Interview
Election 2020: what the data tell us
- Polling accuracy is projected to improve relative to 2016 as adjustments for Trump's support are made correctly, with undecided voters estimated at approximately 5 percent compared to over 15 percent previously, suggesting a robust lead for Joe Biden.
- Higher voter turnout is expected to benefit Democrats, particularly in diverse Sunbelt states such as Arizona, Texas, Georgia, North Carolina, and Florida, providing an upside if participation exceeds 2016 levels.
- Forecasting models prioritize presidential popularity and economic indicators rather than early voting data, which is difficult to analyze due to state-specific availability and discrepancies between registration and actual voting behavior, such as 30 percent of registered Republicans in Pennsylvania not voting for Trump.
- Potential model errors could arise from late voter preference shifts, pollsters contacting incorrect populations, or inflated voter likelihood responses from Trump's base if they are advised to vote in person, which may bias polls against Republicans.
- Structural issues with absentee ballots, including slower postal service delivery in certain areas and signature verification problems, could disproportionately affect younger, non-white, and lower-income voters, potentially causing a ballot rejection rate spike that might drop Biden's standing by one to two percentage points in competitive states.
- Election results are anticipated to be clarified by 9 or 10 p.m. Eastern time, coinciding with 90 to 95 percent of Florida's ballots being cast; a win for Biden in Florida is modeled to indicate a 95 percent probability of election victory barring unforeseen consequences.
- While absentee ballot counting may take longer due to the pandemic and reduced poll worker availability, the probability of the Electoral College outcome changing after significant counting is low, with final results expected to align with partial counts once approximately 80 percent of ballots are accounted for.
- Conspiracies regarding ballot discarding are characterized as rare and unlikely to impact vote totals, whereas widespread unrest or acceptance of violence is more likely if politicians question results or if the Electoral College diverges from the popular vote, though a projected eight to nine point margin for Biden is expected to minimize doubts about the result.
- An upcoming digital event is scheduled to address the economic impact of the election and field questions regarding the remainder of the campaign for Economist readers and viewers.