Historian who correctly predicted every election since 1984 makes his call
Fox News, October 30, 2020 — American University Professor Allan Lichtman joins “The Story” to make his prediction on who will win the 2020 presidential election. (4:07)

By Aubrey Immelman
Unit for the Study of Personality in Politics
November 1, 2020

On July 8, 2019, I announced that the Unit for the Study of Personality in Politics’s Personal Electability Index (PEI) projected that none of the seven front-running Democratic contenders assessed at the USPP — in descending order of PEI scores Joe Biden, Kamala Harris, Bernie Sanders, Amy Klobuchar, Beto O’Rourke, Pete Buttigieg, and Elizabeth Warren — would defeat Donald Trump in the 2020 U.S. presidential election.

At the time, this election-outcome forecast was at variance with contemporaneous polls, for example, an ABC News/Washington Post poll, conducted June 28-July 1, 2019, that showed Trump losing against or statistically tied with the front-runners for the Democratic nomination for president:

Among all adults …, Joe Biden leads Trump by 14 points. But that narrows among the other four Democrats tested against Trump in this poll — an 8-point lead for Kamala Harris, a slight 7 points for Elizabeth Warren, 6 for Bernie Sanders and 4 for Pete Buttigieg.

In the past several months, I have frequently had to field inquiries regarding the status of a projection issued so far in advance of the election — nearly 16 months. For example, on September 29, Cambridge, Mass., psychiatrist John Martin-Joy, M.D., as part of his research for a Psychology Today blog on Joe Biden’s mental state, wrote in an email: “I’m curious if at this point you stand by the PEI’s prediction that Trump will win.”

I responded as follows:

Regarding my simple little PEI heuristic, I have no choice but to “stand by” it; after all, Trump’s personality has not changed significantly since I issued my prediction on July 8, 2019 — nor would one expect it to change at his age, save for organic reasons. … The PEI is blissfully unaware of SARS-CoV-2, the economy, the unemployment numbers, or Trump’s tax returns leaked to the NYT shortly before the election. An implicit assumption of the PEI model is a relatively typical or “normal” election cycle — say (metaphorically) within one standard deviation of the mean. If the PEI projection of a Trump win turns out to be correct, it should be considered strong suggestive evidence for the robustness of the model.

In practical terms, the point I was attempting to communicate to Dr. Martin-Joy was that the 2020 presidential election is taking place under highly unusual circumstances; in fact, some observers have referred to the political environment as “unprecedented.”

In that regard, it should be noted that the PEI heuristic model employs candidate personality traits, as publicly perceived, to predict which contender will resonate most favorably with independent and unaffiliated voters who base their voting choice primarily on the candidate’s personal qualities as publicly displayed rather than on party-political affiliation or allegiance. The implicit assumption is that the election occurs in a relatively typical, stable political environment in which the role of personality is not overwhelmed by strong situational forces (such as a global pandemic, a deep economic recession, high unemployment, social upheaval, or widespread civil unrest).

“Relatively typical” is in the eye of the beholder, but to invoke the normal curve metaphor, typicality can be defined as falling within one standard deviation of the mean (i.e., the general political environment in nearly 7 out of 10 election cycles). Of course, if social and economic circumstances surrounding the current presidential election truly are “unprecedented,” it could even be ventured that the 2020 presidential election is taking place in a political environment fully two standard deviations from the norm (i.e., more unusual than 9 out of 10 presidential election cycles).

Photo credits: Kevin Lamarque and Tom Brenner / Reuters

To conclude, I am not predicting the outcome of the presidential election, in which statistical models infinitely more sophisticated than my simple PEI heuristic give Joe Biden a 9-to-1 edge over Joe Biden. What I am asserting, however, is that — based solely on their respective personality profiles as assessed at the Unit for the Study of Personality in Politics — it’s advantage Donald Trump.

Graphic: FiveThirtyEight (Nov. 1, 2020)

Related reports

Projecting the Winner of the 2020 Presidential Election: The Personal Electability Index (July 8, 2019)

Projecting the Winner of the 2016 Presidential Election: The Personal Electability Index (Feb. 29, 2016)