Between a deadlocked presidential race, dozens of senate and house races, and countless ballot measures, there’s plenty to follow in Tuesday’s election. Stanford social scientists will also be keeping an eye on some issues that might not be on your radar.
How the economy factors into election forecasting
If you ask Americans right now what issue matters most to them this election, they’ll likely tell you it’s the economy – a recent survey conducted by Pew found that eight in 10 registered voters considered the economy to be very important to their vote.
Brandice Canes-Wrone, a professor of political science and a senior fellow at the Hoover Institution, will be paying close attention to just how predictive a factor the economy is in swaying votes in the presidential election.
“There are long-standing models in political economy that predict presidential election outcomes based on a sparse set of factors in which economic indicators such as inflation and gross domestic product (GDP) play large roles,” Canes-Wrone said. “It is admittedly incredible to think that in an election such as this one – with such distinctive and high-profile individuals running for president – historical political economic patterns would be predictive. Yet amazingly, these models are very close to current polls. It will be interesting to see come election day whether this holds or instead, whether the models and polls are missing important elements that are novel to this election.”
Variability in voter list maintenance
America requires that voters be registered in order to cast a ballot, but how voter registration lists are updated can vary dramatically from state to state, says Stanford political scientist Justin Grimmer, the Morris M. Doyle Centennial Professor of Public Policy and a senior fellow at the Hoover Institution. But as he explains, how these voter lists are updated can vary significantly. “Some states enable people to register on election day, others don’t. Some states automatically register eligible citizens when they are at the DMV, other states don’t. And different states have different strategies for maintaining voter rolls, such as updating the addresses for people who move and removing individuals no longer eligible.”
The variability poses potential problems, post-election.
“This might be a big issue in post-election litigation in this election and potentially could be a major source of debate over the next few years,” Grimmer added.
State-level ballot measures with national implications
Meanwhile, there are many issues being voted on at the state level that have the potential to impact democratic processes on a national scale.
Stanford political scientist Jonathan Rodden will be paying special attention to Issue 1 in Ohio, which would create an independent redistricting commission. “This would lead to a substantial change in the redistricting process in a state where extreme partisan gerrymandering has been the norm,” Rodden explained.
“Republican elites in Ohio fought hard to prevent this from making it to the ballot, and they strongly oppose it,” Rodden added. “However, it appears to have a chance of winning, even though Donald Trump is expected to win a large statewide majority in Ohio. Moreover, Sherrod Brown [who is running for reelection] is slightly ahead in polls in the U.S. Senate race in Ohio. Thus, one of the broader areas of interest for me on election night will be split-ticket voting in various forms. In some states and in some races, the correlation between the presidential vote share and the vote shares of co-partisans (or aligned referendum items) might be low enough to lead to some surprising or counter-intuitive outcomes.”
Ranked choice voting as an antidote to polarization
One electoral reform Stanford political scientist Larry Diamond is paying attention to is ranked choice voting (RCV), a type of voting system where voters rank candidates in order of preference. If a candidate receives a majority of first-choice votes, they are the declared winner. But if no candidate receives a majority, the lowest-ranked candidates are eliminated in run-offs until one winner emerges. One of the arguments for RCV, also known as instant runoff voting, is that it offers voters more freedom in how they cast their vote, and as research has shown, it can also lead to more consensus building.
According to Diamond, he thinks it could be an antidote to political polarization.
“The logic is that this induces greater willingness to compromise, less intense partisanship, maybe a slight boost to moderation because you can’t narrowcast as easily, and you need to appeal to a broader share of the electorate,” said Diamond at the panel event, “America Votes 2024: Stanford Scholars on the Election’s Most Critical Questions,” held Oct. 16.
This election, RCV is on the ballot in jurisdictions in at least five states: Alaska, Nevada, Colorado, Idaho, and Oregon.
Exploiting fears about AI misinformation
Stanford legal scholar Nate Persily is tracking the potential for anxiety about AI’s role in the democratic process to be exploited in ways that erode voter trust.
While there is the risk that artificial intelligence could make it easier to flood the internet with fake news, Persily says, the worries surrounding these new technologies could be what ends up hurting democracy.
“I’m more concerned that the panic over AI is going to lead people to stop trusting true news,” Persily said. “We’re seeing politicians around the world mainly using the scare over AI to say, ‘What you think is true is actually fake,’ not ‘What is fake is actually true.’”
Persily took up this topic in an essay, Misunderstanding AI’s Democracy Problem.
Large language models as a tool for empowering voters
One reason for relatively low voter turnout in American elections is the high informational burden placed on voters by lengthy and complicated ballots, says Stanford sociologist Robb Willer. In San Francisco, for example, the voter guide is over 300 pages long.
Willer is testing a possible way to help voters navigate such overwhelming amounts of information: large language model (LLM) voting guide bots.
“While existing voter guides play an important role in reducing these costs, they are more commonly used by individuals already more politically engaged than average,” Willer said. “To better inform a wider subset of people, we are testing whether LLMs with access to human-vetted, non-partisan voter information can offer prospective voters more personalized, comprehensive, and accessible information than traditional voter guides. We investigate whether and how people use LLM-driven voter guides and how they use the information they get from the guides.”
Willer hopes that a rigorous analysis of AI voter guides can help facilitate their responsible implementation.
This post was originally published on here
