The elections in the United States have historically attracted world-wide attention owing to the global implications of the American policies. These electoral contests are also keenly studied by poll experts who try to gauge the subtle undercurrents that have the potential to alter the outcome of these closely fought contests.
As far as the ongoing election is concerned, experts from Stanford University have noted six major fulcrums that they believe to be crucial in a return of the former President Donald Trump or Democrat candidate moving into the White House.
These determinants include the state of American economy, variability in voter lists, state-level ballot measures, ranked choice voting, misinformation by using artificial intelligence (AI) and the importance of large language models (LLMs) in combating this misinformation.
The economy remains a pivotal issue for American voters, with surveys indicating that the majority of registered voters perceive it as crucial to their decision-making.
Brandice Canes-Wrone, a political science professor at Stanford, is particularly interested in how economic indicators — such as inflation and gross domestic product — are influencing election predictions.
Long-standing political economic models, which rely heavily on such indicators, have closely mirrored current polling data, raising the question of whether these historical models will continue to hold true in a highly polarised election.
Canes-Wrone suggested that this election might either validate or challenge these predictive models, especially given the unique personalities of the candidates.
Voter registration processes vary significantly across the 50 American states, with some allowing same-day registration and others using automatic registration through agencies like the Department of Motor Vehicles.
Stanford’s Justin Grimmer highlighted this variability, noting that states have different methods for maintaining voter rolls, such as updating addresses or removing ineligible voters.
This inconsistency could lead to disputes post-election, particularly in states with contentious races, making voter list maintenance a key issue in post-election litigation.
At the state level, several ballot measures have the potential for broader national consequences.
One such measure is Ohio’s ‘Issue 1’, which would establish an independent redistricting commission to address partisan gerrymandering.
Stanford political scientist Jonathan Rodden is closely monitoring this issue and noted the strong opposition from Republican elites, despite growing support for the measure.
Rodden is also interested in the possibility of split-ticket voting, where voters might choose candidates from different parties, leading to unexpected results in some races.
Another area of focus is RCV, a system that allows voters to rank candidates by preference and eliminates candidates with the fewest votes until a majority winner is found.
RCV is believed to reduce polarisation by encouraging candidates to appeal to a broader electorate rather than relying on narrow partisan bases.
Stanford’s Larry Diamond sees RCV as a potential antidote to political division, fostering moderation and compromise.
This election sees RCV on the ballot in several states, including Alaska, Nevada, Colorado, Idaho, and Oregon, and its impact on voter behaviour and election outcomes will be closely watched.
The rise of AI has introduced concerns about its role in spreading misinformation and eroding trust in the electoral process.
Nate Persily, a Stanford legal scholar, warns that while AI could flood the internet with fake news, it is the public’s panic over AI’s potential that may prove more damaging.
Politicians are increasingly using the fear of AI to delegitimise truthful information, which could undermine public confidence in the democratic process.
Finally, Stanford sociologist Robb Willer is exploring how LLMs such as AI-driven voter guides, can help inform voters and reduce the informational burden of complicated ballots.
In cities like San Francisco, where voter guides can exceed 300 pages, LLMs could offer personalised, accessible, and non-partisan information, making it easier for voters to make informed decisions.
Willer’s research is testing whether AI can provide a more user-friendly alternative to traditional guides, aiming to improve voter turnout by making information more accessible to a wider range of people.