Jared McDonald, David Karol, Lilliana Mason. Political Behavior. Volume 42, Issue 3. September 2020.
Introduction
What qualities do voters want in a president and how do they go about determining whether a politician has them? Particularly when the characteristics the public desires in a leader are not outwardly observable, can voters find accurate information about candidates, or are they hopelessly biased in pursuit of this information? In this paper, we draw on the literature on political character to argue that Americans desire a politician who both intuitively understands their struggles and has the competence necessary to get things done. When voters determine if a candidate for office has these qualities, their assessment of the background of the politician will play an important role. A politician born to a blue-collar family can more credibly claim to care for struggling Americans than someone who has been rich her entire life. Likewise, a politician who has worked her way up from nothing will likely be viewed by voters as smarter and harder-working than an individual who was born into great privilege. In short, perceptions of personal history affect subjective evaluations of personal character and individual character traits.
While voters’ understanding of politicians’ backgrounds may affect their evaluations of candidates, we show that these understandings may not be grounded in fact, especially when the characteristic is not a visible one like race or gender. Such is the case, we find, with President Trump. On October 2, 2018, the New York Times revealed in a lengthy exposé that Donald Trump received at least $413 million in today’s dollars from his father’s real estate empire over the course of his lifetime. This story undercut President Trump’s claim that his father’s assistance was limited to “a small loan of $1,000,000.” In this article, we show that when such information becomes known to the American public, it leads to reduced evaluations of personal character, and ultimately to lower levels of support for the president.
The report by the New York Times was perhaps the most comprehensive accounting of Trump’s personal finances to date. Yet a substantial amount of information about Fred Trump’s role in his son’s business success was already a matter of public record, even if it failed to receive widespread public attention. Prior to serving very briefly as President Trump’s Communications Director, Anthony Scaramucci had been one of Trump’s fiercest conservative critics. As he pointed out in August 2015, Trump was “an inherited money dude from Queens County,” and as such, could not credibly run an anti-establishment campaign. Yet it is not clear that voters knew as much about Trump’s background as Scaramucci did. While Donald Trump is well-known for his wealth and gold-plated lifestyle, voters remain surprisingly misinformed regarding the process by which he obtained this wealth. These misperceptions, we find, are consequential in how Americans view the president. More generally, these findings suggest scholars should pay greater attention to voter perceptions of candidate characteristics that do matter to voters, but which are not visible, and about which they may be misinformed.
These results may be particularly consequential to the extent that they produce negative candidate evaluations even within the candidate’s own party. As Lavine et al. demonstrate, changes in contemporary evaluations of a party or candidate are capable of introducing partisan ambivalence. Ambivalent voters are found to see the political world more clearly, and to make political choices more carefully. If a revision of voter evaluations of a candidate can generate partisan ambivalence, it may be able to restore some of the clear-thinking required for a functioning democracy. These findings are particularly relevant when a candidate’s background is widely misconstrued.
Trump’s Family Background
Donald Trump is not the first President born to wealth. Many chief executives, including some rated highly by historians and the public, were born to privilege. Trump may be the richest, but given the opacity of his finances and the difficulty of historical comparisons, that is unclear. Where Trump clearly does differ from other scions who ran for President is that his wealth and business experience were his chief credentials for the Presidency, absent any time in public office or military service. This raises the question of the origins of that wealth and the extent to which voters make inferences about Trump’s views and abilities based on it.
Of course, someone born to wealth may have great business acumen in his own right. Our focus in this study is public perceptions of Trump’s family finances however, rather than the accuracy of their understanding of his business record, for multiple reasons. Trump’s business history is a long and tangled one, including indisputable triumphs in branding and marketing as well as several bankruptcies, failed ventures and costly legal settlements. The exact state of the President’s finances remains murky. He has, to date, refused to release tax returns and has a history of questionable financial claims we cannot adjudicate here. Given these informational constraints, whether Trump is a “great businessman” is at best a subjective question.
Yet while the exact state of Trump’s finances is unknown, this is not true of his family background. The President’s father, Fred Trump, was a very wealthy real estate developer. Beyond the fact that Trump’s family wealth and long-term reliance on his father are unambiguous; voters’ perceptions of this aspect of the President’s background are interesting for another reason. Research shows that—rightly or wrongly—voters do view politicians’ class background as relevant and make inferences about their candidates’ policy positions based on it, even when told those candidates’ party affiliations (Carnes and Sadin).
Fred Trump was a builder and landlord specializing in low and middle-income housing in Brooklyn and Queens. In the early 1920s the elder Trump went into business with his mother, who inherited properties from her late husband. By the 1930s, Fred Trump had built hundreds of homes backed by the newly-created FHA. His mass construction techniques led more than one newspaper to term him the “Henry Ford” of real estate by the late 1930s. Fred Trump had large government construction contracts during World War Two, building thousands of housing units for the Navy in Virginia (Blair). A 1946 news report identified him as “one of Brooklyn’s largest builders.” In short, Fred Trump already enjoyed considerable success by the time his son Donald was born in 1946. The elder Trump continued to prosper throughout his son’s childhood and adolescence during the postwar boom. Over the course of his career, he built 27,000 homes.
After a failed attempt at age 23 to become a Broadway producer financed by his father, Donald Trump went into the family business. President Trump has downplayed his father’s role in his business career, asserting it was limited to a “small loan of $1,000,000”. Yet the elder Trump was instrumental in his son’s rise in real estate and other ventures, including casinos. The precise extent of the paternal assistance Donald Trump received is unknown. The New York Times estimated the figure to be at least $413 million, which Donald Trump disputed, but it is clear that it included direct loans far in excess of $1,000,000. Court documents show the younger Trump owed $14,000,000 to entities controlled by his father in 1985. Fred Trump also loaned his son $3,500,000 in 1990 through the device of buying chips when the latter was about to lose control of a casino, an action that resulted in a fine by the New Jersey Gaming Commission. In the 1990s when he was in financial straits and his father was suffering from dementia he borrowed many millions—$9,000,000 according to his own testimony, $30,000,000 by other reports, from his anticipated share of his father’s estate with his siblings’ consent.
Perhaps most importantly, Fred Trump personally guaranteed the large loans his son received in order to develop his early projects and provided indispensable access to key public officials. When Donald Trump began his first venture in Manhattan, which was dependent on tax abatements, his father was very close to both New York Governor Hugh Carey and Mayor Abe Beame, both products of the Brooklyn Democratic organization and longtime recipients of the elder Trump’s campaign contributions (Barrett).
The extent to which Americans are knowledgeable about this history is unclear. In this paper, we posit that American voters infer character traits based on their understanding of politicians’ personal backgrounds. Whether or not they are justified, perceptions of those characteristics are politically consequential. Yet we show that these assessments may be based on misunderstandings or ignorance about the politicians themselves. Many voters are largely unfamiliar with the personal background of Donald Trump, as we demonstrate below. These facts matter to voters, shaping attitudes and support for the forty-fifth president.
The theory advanced here relies on two assumptions grounded in literature from political behavior. First, Americans are not generally knowledgeable about politics (Kinder and Kalmoe). Many Americans are unaware that President Trump was born into a wealthy family. This is the case despite the fact that Trump’s wealth is the basis for his longtime celebrity status. This misperception in turn affects Americans’ assessments of both Trump’s competence and his compassion, and ultimately affects how much voters support him. Second, political traits such as competence, integrity, and empathy affect the approval of political leaders (Holian and Prysby). Perceptions of character traits are not fixed, but can be changed during the course of an election, and these fluctuations can influence voting decisions (McCann).
Literature Review
Americans’ Political Knowledge
Much of the literature on political knowledge views Americans as largely uninformed (Campbell et al; Converse; Kinder and Kalmoe). Beyond lacking opinions on many issues, Americans are often unaware of basic political facts (Anson). They tend to lack information that would be helpful in formulating an opinion (Delli Carpini and Keeter). Compounding this problem, scholars find that the information environment necessary for one to become informed is often missing (see e.g., Jerit et al.). Many people do not look for opposing views that might bolster their level of political information, and instead seek out poor information environments (Kuklinski et al; Mutz). Scholars argue that these forces, when taken together, lead to suboptimal outcomes. Random events, even those outside the control of elected officials, affect their choices (Achen and Bartels). Furthermore, voters reward politicians for responding to crises, rather than preventing them, distorting incentives and creating inefficient outcomes (Healy and Malhotra). Scholars have also found that misinformation is easy to spread and resonates in particular among individuals predisposed to believe in a given lie (Berinsky; Lenz; Nyhan and Reifler). Partisans, in particular, tend to rate themselves as more knowledgeable than they are when their partisan identities are made salient (Anson). Recent research on fact-checking suggests that misperceptions can be corrected without worry of a potential “backlash effect,” but that these updates do not change overall evaluations of politicians (Guess and Coppock; Nyhan et al.)
These features of the American electorate inform the theoretical argument advanced here. We argue that correcting misperceptions regarding the personal history of a politician is different from correcting factual matters regarding policy, since a politician’s biography speaks directly to the type of person a politician is. We therefore examine the ways in which a lack of knowledge, specifically on matters of biography, shape character evaluations and support for political leaders. We find that information that many take for granted, such as Donald Trump’s privileged background, is not known by large portions of the electorate. This knowledge significantly influences how Americans view the President. Our findings contribute to the literature on voter responses to candidates’ personal characteristics and background, which ignores the possibility that voters are misinformed rather than simply ignorant of key candidate characteristics.
The Importance of Political Character
The idea that Americans lack political knowledge and may be unable to hold politicians accountable is not new. Yet unlike prior work, we seek to connect this political knowledge to perceptions of political character, something that has not received much attention in the literature. In light of voters’ informational deficiencies, scholars have focused on the importance of static partisan and group attachments in determining vote choice and approval of elected officials (Campbell et al; Green et al.; Kinder and Kalmoe; Mason). Yet while factors such as partisanship and global evaluations can influence perceptions of political character, prior scholarship has argued convincingly that the causal arrow also runs in the opposite direction. McDermott et al, for example, find that the personal scandals of politicians generate a negative effect on trait evaluations specific to the scandal itself and do not necessarily translate to lower global evaluations of the politician.
Studies looking specifically at voter perceptions of candidate character, however, have been sparse in recent years. Kinder posits that voters would like to attribute some sort of motivation to the actions of a president. This exercise is made easier if the voter believes she understands the type of character the president has. Kinder also produces a framework for classifying important candidate traits, including competence, leadership, integrity, and empathy.
Scholars in subsequent decades have argued over how many dimensions of character truly exist, but they have agreed that competence and empathy are important traits (see e.g., Aaldering and Vliegenthart; Greene; Holian and Prysby). Aaldering and Vliegenthart summarize the literature on political character and find that perceptions of these character traits have a clear and sometimes pivotal impact on vote choice and election outcomes (see also Goren; McCann; Miller and Shanks).
Hayes argues that parties “own” traits much in the same way they “own” issue areas. While trait ownership advantages Democratic candidates on matters of compassion, Hayes and others (e.g. Funk; McCann) find that perceptions are still malleable and candidate-dependent. Johnston et al. show that voter evaluations of Al Gore fluctuated during the 2000 campaign. Negative campaigning, filtered through a media that amplified negative messages, led voters to view Gore as less honest and trustworthy. The argument we make here does not deny the importance of static partisan attachments, nor does it discount the ability of campaigns to develop effective messages around questions of character. Instead, we show how biographical information can influence perceptions of a leader’s character. This in turn affects public approval of that leader beyond the effects of partisanship alone.
We also know that voters infer traits from candidate characteristics beyond party affiliation. A literature too extensive to cite fully explores voter reactions to candidate characteristics and the use of stereotypes in forming judgments. The most frequently explored characteristics include gender (Fox and Oxley; Dolan; Hayes; Huddy and Terkildsen; Koch; Lawless; Sanbonmatsu) and race (Karl and Ryan; Jones; McConnaughy et al; McDermott; Piston; Sigelman et al; Terkildsen). Religion (Berinsky and Mendelberg; Campbell et al), occupation (Campbell and Cowley; McDermott) and military service (McDermott and Panagopoulos; Teigen) have also been shown to affect voters’ judgments of candidates. Some recent studies especially relevant to our research concern voters’ reactions to candidates based on their class background (Carnes and Sadin). To varying degrees, scholars find that candidate characteristics have been relevant for voters, albeit outweighed by party affiliation in many cases (Dolan; Hayes).
However, in general this literature gives much less attention to the effect of voters being misinformed about candidate characteristics. Many studies use experimental designs in which voters cannot be misinformed about the characteristics of candidates who are fictitious. In other investigations observational data is used and the characteristic in question is visible (race and gender) or listed on a ballot, as candidate occupation has been in California (McDermott).
Yet candidate characteristics that are not visible or inferable with a high degree of accuracy from candidates’ names also influence voters. This is true, for instance, for religion and military backgrounds as well as class. Voter knowledge of non-visible characteristics is highly uneven. By July of 1960, 84% of respondents knew that Senator John Kennedy of Massachusetts, the Democratic Presidential nominee, was a Catholic. Yet in July 2004 only 28% knew that Senator John Kerry of Massachusetts was one. Perhaps in this case voters’ lack of information was unimportant, because while Kennedy’s Catholicism repelled many voters and attracted others, this identity had lost political salience by 2004 (Putnam and Campbell).
Yet as we show here using a timely example, voters are not just uninformed, but misinformed about characteristics that do affect their candidate evaluations.
A Theory of Knowledge and Political Character
Of the non-visible characteristics politicians possess, biography is of special importance. While some aspects of candidate biographies are publicized over the course of a campaign, we note that American voters remain ignorant or misinformed regarding important facts about the candidates. This is true even for Donald Trump, who has been famous for decades.
This claim runs counter to the narrative that elections center too much on candidates’ personalities and too little on the issues. In a news cycle that normally provides daily sound bites from the campaigns and focuses on the personal narratives being driven by the candidates, generating an impression as to what kind of person each candidate is should be easy. While pundits can reasonably lament the lack of substance during campaigns, we point out that the emphasis the media places on personal narratives does not necessarily leave voters well-informed about the candidates’ lives. Generating an impression of a politician’s character is dependent on the types of narratives the campaigns choose to emphasize, the kinds of stories the media choose to report, and the information voters seek. As a result, there is little reason to believe that voters are all operating on similar assumptions as to the character of each politician.
Despite these problems, voters strive in every election to develop an image of the type of person a politician is in real life. Holian and Prysby provide an in-depth examination of the importance of character traits in voting decisions. They argue that these traits should matter for three reasons: (1) they are easy for voters to use, (2) voters are encouraged to use them by both the campaigns and a news media that emphasizes candidate character, and (3) it is rational for voters to use them. Taken together, these reasons provide a compelling rationale for why voters might look to candidate character when deciding for whom to vote. While most Americans may not have a strong opinion about which policy will provide the greatest benefit, they can get an idea over the course of a campaign whether a politician is smart, motivated, and understands the problems facing people like them. In this way, character traits such as empathy and competence act as voting heuristics. A vast literature exists on the importance of political cues and heuristics (see e.g., Lupia; Popkin), yet a great deal of it notes how heuristics can be manipulated or lead to incorrect voting decisions (Gilovich et al).
While voters may rely on character traits because they are easy to use, we question the degree to which they can be used reliably. Perceptions of character traits, just like any other heuristic, are malleable and susceptible to manipulation. This is especially true if particular aspects of a candidate’s character are not thoroughly vetted or if voters are misinformed. As we show in the case of Trump, important pieces of the president’s background have gone unnoticed by much of the American electorate. Despite Trump’s omnipresence in the media since 2015, misperceptions still exist that are materially important to his overall image.
The Case of Donald Trump
While Donald Trump’s wealth and business dealings were discussed at some length throughout the 2016 campaign, relatively little attention was paid to his experiences as a child or the assistance he continued to receive from his father well into adulthood. This is in line with the consensus among political communications scholars who note that journalists are not likely to focus on a story unless (1) the campaigns themselves are pushing it (Gans), (2) it fits a normal newsworthiness value such as conflict, simplicity, unusualness, or timeliness (Cook), or (3) it fits a predetermined narrative journalists prefer when describing a particular politician or other prominent person (Graber; Kendall).
The lack of media focus on the Trump family fortune is reflected in what Americans know about the President’s financial history. From 2016 to 2018 across national surveys of Americans, one fielded by Survey Sampling International and the other two through the University of Maryland’s (UMD) Critical Issues Poll (administered by Nielsen Scarborough), we probed the electorate’s awareness of the Trump family wealth around the time Trump was born. While we describe the data more completely in a later section, Fig. 1 reveals that, across all three surveys, many were not aware of Donald Trump’s upbringing.
How might this phenomenon affect political approval? In this paper, we draw on the candidate character literature to identify two pathways through which misperceptions about the president’s biography should affect support. First, we posit that individuals who are unaware of Trump’s privileged background are going to see him as better able to understand, relate, and care about the problems facing everyday Americans. Second, we argue that this misperception should also bolster Trump’s claim that he is a skilled businessman. People who believe Trump was not born wealthy will be more likely to view his success as the consequence of his own hard work and intellect rather than the product of familial privilege.
Trump and Empathy
In discussing the importance of candidate character, Holian and Prysby pay special attention to the perception that a candidate truly cares for others. They define empathy as “the recognition of another person’s emotions, to feel what another feels” (p. 29). Individuals who believe an authority figure truly understands their struggles will also trust that authority to do right by them. An empathetic leader intuitively understands who voters are and what they need. They do not need to be informed of the problems facing Americans by an advisor or pollster. As a result, any leader who can persuade voters that they are able to walk a mile in their shoes will be better able to convince voters that they will enact positive change once in office. This appeal goes beyond any particular policy and hinges instead on a question of character. Bill Clinton, who grew up poor in Arkansas, was famous for his ability to build empathic bonds with others. When he famously told one skeptic, “I feel your pain,” many Americans believed him.
Examining empathy with regard to Trump might strike some as odd. In the media, Trump has been lampooned as incapable of relating to those dealing with hardship. After speaking with survivors of the Marjorie Stoneman Douglas High School shooting, Trump was ridiculed when pictures revealed that staffers had provided him notes on what to say to the students, including lines such as “I hear you” and “What would you most want me to know about your experience.” During the 2016 campaign, he got into a war of words with a gold star family who supported Hillary Clinton. Khizr Khan, the father of fallen U.S. Army Captain Humayun Khan, accused Trump of being “without empathy” for Americans. Even Trump defenders have not been able to argue that Trump has a particularly empathetic style. In defending Trump, Gary Abernathy of the Washington Post admitted, “It is true that Trump displays little outward empathy.” Yet Abernathy goes on to say that many Americans can relate to the notion that we shouldn’t be overly demonstrative with our feelings. In the heartland, Abernathy claims, “they do not wear their emotions on their sleeves. They greet strangers with a subtle nod rather than a showy hug or kiss on the cheek.” While Trump may not be particularly demonstrative, then, it does not mean to voters that he lacks empathy.
This notion is key. Despite the frequency with which pundits discuss Trump’s lack of empathy, many Americans view him as truly caring for others. One source of this view, we argue, lies in misperceptions regarding his background. We expect that Americans who believe Trump was not born rich should view him as more compassionate. Few voters view themselves as rich, so any misperception that Trump was not born wealthy should represent a commonality the voter shares with the president. This bond fuels the perception that Trump can relate to the issues that matter most to everyday Americans. This leads to our first set of hypotheses:
H1a
As Americans become better informed of Donald Trump’s privileged background, they will view him as less compassionate.
H1b
As Americans view Donald Trump as less compassionate, they will be less likely to support him in office.
Trump as a Good Businessman
Empathy is not the only trait that matters to voters. Research also shows that voters want a strong leader who is competent (Funk; Holian and Prysby; Kinder). The reason for this is straight-forward: a competent leader should be more effective in office than one who only “cares” about people but lacks the know-how to accomplish anything. For both Clinton and Trump in 2016, competence was a trait that the campaigns tried to emphasize, leaning on personal stories to support their claims of intelligence and knowledge.
We argue that there is a second pathway by which misperceptions of Trump’s background should influence approval. While believing Trump was born rich should affect perceptions of empathy, it should also affect the perception that Donald Trump is a skilled businessman. After all, it takes more ingenuity to build a business from nothing than to grow an already considerable family fortune. Stories of Trump’s privileged background and the numerous times Fred Trump intervened to help his son’s business went largely unnoticed by the media and the voters in 2016. When Americans are informed of these stories, as we later show, they undercut perceptions of Trump as a strong businessman.
This notion appears to be something of which Donald Trump himself is keenly aware. A prominent storyline pushed by his campaign in 2016 was the idea that he built the Trump Organization mostly on his own, demonstrating his brilliance in negotiating and business. “It’s not been easy for me,” Trump argued, “and you know I started off in Brooklyn, my father gave me a small loan of a million dollars.” Trump continuously pointed out that the loan was “very, very small” and he “built that into a massive empire.” While getting substantive and sustained backing from his father likely undercut perceptions of him as a shrewd businessman, his childhood experience may have played just as big a role. Americans are no doubt aware that children of rich families are afforded many advantages that make it easier for them to be successful in later life. Americans who know that Trump was born into a wealthy family should be less likely to view his success as a result of his own abilities. We therefore reason that knowledge of the Trump family wealth will diminish the perception that he is a skilled businessman. This leads to our second set of hypotheses:
H2a
As Americans become better informed of Donald Trump’s privileged background, they will view him as less competent in business.
H2b
As Americans view Donald Trump as less competent in business, they will be less likely to support him in office.
Research Design
In order to test these hypotheses, we use data from three surveys that use national samples of the American eligible voting public. The first sample was collected in August 2016 by Survey Sampling International (SSI). The remaining two samples were collected from the UMD Critical Issues Poll using Nielsen Scarborough’s probability-based panel, which was originally recruited by mail and telephone using a random sample of adults provided by SSI. Responses in the Critical Issues Polls are weighted by age, gender, income, education, race, geographic region, and partisan identification.
In Study 1, we examine the possibility that knowledge of Trump’s background represents expressive motivations. That is, we assess whether Republicans may be motivated to say that Trump was born less wealthy simply because they like him, and Democrats conversely may be motivated to say he was born wealthy even if they are not sure. In Study 2, we employ a mediation analysis (described in more detail in Imai et al.) using observational data to examine the impact of knowledge about Trump’s background on presidential approval, mediated through perceptions of him as both empathetic and good at business. In Study 3, we seek to reduce any concerns of endogeneity by employing an experimental design. Leveraging Americans’ ignorance of the President’s background, we provide some respondents with information regarding Fred Trump’s role in his son’s success and use that treatment to examine the effect of knowledge on evaluations of Trump’s character traits. All dependent variables in subsequent analyses have been standardized, such that 0 is the lowest value for any response and 1 is the highest value. Differences between groups, then, can be interpreted as the percent change across the response scale.
Study 1: August 2016 Poll
To assess the factors beyond partisanship that might influence knowledge of President Trump’s background, we had 2500 respondents answer a web-based survey conducted by SSI from August 18-23, 2016. SSI maintains a panel of respondents, which it recruits through a website in return for incentives. Since recruitment into the panel is voluntary, the sample may be unrepresentative of the national population. However, sample matching on census region, age, gender, and ethnicity was employed to draw a close to nationally representative sample from the larger, non-representative sample. Respondents were asked, “When Donald Trump was a child, his family was…” Answer options included poor, working class, middle class, upper-middle class, and very wealthy. They were similarly asked four questions intended to measure their political knowledge broadly. Finally, they were asked items gauging the degree to which they paid attention to political news as well as a host of demographic questions.
Study 2: November 2017 Poll
For the mediation analysis in this study, we rely on a survey of 977 respondents collected by the UMD Critical Issues Poll from November 1-6, 2017. Respondents were asked the same question as the item from Study 1 that gauged knowledge of the Trump family wealth during Donald Trump’s childhood. Respondents were then asked to describe how well or poorly particular statements described Donald Trump, including “he really cares about people like me” and “he is a good businessman.” Survey subjects also rated the degree to which they approved or disapproved of the job Donald Trump was doing as president. Because we theorize that knowledge of Trump’s background influences approval through perceptions of his character traits, we follow the framework set forth by Imai et al. and present our results in the next section by looking at the average causal mediation effect (ACME) and the proportion of the total effect that is mediated. Both mediation analyses use the same controls for partisanship, age, education, income, gender, and race.
Study 3: June 2018 Survey Experiment
In addition to the mediation analysis, we examine the impact of treating respondents with information regarding Fred Trump in order to determine the degree to which Americans’ perceptions of Trump change in light of this knowledge. The UMD Critical Issues Poll collected a sample of 809 respondents from June 1-5, 2018. Respondents were randomly assigned to one of two conditions. In the control condition, respondents received no treatment. They were not asked anything regarding Trump’s personal history, so the advantages afforded Trump by his father’s wealth should not dominate their evaluations of Trump. In the treatment condition, respondents were asked the following:
To what extent were you aware that Donald Trump grew up the son of wealthy real estate businessman Fred Trump, started his business with loans from his father, and received loans worth millions of dollars from his father in order to keep his businesses afloat?
This question was designed to impart critical information that we theorize should influence perceptions of Trump as an empathetic person and a good businessman. After the treatment, all respondents were asked the character trait questions from Study 2, involving the degree to which Trump “cares about people” and “is a good businessman.”
Results
In order to establish the importance of background knowledge in the formulation of attitudes toward political actors, we first establish that the expression of this knowledge is not simply attributable to motivated reasoning. Second, we show that knowledge of the President’s background affects public approval, mediated through perceptions he is empathetic and good at business. Finally, we show that when Americans are exposed to evidence that Donald Trump’s success is at least in part a result of his father’s help, they see him both as less empathetic and less competent at business.
Is “Knowledge” Just Expressive Responding?
To assess the impact of knowledge on public approval, we first consider a possible alternative hypothesis to the one we have presented here. It is quite likely that Democrats will be motivated to say Donald Trump was born rich even if they are not sure, while Republicans will conversely feel the pressure to say he was not born rich even if they know he was. Scholars in the misinformation literature have argued that partisans feel pressure to express attitudes even on factual questions that appear to support their party or leader (Schaffner and Luks). Further, the motivated reasoning literature explains that partisans will intentionally look for the most positive information they can find about their own party (see e.g., Redlawsk; Taber and Lodge). If expressive responding or motivated reasoning were the only influence on presidential character perceptions, it is possible that any effect we find for knowledge of Trump’s background may instead simply be picking up the effect of party.
In spite of the powerful influence of partisanship, a number of important characteristics beyond party can influence political knowledge, including education, age, income and gender (Delli Carpini and Keeter; Jerit et al). Using the 2016 SSI dataset, we model knowledge of President Trump’s upbringing using both political and demographic features known to influence political knowledge. Table 1 presents a probit model predicting whether respondents correctly answer the question about the Trump family wealth. Here, even after controlling for the powerful effects of partisanship, factors such as the closeness with which one generally follows politics as well as demographic factors such as income and gender still exert a significant influence on misperceptions regarding Trump’s background.
Probit model: knowledge of Trump’s family wealth by demographic predictors
| Variable | Coefficient | Std. error | p value |
| Party ID (Democrat) | 0.282** | 0.053 | 0.000 |
| Follow politics | 0.195** | 0.031 | 0.000 |
| Education | − 0.007 | 0.019 | 0.354 |
| Age | − 0.001 | 0.002 | 0.192 |
| Family income | 0.017** | 0.007 | 0.008 |
| Gender (male) | 0.127** | 0.052 | 0.007 |
| Constant | − 0.903** | 0.131 | 0.000 |
| Prob > χ2 | 0.000 | ||
| N | 2446 |
**Statistically significant at p < 0.01, one-tailed test
To further assess the sources of misperceptions regarding Trump’s background, we look at their relationship with less politically motivated measures of knowledge. In the 2016 SSI survey, we asked respondents four general political knowledge questions: which party was in charge of the House of Representatives, whether the federal budget deficit had grown or shrunk compared to the 1990s, the length of a U.S. Senate term, and the job title of John Kerry at the time. We present the predicted probability of correctly answering the Trump family wealth question, by level of general political knowledge and partisanship in Fig. 2.
Figure 2 suggests that being a Democrat is in fact related to a significantly higher likelihood of expressing the belief that Trump was born rich. However, despite the effect of partisanship, general political knowledge is also strongly associated with knowledge specific to Trump’s upbringing. Among Democrats, those who answered none of the four questions correctly had about a 35% probability of saying Trump was born rich, compared to a more than 65% probability among those who answered all 4 items correctly. Republicans similarly shift about 30 percentage points moving across the scale of political knowledge. These findings indicate that misperceptions of Trump’s background correlate with other, less partisan measures of political knowledge. Chalking up misperceptions of Trump entirely to expressive answering or motivated reasoning, then, misses an important piece of the story. Even within a single party, knowledge of politics is associated with knowledge of the President’s background.
Mediation Analyses
We have argued that knowledge should affect president Trump’s approval precisely because supporters believe he is both empathetic and understanding of their needs as well as astute in his business dealings. Believing Trump may be advantaged by his family’s wealth should depress perceptions of both, leading to a lower overall opinion of Trump. To test this argument, we turn to the November 2017 UMD Critical Issues Poll. We regress the standardized measures (on 0-1 scales) for Trump job approval, Trump empathy, and Trump business acumen on perceptions of Trump’s wealth (on a 1-5 scale) and a host of political and demographic controls. Figure 3 provides evidence in support of our theory.
The analyses in Fig. 3 suggest that the effect of knowledge of Trump’s family wealth on Trump’s job approval is mediated through both the perception that he is compassionate and the belief that he is a smart businessman. The mediation effects are all statistically significant at conventional levels and are substantively important. The coefficients suggest that, for every one unit shift in perceptions of the Trump family wealth, from poor to working class to middle class and so on, approval of Donald Trump’s performance in office declines by 5 percentage points. This is true even after controlling for critical factors in evaluations of Trump, such as demographic characteristics and partisanship.
Looking first at the model for empathy, we find that most of the effect for information on approval is mediated through perceptions of empathy. Nearly 3 points of the 5-point effect, or 56%, is mediated. This suggests that those who believe Trump was not born wealthy also believe he is better able to understand the struggles average Americans face.
Second, we find that knowledge of the Trump family wealth also affects Trump’s approval through the perception that he is a good businessman. Just less than 2 points of the 5-point effect, or 31% of the total effect, is mediated through the evaluation of this key character trait. As people learn about Trump’s family wealth, they approve of him less due to his declining credentials as a savvy businessman. This is unsurprising, since Trump repeatedly played up his business acumen as a selling point for his 2016 candidacy.
Experimental Results
Because perceptions of the Trump family fortune are sometimes hard to disentangle from partisanship, we augment the prior observational analyses with an experimental component from our June 2018 survey. If, as we show in the previous section, significant portions of the American electorate are unaware of the Trump family history, and if this information substantially alters perceptions of Trump as a person, then exposing Americans to this information should lead to substantially altered views toward him.
Figures 4 and 5 present the results of our experimental analysis. Respondents in the control condition received no information about the Trump family, while respondents in the treatment condition were told that Trump was born wealthy and was rescued by his father multiple times in his career as his business ventures were failing. Breaking these results out by partisanship is illustrative, as it not only shows the power of partisanship to act as a strong perceptual screen, but demonstrates that this information can have an impact even on those who likely already have strong opinions about Donald Trump.
Among Democrats, the treatment group reports lower evaluations of Trump’s capacity for empathy by a margin of 0.04 points. This margin is not only statistically significant at conventional levels, but substantively large considering the floor effect that is likely at work. On the 0-1 scale of empathy, Democrats in the control condition were already feeling overwhelmingly negative toward President Trump, giving him an average rating of only 0.1. The seemingly small movement of 0.04 in the treatment condition still represents a 40% decrease in perceptions of Trump’s compassion.
With Republicans, no floor effect is at play. Instead, evaluations of Trump’s empathy drop from a fairly high baseline of 0.79 to 0.67, constituting a drop of more than 10 percentage points from the control. This result is important to highlight. Providing candidate background information changes evaluations of the President even among co-partisans. During a time when presidential approval ratings appear to be static and polarized, it is possible to change partisans’ opinions about their own party leader by correcting background information about him.
Figure 5 shows that similar effects are found for perceptions of Trump as a skilled businessman. Because perceptions of Trump’s business acumen are generally higher across the entire sample, there is less of a floor effect among Democrats. Democrats in the treatment condition are roughly 0.06 points more negative on Trump’s business record than Democrats in the control condition. Republicans drop similarly, from a very high baseline of 0.91 to 0.82, representing a drop of roughly 10 percentage points from the control. Again, the treatment effects across both Democrats and Republicans are statistically significant at conventional levels.
While we could not assess the direct effect of the treatments on Trump’s job approval with the Critical Issues Poll, a replication using Amazon’s Mechanical Turk suggests the treatments do, in fact, directly reduce overall job approval. Taken together, these results suggest that individual knowledge of a politician’s background can have significant effects on voters’ evaluations and approval of a candidate or leader. Possibly even more important, these effects can occur within the politician’s party. If these changes result in increased ambivalence toward the candidate or the party, they may have significant electoral results. As Lavine et al. explain, ambivalent voters are able to walk a fine line between apathy and total partisan bias, using more discerning reasoning in making their vote choice. Informing the electorate about the backgrounds of the candidates running for office, then, becomes critically important, particularly when these backgrounds clash with the prevailing story.
Discussion
The 2017 Alabama special election for U.S. Senator involved a Republican candidate, Roy Moore, who had been widely accused of pedophilia. This was a character failing (and a crime) that would destroy nearly any candidate. This failing turned a Senate race in deep-red Alabama into a tight contest. Moore’s deep character flaw, however, was broadly publicized. Voters were aware of these allegations. On Election Day, turnout in Moore’s most loyal districts dropped significantly. Some Republicans even voted for Moore’s opponent. While this is an extreme case, it underscores the point that character can matter in an election, as long as candidate background is well-known.
The current political climate is one in which partisan bias, motivated reasoning, and polarization seem to tell the entire story of American political perceptions. However, the findings presented here suggest, perhaps, a small glimmer of hope for those who believe in the power of information to persuade voters. While our results indicate very large and powerful effects of partisanship on perceptions of the President, we also find evidence that information about the background of elected officials can be consequential to voter perceptions. These results are all the more surprising in an electorate that seems, by all other measures, to be unwavering in its divided partisan loyalties.
According to our results, knowledge of Trump’s family background is capable of changing evaluations and approval of the President, even (or especially) among his own partisans. What does this mean for the study of voter perceptions and candidate evaluations in a time of polarized identities and attitudes? For one thing, it means that accurate information can be politically consequential. At a time when the political media are clashing over the nature of truth, these results provide evidence that this is a worthy fight. As exposés such as the one published by the New York Times gain traction, public perception of President Trump may change. Admittedly, such stories would have likely had a greater impact in the run-up to the 2016 election, when Donald Trump’s record was based entirely on his career as a businessman. In future elections, retrospective considerations concerning Trump’s tenure in office will likely make considerations regarding his character less consequential for vote choice. Despite this, our results suggest that public perceptions are still malleable two years into his presidency. Of course, for these stories to have the expected impact, they must be both seen and believed. Identifying precisely the impact of biographical information in more complex information environments is a topic worthy of future investigation.
During this period of social (Mason) and affective (Iyengar et al) polarization, it is generally understood to be the case that partisans will always vote for their own party’s candidate. This is partly because the opposing party’s candidate is seen as more unacceptable than even a flawed in-party candidate. To a large extent, this is true. However, the results presented here suggest that despite this entrenchment, correcting candidate character traits can make an electoral difference. Simply by introducing some ambivalence in partisan approval, correcting candidate background information can lead voters to think more carefully.
Although the total magnitude of changes in opinion presented here are not enough to switch most voter loyalties, they certainly could make a difference among undecided voters, or those who are wavering in support of a candidate. For those whose support for Trump has declined in the presence of retrospective evaluations of real governing, information that undercuts the main narrative about his talents could have real influence on their ultimate sense of approval. In a close election, this could be pivotal.
The effects presented here may be particularly powerful because they offer contradictory information about a prevalent narrative (Trump’s self-made success). We cannot say here whether or how other types of information relevant to character trait evaluations might change support for other politicians. We only examine one. But we suggest that these findings are sufficient to demonstrate that voter perceptions of candidates can be changed, and that these perceptions are meaningful for voters’ overall approval of those candidates. Future research should examine the effects of other character traits in differing political contexts.