bigrus

Which election forecasting models can you trust?

Harry November 3, 2024

WASHINGTON – As voters look to polls and political analysts for insight into who might be elected Win the presidency on TuesdayThe fight between two of the country’s leading election forecasters Allan Lichtman And Nate SilverWill be put to the test soon.

Lichtman, American University professor correct guess Nine out of 10 predict victory in recent presidential election Vice President Kamala Harris.

Silver, the statistician and pollster who founded FiveThirtyEight, recently wrote: New York Times He says the race is a virtual draw but his “gut instincts” tell him former President Donald Trump will probably prevail.

Lichtman and Silver disagreed on methods

you have a couple got into a fight on social media about the validity of their methods.

In September, Silver questioned whether Lichtman had accurately evaluated the “13 keys” he used to project election results and suggested the professor’s system actually favored Trump. Lichtman responded by saying Silver, whose background is in economics, “is not a historian or a political scientist” and has been wrong in the past.

“At least 7 of the keys, maybe 8, are clearly in Trump’s favor. Sorry bro, but that’s what the keys say. Unless you agree they’re completely arbitrary?” Gümüş shared on social media.

So whose guess is more accurate? So how do they reach these conclusions in the first place?

Forecasting approaches

Lichtman devised the measurements he used for his election forecast more than three decades ago with the help of an earthquake expert and mathematician from Moscow named Vladimir Keilis-Borok.

The system, called “13 Keys to the White House”, uses, as you might guess, the following: thirteen true or false statements It relies on historical analysis of the country’s situation, parties and candidates to determine who will win.

This survey includes questions such as whether there is a third opponent, “whether the White House party is avoiding a primary” and whether any of the candidates are charismatic.

The method does not take into account how campaign messages or key events such as debates affect voter sentiment. Lichtman often makes his assessment several months before an election and does not change his assessment unless major foreign policy events occur.

When six or more of the statements are correct, the challenging side is expected to win. If five or fewer are incorrect, the incumbent party is expected to win. Lichtman said at least eight of the keys are in Harris’ favor in 2024.

But Silver uses an entirely different strategy and set of data points to examine the state of the election.

It creates probabilistic statistical models based on national and state surveys, economic data points, likely voter turnout and other factors. Model It also adjusts for differences in the surveys it collects, giving more weight to pollsters it deems more reliable.

Forecast records

Lichtman has correctly predicted the outcome of nine of the 10 most recent presidential elections dating back to 1984. Did he guess wrong? The 2000 presidential race in which George W. Bush defeated Al Gore.

Silver became nationally known in 2008 when his statistical model correctly predicted the outcome of the presidential election in 49 of 50 states. His model has since predicted the outcome of the presidential race in 2012 and 2020. During the 2016 election, Silver’s model suggested a possible victory for Hillary Clinton but gave Trump a roughly 30% chance of winning; This is much higher than most other forecasters.

Which model is better?

It depends on who you ask.

Thomas Miller, director of Northwestern University’s data science program, argued that both Silver’s and Lichtman’s strategies are “wrong in different ways.” Miller created something election prediction system Combining 60 years of historical analysis and betting market data, Predict It.

He argued that Lichtman’s model fails to explain how campaign messages and key events changed public sentiment in the final months of the election.

“To Lichtman, nothing campaigns do really matters. “The message doesn’t matter, the positioning doesn’t matter… because everything is kind of predetermined by history,” Miller said. He also questioned whether Lichtman’s economic metrics, which look at U.S. gross domestic product, are valid, target accurately perceptions about the economy.

For example, inflation is a major issue for many voters this year. The U.S. economy is performing relatively well, but voters aren’t quite feeling it.

However, Lichtman disputed these claims and said his economic analysis was objective and had roots dating back to 1860. Each key is tightly defined based on this analysis, Lichtman said. He argued that the lack of campaign events in their keys was one of the reasons they were so successful.

“What some people say is the weakness of the keys… the strength of the keys because it looks at the fundamentals and not the ephemera of the campaign,” Lichtman said. he said. He said the structural model reflects how American presidential elections actually work.

Miller also saw flaws in Silver’s approach; that is, it relied too heavily on survey data, which was variable and fallible. If the polls prove wrong, Silver’s predictions will be wrong, too.

Weighting polls by which groups of people are more likely to vote can also be complicated, Lichtman suggested. Polls can and do underestimate the number of Democrats and Republicans participating in the election, for example.

Election analyst David Wasserman Cook Political ReportHe said that despite the variability, he found Silver’s approach “more methodologically rigorous.”

“Lichtman is comically overconfident and doesn’t acknowledge the subjectivities in his method, but you’ll learn a lot about presidential elections by reading his work, and he at least puts himself out there,” Silver said in late September. making testable predictions.”

Wasserman said he believes Silver’s approach is better at “communicating the state of the election to the public,” in part because it “recognizes that there is uncertainty inherent in elections and future events.”

“I believe campaigns matter and a candidate’s choices influence voters’ opinions,” he said. “I give more weight to Silver’s approach because it can take these factors into account.”

But in essence the models are completely different.

While Lichtman’s model looks at established patterns of past elections to predict future presidential votes, Silver’s model provides insight into how American voters’ views change over weeks and months.