The popularity of the GOP front-runner can be explained by the Dunning-Kruger Effect.
Many commentators have argued that Donald Trump’s dominance in the GOP presidential race can be largely explained by ignorance; his candidacy, after all, is most popular among Republican voters without college degrees. Their expertise about current affairs is too fractured and full of holes to spot that only 9 percent of Trump’s statements are “true” or “mostly” true, according to PolitiFact, whereas 57 percent are “false” or “mostly false”—the remainder being “pants on fire” untruths. Trump himself has memorably declared: “I love the poorly educated.”
But as a psychologist who has studied human behavior—including voter behavior—for decades, I think there is something deeper going on. The problem isn’t that voters are too uninformed. It is that they don’t know just how uninformed they are.
Psychological research suggests that people, in general, suffer from what has become known as the Dunning-Kruger Effect. They have little insight about the cracks and holes in their expertise. In studies in my research lab, people with severe gaps in knowledge and expertise typically fail to recognize how little they know and how badly they perform. To sum it up, the knowledge and intelligence that are required to be good at a task are often the same qualities needed to recognize that one is not good at that task—and if one lacks such knowledge and intelligence, one remains ignorant that one is not good at that task. This includes political judgment.
We have found this pattern in logical reasoning, grammar, emotional intelligence, financial literacy, numeracy, firearm care and safety, debate skill, and college coursework. Others have found a similar lack of insight among poor chess players, unskilled medical lab technicians, medical students unsuccessfully completing an obstetrics/gynecology rotation, and people failing a test on performing CPR.
This syndrome may well be the key to the Trump voter—and perhaps even to the man himself. Trump has served up numerous illustrative examples of the effect as he continues his confident audition to be leader of the free world even as he seems to lack crucial information about the job. In a December debate he appeared ignorant of what the nuclear triad is. Elsewhere, he has mused that Japan and South Korea should develop their own nuclear weapons—casually reversing decades of U.S. foreign policy.
Many commentators have pointed to these confident missteps as products of Trump’s alleged narcissism and egotism. My take would be that it’s the other way around. Not seeing the mistakes for what they are allows any potential narcissism and egotism to expand unchecked.
In voters, lack of expertise would be lamentable but perhaps not so worrisome if people had some sense of how imperfect their civic knowledge is. If they did, they could repair it. But the Dunning-Kruger Effect suggests something different. It suggests that some voters, especially those facing significant distress in their life, might like some of what they hear from Trump, but they do not know enough to hold him accountable for the serious gaffes he makes. They fail to recognize those gaffes as missteps.
Here is more evidence. In a telling series of experiments, Paul Fernbach and colleagues asked political partisans to rate their understanding of various social policies, such as imposing sanctions on Iran, instituting a flat tax, or establishing a single-payer health system.
Survey takers expressed a good deal of confidence about their expertise. Or rather, they did until researchers put that understanding to the test by asking them to describe in detail the mechanics of two of the policies under question. This challenge led survey takers to realize that their understanding was mostly an illusion. It also led them to moderate their stances about those policies and to donate less money, earned in the experiment, to like-minded political advocacy groups.
Again, the key to the Dunning-Kruger Effect is not that unknowledgeable voters are uninformed; it is that they are often misinformed—their heads filled with false data, facts and theories that can lead to misguided conclusions held with tenacious confidence and extreme partisanship, perhaps some that make them nod in agreement with Trump at his rallies.
Trump himself also exemplifies this exact pattern, showing how the Dunning-Kruger Effect can lead to what seems an indomitable sense of certainty. All it takes is not knowing the point at which the proper application of a sensible idea turns into malpractice.
For example, in a CNBC interview, Trump suggested that the U.S. government debt could easily be reduced by asking federal bondholders to “take a haircut,” agreeing to receive a little less than the bond’s full face value if the U.S. economy ran into trouble. In a sense, this is a sensible idea commonly applied—at least in business, where companies commonly renegotiate the terms of their debt.
But stretching it to governmental finance strains reason beyond acceptability. And in his suggestion, Trump illustrated not knowing the horror show of consequences his seemingly modest proposal would produce. For the U.S. government, his suggestion would produce no less than an unprecedented earthquake in world finance. It would represent the de facto default of the U.S. on its debt—and the U.S. government has paid its debt in full since the time of Alexander Hamilton. The certainty and safety imbued in U.S. Treasury bonds is the bedrock upon which much of world finance rests.
Even suggesting that these bonds pay back less than 100 percent would be cause for future buyers to demand higher interest rates, thus costing the U.S. government, and taxpayer, untold millions of dollars, and risking the health of the American economy.
This misinformation problem can live in voters, too, as shown in a 2015 survey about the proposed Common Core standards for education. A full 41 percent claimed the new standards would prompt more frequent testing within California schools. That was untrue. Only 18 percent accurately stated that the level of testing would stay the same. Further, 35 percent mistakenly asserted that the standards went beyond math and English instruction. Only 28 percent correctly reported that the standards were constrained to those two topics. And 34 percent falsely claimed that the federal government would require California to adopt the Common Core. Only 21 percent accurately understood this was not so.
But what is more interesting—and troubling—were the responses of survey takers who claimed they knew “a lot” about the new standards. What these “informed” citizens “knew” trended toward the false rather than the true. For example, 52 percent thought the standards applied beyond math and English (versus 32 percent who got it right). And 57 percent believed the standards mandated more testing (versus 31 percent who correctly understood that it did not). These misconceptions mattered: To the extent that survey takers endorsed these misconceptions, they opposed the Common Core.
My research colleagues and I have found similar evidence that voters who think they are informed may be carrying a good deal of misinformation in their heads. In an unpublished study, we surveyed people the day after the 2014 midterm elections, asking them whether they had voted. Our key question was who was most likely to have voted: informed, uninformed, or misinformed citizens.
We found that voting was strongly tied to one thing—whether those who took the survey thought of themselves as “well-informed” citizens. But perceiving oneself as informed was not necessarily tied to, um, being well-informed.
To be sure, well-informed voters accurately endorsed true statements about economic and social conditions in the U.S.—just as long as those statements agreed with their politics. Conservatives truthfully claimed that the U.S. poverty rate had gone up during the Obama administration; liberals rightfully asserted that the unemployment rate had dropped.
But both groups also endorsed falsehoods agreeable to their politics. Thus, all told, it was the political lean of the fact that mattered much more than its truth-value in determining whether respondents believed it. And endorsing partisan facts both true and false led to perceptions that one was an informed citizen, and then to a greater likelihood of voting.
Given all this misinformation, confidently held, it is no wonder that Trump causes no outrage or scandal among those voters who find his views congenial.
But why now? If voters can be so misinformed that they don’t know that they are misinformed, why only now has a candidate like Trump arisen? My take is that the conditions for the Trump phenomenon have been in place for a long time. At least as long as quantitative survey data have been collected, citizens have shown themselves to be relatively ill-informed and incoherent on political and historical matters. As way back as 1943, a survey revealed that only 25 percent of college freshmen knew that Abraham Lincoln was president during the Civil War.
All it took was a candidate to come along too inexperienced to avoid making policy gaffes, at least gaffes that violate received wisdom, with voters too uninformed to see the violations. Usually, those candidates make their mistakes off in some youthful election to their state legislature, or in small-town mayoral race or contest for class president. It’s not a surprise that someone trying out a brand new career at the presidential level would make gaffes that voters, in a rebellious mood, would forgive but more likely not even see.
But the Dunning-Kruger perspective also suggests a cautionary tale that extends well beyond the Trump voter. The Trump phenomenon may provide only an extravagant and visible example in which voters fail to spot a political figure who seems to be making it up as he goes along.
But the key lesson of the Dunning-Kruger framework is that it applies to all of us, sooner or later. Each of us at some point reaches the limits of our expertise and knowledge. Those limits make our misjudgments that lie beyond those boundaries undetectable to us.
As such, if we find ourselves worried about the apparent gullibility of the Trump voter, which may be flamboyant and obvious, we should surely worry about our own naive political opinions that are likely to be more nuanced, subtle, and invisible—but perhaps no less consequential. We all run the risk of being too ill-informed to notice when our own favored candidates or national leaders make catastrophic misjudgments.
To be sure, I don’t wish to leave the reader with a fatal hesitation about supporting any candidate. All I am saying is trust, but verify.
Thomas Jefferson once observed that “if a nation expects to be ignorant and free in a state of civilization, it expects what never was and never will be.” The Trump phenomenon makes visible something that has been true for quite some time now. As a citizenry, we can be massively ill-informed. Yet, our society remains relatively free.
How have we managed so far to maintain what Jefferson suggested could never be? And how do we ensure this miracle of democracy continues? This is the real issue. And it will be with us far after the Trumpian political revolution or reality TV spectacle, depending on how you see it, has long flickered off the electronic screens of our cultural theater.
David Dunning is an experimental social psychologist at the University of Michigan.