Who do Americans actually want in politics? Part 1
With analysis by Michela Arena
It’s time to talk about the unpopular elephant in the room. It’s now unavoidable that politicians, across the board, are not popular with the general public. Congress has had a net negative rating for the last twenty years. Both parties’ candidates for US President are distinctly unpopular, even with a chunk of their own voters. Even governors, often among the more popular elected officials, are very rarely garnering approval ratings north of 60% in recent years.
While the public may be down on politicians, we wanted to figure out who they would rather have in positions of leadership. What kind of people does the American public want to see run for office in the next few years?
To investigate this, we constructed an experiment, featuring profiles of regular American citizens, including details on their demographics, their family life, their ideology, but importantly, not their partisan affiliation. By not labeling them with Democratic or Republican party affiliation, we’re ensuring that people have to evaluate other aspects of their background and ideology. We wanted to see if any kinds of people could appeal to the starkly divided American electorate.
This is the first of a series on what kind of people Americans want in office. In this first part, we investigated the overall effects and patterns apparent in the general American population.
Methodology and conjoint analysis
This kind of experimental design allows us to assess how different aspects of a profile impact the way people view it. Conjoint experiments like this can focus on a product, a job applicant, or in this case, a potential politician. First we created a list of attributes that could hypothetically affect the way that people would form their opinions and generated multiple options for each of those attributes. Then, when the participant sees the experiment in the survey, they see a profile with various attributes, which have been randomly selected for each profile. This kind of experimental design allows us to identify which attributes are the most influential on shaping participants’ perceptions and determine which features of a profile have the greatest effect on how people form those opinions.
In this experiment, we created profiles of regular Americans, and we investigated whether any of these attributes would make them more or less desirable as political candidates. We included variations of these attributes:
Gender
Age (between 27 and 85)
Ethnicity
Current Job
Including Doctor, Lawyer, Accountant, Mayor of a small city, Business consultant, Small business owner, Investment banker, Teacher, Nurse, Delivery-truck driver, Janitor, Farm worker, and Bartender
Upbringing
Class, including working class, middle class, upper middle class, and upper class
Region, including the Midwest, South, Northeast, Pacific Northwest, Mountain West, and West Coast
Marital status
Including married, unmarried, and living with partner (this attribute accounted for gay and straight individuals)
Parental status, including 0 to 3 children
Ideological statements on immigration, economics, environment, and national history
We fielded this survey experiment to 2,006 adults in the US, between 7 and 15 February 2024. Each participant viewed 4 potential candidate profiles and rated them in terms of whether or not they’d like them to run for office, how Democratic or Republican they seem, their ability to represent their interests, whether or not they thought this could be a real candidate, and whether they would like to vote for them.
This is an example of the profiles rated by participants in our experiment:
Results
The results of this study illustrate what matters most to Americans, in the aggregate, when evaluating their next leaders:
Ideology (generally) overwhelms demographics. The most influential factors on whether or not we want this hypothetical candidate to run for office are related to their ideological perspectives rather than their demographic characteristics.
Criticism of American history is not a vote winner. Candidates who said there was nothing to admire in America’s national history, just violence and oppression, were not popular. This ideological statement was the strongest negative influence on voters’ appetite to see that candidate run for office.
Environmental denialism is not desired. Hypothetical candidates who deny humans’ impact on the climate were much less attractive to American voters. Overall, Americans were much less likely to see climate-change-denying candidates as someone who would represent their interests.
Immigration is complicated. Overall, the least desirable candidate statement on immigration was the most welcoming towards immigrants, and the most desirable statement on immigration emphasized a need for control, safety, and security at our borders. Stopping the flow of illegal immigration was a popular statement popular overall, and yet calling the treatment of immigrants “immoral and impractical” also made potential candidates more appealing. In our next post on this topic, we’ll dig into this area more.
The young should run. Americans want younger candidates. This is the only demographic factor that was nearly as influential as the ideological statements referenced above. What we don’t know is whether this is a general (and consistent) preference among American voters or whether this is a result of the electorate’s disillusionment with the two oldest candidates for US President in American history. This age effect could potentially be a quirk of this specific electoral context. Comparison with other timeframes and national contexts would be necessary to conclude whether voters in general want younger candidates or if the current American electorate specifically has a prejudice against older candidates.
Mountain West is best. Somewhat surprisingly, the most desirable region for a candidate to be from is the Mountain West. None of the other regions had a significant effect on whether voters wanted them to run for office, meaning no region faced discrimination or received a particular benefit.
Marital status might still matter. Another surprising result relates to the marital status of the candidate. Those who were unmarried and living with a male partner were less desirable as candidates to voters. This variable might be picking up on both homophobia towards gay men and traditional judgment of unmarried women living with partners. No other marital status had a significant effect on whether or not voters wanted these candidates to run for office.
There’s a lack of explicit class judgment. Overall, voters didn’t reward or punish hypothetical candidates based on their class, in terms of job and class upbringing. White collar jobs, like lawyer and business consultant, didn’t receive preferential treatment or discrimination in this experiment. Likewise, working class jobs, like delivery truck driver and farm worker, also didn’t affect how voters assessed these hypothetical candidates. While these treatments were not statistically significant (at the 0.05 threshold, but they are at the 0.10 threshold), these specific jobs were likely influential (and future research with larger samples sizes should confirm). These effects illustrate a slight class bias. Mayors of a small city (e.g. Pete Buttigieg) were judged more positively, as were doctors. On the other hand, janitors and bartenders were judged more negatively (e.g. AOC).
Some working class professions aren’t seen as realistic backgrounds for a candidate. We also asked participants if there will ever be “a political candidate like this in real life”. The only attributes that made people less likely to see them as a realistic candidate were specific working class jobs. People who worked as janitors, delivery truck drivers, bartenders, and to a lesser degree, farm workers were less likely to be seen as a political candidate in real life. This illustrates the underlying expectation of who will become elected leaders and reveals another implicit anti-working-class attitude.
Demographics are not as influential as one might imagine. When presented with all of these other attributes, the gender and ethnicity of the candidates did not have a consistent effect on whether or not voters wanted that hypothetical candidate to run for office. That’s not to say that within pockets of the population those factors may matter significantly - both in positive and negative ways. However, looking at the overall population, there isn’t a clear pattern of support or discrimination along those lines in this experiment.
This chart illustrates the estimated effect each candidate attribute has on how much participants wants the hypothetical candidate to run for office. The blue bars indicate statistical significance (at the 0.1 level), and the red bars indicate those treatments were not statistically significant. Bars to the right indicate that treatment had a positive effect on participants wanting them to run for office, and bars to the left indicate a negative effect. The relative size of the bars illustrate how strong of an effect each attribute had on how participants viewed the hypothetical candidates with those characteristics.
These results echo other recent research that suggests demographics are less influential on how we judge politicians than the narrative of “identity politics” might suggest. Instead, voters pay much more attention to potential leaders’ ideology, which tells us what kind of society they’d like to bring about. Specific aspects of ideology have been found to be especially influential, such as ones that tap into issues of national identity, right and wrong, as well as power and resource distribution. That being said, judgements based on ethnicity, class, and gender clearly persist in contemporary politics, and such judgements impact the lives of elected officials. Experiments like this illustrate that such judgements and prejudices are not uniform and consistent across the general population, but do not discount the considerable effect that a relatively small but loud group of people can have on the lives of elected officials.
Ultimately, our research shows that in this current world of unpopular politicians, the solution isn't as simple as providing voters with candidates who match them demographically or even tick a number of demographic boxes. The only demographic that had any real impact on voters’ perceptions was consistent for the whole population, and this might be a quirk of the 2024 US electoral context. Regardless of the voters' age themselves, there was a consistent desire for slightly more youthful and spry candidates.
As we will see in our next post, voters want candidates who match them ideologically, not necessarily demographically. When the electorate is as complex and dynamic ideologically as the American population, it is increasingly impossible to please everyone, let alone a majority of the population.
In our next post in this series, we will dig into how different groups of Americans want different types of people in elected office.
Photo by Element5 Digital on Unsplash