A Better Way to Vote: Applying Selection Science to the Presidential Election

What if we hired our next President the same way we hire for other important jobs? In this post, I’ll cover how selection science could help us make smarter choices at the ballot box.

 One of the foundational elements of a representative democracy is that citizens “hire” their public officials at the ballot box. But, unlike the hiring processes in most organizations, political candidates aren’t subjected to a formal evaluation before election day. Some might say that political campaigns are like one long interview, but they have little in common with how interviews should be conducted. This is unfortunate because decades of research show that certain hiring procedures, such as structured interviews, can strongly predict a person’s future performance. With the 2024 Presidential election looming on the horizon, what can we learn from selection science that might help us make more informed decisions at the ballot box?

Evaluating Presidential Candidates Today

First, let’s recap how we currently evaluate Presidential candidates. Although voters often claim to support the candidate whose views align with their own, studies show this is not often the case. Instead, when people are surveyed about their policy positions, their views often differ from those of their preferred candidate. This suggests that many of us vote for the candidate who best aligns with our political identity, such as whether we view ourselves as a Democrat or a Republican. But if we put those identities aside, what information could we currently use to guide our votes?

One existing data point is candidates’ past accomplishments, such as their success in a previous political office or, for political outsiders, their achievements in business or other fields. But it’s difficult to compare apples-to-apples using this information, even if the candidates have held the same political office (as Biden and Trump have). Take their impact on the economy, for example—the economy is influenced by numerous factors outside of the President’s control, from the decisions made by their predecessor to global events like wars and pandemics.

What about the Presidential debates? Unfortunately, as many political experts have conceded, the debates are more noise than signal. Responses are often rehearsed, or candidates revert to their preferred talking points instead of answering the moderator’s questions. Similarly, the winners and losers are often determined by style over substance, such as audience reactions or whether they make a gaffe as benign as looking at their watch.

But what if we asked candidates to complete a series of procedures that would yield valuable information about their ability to perform the job? We already require this for most jobs, so why not for the most important job in the land?

Improving the Evaluation Process

Before I cover some of these procedures, one important note—the research in this post is based on meta-analysis, where researchers calculate a weighted average from a collection of studies involving different jobs and applicant pools. Some argue that jobs are too situationally specific, and we should analyze how each procedure performs in a specific job context instead. But the pool of past U.S. Presidents is too small to do this. Even if it were large enough, determining our performance measure (our outcome of interest) would be challenging, if not impossible. Therefore, this post assumes the concept of validity generalization, where we believe the weighted average from multiple, diverse samples is a better estimate of the true underlying relationship between these procedures and a candidate’s future success.

The following are listed in order of decreasing effectiveness according to their validity estimate—a number ranging from 0 to 1, with higher values indicating improved prediction of future performance.

Structured Interviews (.42)
Most people have experienced an interview, but a structured interview involves one or more forms of standardization, such as asking every candidate the same questions and relying on situational or behavioral questions (questions that gauge how a person would react or has reacted in job-relevant scenarios). An example question might be, “Tell us about a time you made a mistake. How did you handle it?” Although debates do this to some degree, they introduce too much noise when moderators do not ask every candidate the same question or allow candidates to respond to others’ responses. Even the choice of moderator can introduce noise. Structured interviews could be conducted by an anonymous person or even by computer using pre-recorded questions. Voters could then receive recordings of candidates’ answers and an anchored rating scale for evaluating them. Though impractical, an even more ideal process would be to provide candidates’ answers without revealing their identities, allowing voters to rate candidates based solely on the interview. 

Job Knowledge Tests (.40)
As their name suggests, job knowledge tests would assess candidates’ knowledge of the role of President. Questions would need to focus on objective facts, such as the names of world leaders, historical events, and the basic functions of the federal government, particularly the Executive Branch.

Biographical Data (.38)
Biographical data (biodata) is information about a person’s past experiences. Of the procedures discussed in this post, biodata most closely represents the information we already use to make voting decisions. But, for biodata to be effective, the procedure must be standardized across all candidates. The biggest obstacle to creating a standardized biodata form for the job of President would be determining which information to assess, as each item should either be shown to predict performance in past Presidents (which is difficult to do, as mentioned above) or at least be conceptually related to the job of President. Possible examples could include, “How many executive leadership positions have you held in your life?” or “How many hours of volunteer work have you done in the last year?”

Work Samples and Other Simulations (.26 to .33)
In a work sample, candidates are evaluated on their ability to perform a simulated work task. For example, someone applying for a teaching job might be asked to develop a lesson plan based on a given set of parameters. Work samples are sometimes integrated into a battery of tests known as an assessment center. Situational judgment tests are a form of low-fidelity simulation in which candidates are given a series of job-relevant situations and must select how they would respond from a series of options (or describe their response in an open-ended format). Regardless of format, simulations are the procedures I would most want to see included in the process of evaluating Presidential hopefuls. Candidates would be given the same series of tasks or situations, and voters could evaluate their responses using an anchored rating scale. For example, candidates could be given an urgent memo describing a major earthquake or international conflict and would need to describe their response or select from a set of available options provided by an aide or cabinet official.

Cognitive Ability Tests (.22)
Much has been said about the two candidates’ cognitive abilities. Although Trump has touted his performance on the Montreal Cognitive Assessment (MoCA), this test assesses cognitive impairment rather than cognitive ability, which is why it is often used as a screening for dementia. Cognitive ability tests assess general intelligence or specific mental abilities, such as quantitative reasoning or reading comprehension. For many years, cognitive ability tests were recognized as the strongest predictors of job performance. However, the strength of this relationship has decreased over time, with recent estimates being less than half of those from earlier studies (.22 versus .51, respectively). This can be attributed to numerous factors. Technology has reduced the need for certain cognitive abilities, such as performing complex calculations by hand. Jobs are also increasingly team-oriented, where performance is often based on collective decision-making rather than any one person’s abilities. Additionally, jobs are increasingly service-based and may rely more on the interpersonal skills needed to work with clients and customers. Despite these trends, Presidents are expected to make important and often time-sensitive decisions, so a cognitive ability test would still be reasonable to include as part of the evaluation process.

Personality Tests (up to .31, depending on the trait assessed)
Although candidates’ personalities are often inferred based on their actions, personality testing provides a more accurate and reliable way to measure dispositional traits. Several types of personality tests could be valuable in evaluating Presidential candidates. The most well-established personality framework is the Big Five, consisting of conscientiousness, agreeableness, neuroticism (or emotional stability), openness, and extraversion. Studies have shown conscientiousness—the tendency to be hardworking, dependable, and achievement-oriented—to be the trait that best predicts job performance overall. Items assessing conscientiousness might include asking a person how much they “do things according to a plan” or “follow a schedule.” Other Big Five traits are likely also important for Presidents, such as neuroticism, which can predict how a person generally responds to stressful situations. Personality-based assessments of emotional ability could also be useful for understanding candidates’ tendencies to express empathy or compassion. Alternatively, assessments of “dark” personality traits, such as narcissism, Machiavellianism, and psychopathy, could help identify potentially destructive leaders.

A final type of personality test that might be useful for evaluating Presidential candidates is personality-based integrity tests, also known as covert integrity tests because they assess traits related to integrity (agreeableness, conscientiousness, and emotional stability) rather than asking about troublesome behaviors outright. This distinguishes them from overt integrity tests, which may include items such as, “Have you ever taken anything from work without telling anyone?” As you might expect, overt tests are more susceptible to faking than their covert counterparts.

Conclusion

As this post has hopefully indicated, there are multiple ways to improve how we evaluate Presidential candidates. As with the hiring processes used in many modern organizations, several complementary procedures should be considered, as no single procedure can predict future job performance perfectly. Some might argue that Presidential candidates should undergo all of these procedures given the importance of the job, but I worry that all that information could potentially overwhelm voters. Therefore, if I had my pick, I would choose a structured interview, a work sample (or other simulation), and a personality test, as I believe these would capture the widest range of important information. If you had yours, which would you choose?

Photo by Sora Shimazaki: https://www.pexels.com/photo/badge-in-ballot-box-5926402/