of the performance of a party in office or because they change their ideology or issue preferences.
In The American Voter, Miller and Converse wrote about partisanship as a result of socialization; they argued that partisanship shapes the perceptions of events, of candidates, and the vote. Kollman and Jackson don’t necessarily argue with this. The American Voter portrayal of partisanship remains robust and is good at predicting the vote. In contrast, Kollman and Jackson focus on the dynamics of partisanship and how partisanship changes. Group memberships based on interests and elements of socialization determines partisanship. It’s malleable and they model it as a form of what’s called Bayesian updating, a method of modeling how people incorporate new information in their decisions.
Kollman and Jackson are continuing to analyze the patterns from the past to predict what’s going to happen in the future if the two parties take different positions. The chart below shows how different groups of voters would respond if the Democratic Party moved to the left. What they find, first, is that partisanship becomes more Democratic for every group as the Republicans become more extreme. African Americans are complex in that they prefer the state of the Democratic Party in 2016, but their partisanship actually drops away if the Democratic Party moves to the left or moves to the right. Among white voters, the Democratic Party would lose partisans (and votes) if it moved to the left.
The trends of both parties away from the political center are worrisome for many people. Extreme party positions, including the pursuit of extralegal strategies to either pass policies or hold and maintain power, could become more likely as parties become more extreme.
Rob Franzese: What Causes People to Become Political Extremists?
What explains the rise of far-right nationalist-xenophobic and rightwing populism in the United States and other developed demoocracies? Robert Franzese presented research to address this question.
Scholars have noted that the rise of anti-immigrant, anti-globalization, anti-elite, anti-government sentiments correspond to a sea-shift of white working class voters to the right. One explanation for this shift is the notion that people have been left behind socioeconomically, and experience angst as a result. While support for parties farther to the right increased everywhere, it is especially notable in regions experiencing economic hard times, demonstrating support for these economic explanations of voting behavior.
Surveys have examined whether the shift to the right was attributable to socioeconomic malaise and decline, or whether it was due to cultural status threat. The data from these surveys seem to suggest that the political shift resulted from preceved xenophobic threats and it doesn’t have anything to do with the economic conditions.
However, Franzese contends that this conclusion is both wrong and wrong-headed. Instead of either/or explanations for political shifts, he suggests that we think in terms of both/and. Both neighborhood socioeconomic malaise and xenophobic anxiety associated with cultural change are both part of a broader sense of socioeconomic and cultural threat, as described in the graphic below. Franzese emphasized the importance of heterogeneity of perceptions. Some people are more susceptible to demagogic railing against the elite, the media, and foreigners. Other people will be immune, and may even become more repulsed by populist appeals.
This approach shows that the socioeconomic conditions the individual experiences are partly contributing to social-cultural threat perceptions that produce support for extremism. It’s not just economic hardship that creates the response, but economic hardship contributes to the sense in some respondents that their group is being left behind.
Extremism, especially far-right extremism, is a rising threat to democratic society. Therefore, understanding better the provenance of this rising far-right extremism and concomitant rise in rightwing populism is urgently essential. Casting the possible causal processes as some xenophobic or socioeconomic threat perception is unhelpful. These processes are better understood as complementary.
Pauline Jones: Democratic Survival, Using Lessons from the Muslim World
Pauline Jones notes that many people think that democracy is either unlikely or impossible and due to familiar tropes that Islam and democracy are somehow incompatible.However she contends that democracy and Islam are not incompatible at all. Muslim democracies exist all around the world. Several Muslim-majority countries have transitioned to democracy in the latter half of the 20th cenury, and there are Muslim-majority democracies in multiple diverse regions across the world.
Survey research shows popular support for democracy among Muslims, and that Muslims are mostly supportive of democracy as a form of government, and they do not view democracy as incompatible with their religious principles or institutions. Furthermore, democracy itself is in a constant state of struggle to survive. Jones describes democracy not as an outcome, but a process toward resilience. Democracies are constantly undergoing a test of vitality.
There are two key dimensions to typologizing varieties of democratic vitality. The first is duration: the length of time that a country maintains a certain level of democracy since its initial transition to democracy. Both geographically and temporarily, it’s important to consider the context of that particular democratic state. The second dimension is trajectory: the overall trend in a country level of democracy since its initial transition to democracy. Trajectory measures how consistently a country has improved or maintained the level of democracy over time, since its transition.
To measure the level of democracy, Jones uses the Varieties of Democracy Electoral Democracy Index (DDI). This score focuses on the role of elections as the core feature of democracy, and includes aspects of the political system that increase the likelihood that elections will result in democratic outcomes. She then created a typology based on the dimensions of duration and trajectory, which describes four modes of democratic survival, depicted in the graphic below. Democracies are grouped into categories including striving, thriving, waning, and backsliding.
Striving democracies have short duration, but an upward trajectory. The thriving category is the best case scenario: long duration and upward trajectory. In the waning category there is neither duration, nor trajectory. Democracy is just not taking hold, and this is where you might see the transition away from democracy. Democracies in the backsliding category have long duration, but have a downward trajectory.
Jones investigated eight Muslim majority countries and fit them to these modes: Albania, Malaysia, Mali, Tunisia, Indonesia, Senegal, Kyrgystan, and Turkey. She found, surprisingly, that for that most of the Muslim-majority countries in the sample were striving are thriving.
The key takeaway from this research is that democracy is an ongoing struggle to survive. Jones challenged the audience not to think about democracy as meeting some threshold, but rather as a sort of ongoing struggle, and to think about it as varying degrees of vitality, as opposed to focusing on the mortality of democracy. This, she concludes, allows us to have some degree of cautious optimism. Democracy faces constant challenges; survival is just a matter of the degree of the threat and the strength of the institutions meeting that threat.
Vincent Hutchings delivered the inaugural lecture of the Hanes Walton Jr. Collegiate Professorship in Political Science and Afroamerican and African Studies on March 31, 2021.
Vincent Hutchings and Anne Curzan at the inaugural lecture on March 31, 2021.
Swedish economist Gunnar Myrdal, the author of the landmark study An American Dilemma, believed that education was the key to racial progress. Myrdal wrote “There is no doubt, in the writer’s opinion, that a great majority of white people in America would be prepared to give the Negro a substantially better deal if they knew the facts.”
Vincent Hutchings began his inaugural lecture of the Hanes Walton Jr. Collegiate Professorship in Political Science and Afroamerican and African Studies by citing this quote, and immediately calling it into question. Is it true that policies aimed at alleviating the racial wealth disparities would gain more support if only whites knew the true extent of the gap?
A noteworthy aspect of the recent protests against police brutality and racial inequality has been the extent to which white Americans were participants in greater numbers than in the past. Even before the events of 2020, numerous surveys show whites becoming more racially progressive and acknowledging the reality of racial bias against Blacks. But does that mean that whites will support policies to reduce racial disparities? Professor Hutchings suggests that this is an example of a principle-policy gap; whites are prepared to acknowledge that injustice exists in principle, but they may not be willing to support redistributive policies.
Research has shown that both Blacks and whites underestimate the racial wealth gap. Hutchings and his research team, which includes Sydney Carr, Kamri Hudgins, Zoe Walker, wanted to find out what would happen if they could correct misperceptions of the racial wealth gap. Would informing people about the size of the racial wealth gap influence opinions about policies to address the gap?
The research team posited that the way the information was framed would be important. Past studies and media headlines about wealth disparities tend to frame the issue in terms of Black disadvantage. It’s less common to see a focus on the ways whites are advantaged.
A first study, conducted in June 2020, surveyed 2,024 online respondents. Respondents were randomly assigned either to a control group, where they were merely provided a definition of the racial wealth gap, or to one of two treatment conditions that provided a definition of the racial wealth gap, as well as textual and visual information on the current size of the Black/white racial wealth gap based on information from the 2017 Survey of Consumer Finances. The first treatment group received information framed in terms of Black disadvantage; the second treatment group received information with a white advantage frame.
The experiment was informative, but not persuasive. White participants in the treatment groups that received more information about the size of the racial wealth gap subsequently showed a better understanding of the issue. Black participants in the treatment groups also showed a fairly accurate assessment of the wealth gap, with less difference between the control and treatment groups. Neither whites nor Blacks in the treatment groups adopted different positions on racial policy. The one policy area where the study found some change was reparations; white respondents who received more information about the wealth gap were slightly more supportive of reparations.
If the study participants were informed about the wealth gap, but did not change their support for policies to address it, perhaps the treatments were not strong enough? In open-ended questions in the first study, some participants responded that they believed the wealth gap existed because whites worked harder than Blacks. To address these issues, the research team launched a second study in March 2021, building on the first study and specifically highlighting the limited impact of education on reducing the racial wealth gap.
The second study included four treatment groups and a control group.
Treatment group 1 received information about the wealth gap framed in terms of Black disadvantage.
Treatment group 2 received information about the wealth gap framed in terms of white advantage.
Treatment group 3 received information about the wealth gap framed in terms of Black disadvantage, as well as information that Black household heads with a college degree have the same wealth as white household heads without a high school diploma.
Treatment group 4 received information about the wealth gap framed in terms of white advantage, as well as information that white household heads without a high school diploma have the same wealth as Black household heads with a college degree.
Whites across the ideological spectrum are informed by the treatments in the second study. Surprisingly, the additional information about education was less effective than the general population frames.
Again, the results showed that a better understanding of the racial wealth gap does not make participants more progressive on racial policies. However, those in the first treatment group (receiving information framed in terms of Black disadvantage) showed mild increase in support for reparations.
Surprisingly, white liberals in the fourth treatment group (receiving information about the wealth gap and education, with a frame of white advantage) had a threat response to the treatment. When responding to questions about competition for jobs and political influence, white liberals in the fourth treatment group perceived more competition between racial groups. They responded to questions about racism, such as whether whites have greater wealth because they are naturally superior and whether the use of racial epithets by whites was acceptable, in less progressive ways.
These studies show the limits of providing information about racial disparities. People are not necessarily persuaded to close the gaps when they know more about them. When asked if he was discouraged by the outcome of the studies, Professor Hutchings said that as an educator, he is an optimist. He said that he is motivated by his students to keep asking questions about political opinions and behavior. Hutchings and his students will field new surveys to further examine these questions in the fall of 2021.
Post developed by Nicholas Valentino and Katherine Pearson
Survey research is an ever-evolving field. Technology has increased the number of ways to reach respondents, while simultaneously reducing response rates by freeing people from the constraints of one land-line telephone per household. Surveys remain an essential tool for making inferences about societal and political trends, so many survey researchers offer incentives to survey respondents in order to ensure a large and representative sample. Financial incentives to complete surveys, in turn, entice some people to respond to a large number of online surveys on a regular basis, essentially becoming professional survey respondents.
Survey methodologists have carefully considered the ways that survey modes may impact the way people answer questions. Talking to a real person is different than answering questions online. But less is known about how individual factors bias participation in surveys in the first place. For example, might personality traits shape your willingness to agree to answer a survey online versus someone who comes to your door? New work from researchers at the University of Michigan and Duke suggests in fact this is the case.
In order to examine the personality traits of survey respondents, the research team used data from the 2012 and 2016 American National Election Studies (ANES). During these two study periods, the ANES ran parallel and face-to-face surveys. In both years, the ANES included the 10-item personality inventory (TIPI), which consists of pairs of items asking respondents to assess their own traits. Based on the responses, respondents build a profile of “the Big Five” personality traits: openness to experience, conscientiousness, extraversion, agreeableness, and emotional stability.
Big Five traits with corresponding TIPI qualities
Trait
TIPI Qualities
Coding
Openness to experience
Open to new experiences, complex
Conventional, uncreative
Original
Reversed
Conscientiousness
Dependable, self-disciplined
Disorganized, careless
Original
Reversed
Extraversion
Extraverted, enthusiastic
Reserved, quiet
Original
Reversed
Agreeableness
Critical, quarrelsome
Sympathetic, warm
Reversed
Original
Emotional Stability
Anxious, easily upset
Calm, emotionally stable
Reversed
Original
Researchers were able to compare responses to the TIPI with measures of political predispositions and policy preferences, based on responses to questions on the ANES. These include partisanship, liberal–conservative ideology, issue self-placements, and other measures of political orientation.
Based on these data, the authors found that respondents in the online samples were, on average, less open to experience and more politically conservative on a variety of issues compared to those responding to face-to-face surveys. They also found that the more surveys a respondent completed, the lower they scored on measures of openness. Given that professionalized survey respondents comprise the majority of online survey samples, these results suggest caution for those who would like to generalize results to the population at large. It is not enough to balance samples on simple demographics. Attitudinal and personality based differences might also lead online sample estimates to diverge from the truth.
It is difficult to say whether online survey respondents or face-to-face respondents are more representative of personality traits in the general population. If personality is a factor in whether someone will participate in a survey, that might bias both types of samples. However, the authors note that the data suggest that professional online samples are the outlier. They find “that samples based on fresh cross-sections, both face-to-face and online, yield better population estimates for personality and political attitudes compared to professionalized panels.” While it may be possible to mitigate the potential sampling bias of personality traits, it is important that survey researchers understand the role that personality traits play in professional online samples.
