The Media Bias Fact Check (MBFC) 2024 Electoral College Polling Map is developed through a systematic and data-driven approach, focusing on providing an accurate and current snapshot of the electoral landscape for the upcoming presidential election. The methodology employed for this map is outlined as follows:
Poll Selection
- Credible Polling Sources: Only polls from pollsters with a pollster rating of 2.0 to 3.0 as rated by FiveThirtyEight are used, ensuring reliable and accurate data. See List.
Time Frame for Polls
- Emphasis on Recent Polls: State polling data considered for averaging includes only the most recent 14 days, ensuring current political trends and public opinion are accurately reflected.
Polling Data Averaging
- Calculating State Polling Averages: An average is derived for each state based on selected recent polls, representing the current polling status in the state.
Incorporation of 2020 Election Results with Weighted Averaging
- Combining Polls and Historical Data: The polling average is combined with the 2020 presidential election results between Donald Trump and Joe Biden. This is done using a weighted average approach:
- 70% Current Polls: Reflecting more recent shifts in public opinion.
- 30% 2020 Election Results: Integrating established voting patterns.
- Comprehensive Political View: This combination offers a fuller picture of each state’s political inclination, balancing current trends with historical voting behaviors.
Classification of States
- Categorization Based on Averaged Data:
- Solid Trump/Harris: States with a lead greater than 5 points for either candidate.
- Lean Trump/Harris: States with a margin of less than 5 points.
Daily Updates
- Regular Data Updates: The map and its data are updated daily or whenever polling data changes to maintain relevance and accuracy, capturing the latest polling data and public opinion shifts.
Example Application
- Illustrative Calculation: For example, if current polls show Harris leading by +15.0 points in a state, but the 2020 election results had Biden winning by +5.0 points, the MBFC average would be calculated as follows:
- 70% of Current Poll Lead: 70% × +15.0 points = +10.5 points
- 30% of 2020 Election Result: 30% × +5.0 points = +1.5 points
- Combined MBFC Average: +10.5 points + +1.5 points = +12.0 points lead for Harris
This methodology provides a nuanced approach to mapping the electoral landscape, incorporating both the immediate sentiment of the electorate and the inertia of past voting trends, giving a more rounded perspective for electoral projections.
Win Probability (recalculated each time the MBFC average is adjusted)
Data Preparation
- Load the polling data, which includes information like the number of Electoral College votes per state and the MBFC’s current average lead percentage for the leading candidate in each state.
Adjust Polling Leads with Margin of Error
- For each state, adjust the polling lead randomly within a ±5.1% margin of error during each simulation. This adjustment accounts for the inherent uncertainties in polling data.
Recalculate Win Probabilities
- Determine the win probability for each state in every simulation based on the adjusted lead:
- Lead > 5.1%: 100% certainty.
- 5.1% ≤ Lead Simulation run.
Obtaining National Poll Average
Prior to running simulations, we obtain the latest national poll average over the last 10 days. This figure represents the overall lead of one candidate over the other on a national level.
Adjusting State Polling Data:
- For States Where the Leading Candidate Matches the National Poll Leader: We adjust the state’s “MBFC’s Current Average % Lead” by adding a fraction ( 10%) of the national poll lead to the leading candidate’s advantage in that state.
- For States Where the Leading Candidate is Opposite to the National Poll Leader: We subtract the same fraction from the leading candidate’s advantage in the state’s “MBFC’s Current Average % Lead”.
- Additionally, we either add or subtract 8% of the differential in Candidate favorability ratings. For example, if Trump has a favorability rating of 42% and Harris 40%, the differential would be Trump +2 x 0.08 = 0.16 added to MBFC’s Current Average % Lead for Trump and subtracting 0.16 from Harris. Finally, we add or subtract 20% from the differential in congressional polling averages over the last 10 days.
- Determining the overall odds of winning the election: To calculate the overall odds of winning, we factor in the Candidates’ money raised differential and then add or subtract 1.2% from the difference. For example, if Harris had raised 100 million and Trump had raised 50 million, it would look like this 100 – 50 = 50 million difference in favor of Harris. Harris would gain 50 x 1.2% = +0.6 added to her polling averages. Trump would have -0.6 subtracted from his polling averages. Finally, the overall odds subtracts 1.5 full points from the leading candidate in each state. This disaster formula considers something bad happening to the leading candidate before the election (controversy, investigation (Clinton 2016), bad press, economic downturn, war, natural disaster response, polling over calculation etc.)
This adjustment reflects the influence of money, national momentum, unforeseen events or shifts in voter preference that might not be fully captured in individual state polls.
Monte Carlo Simulation
- Perform a large number of simulations (5,000 in this case). In each simulation:
- Adjust the lead for each state within the margin of error.
- Recalculate the win probability for each state.
- Use a random number within the margins of error -5.1 to +5.1 to determine if the candidate wins each state based on the recalculated probability.
- Sum the electoral votes for the states won by the candidate.
Calculate Overall Win Probability
- Determine the proportion of simulations in which the candidate wins the majority of the Electoral College votes (270 or more).
- Express this proportion as the overall probability of the candidate winning the election.
- Calculate the average Electoral votes won over 5000 simulations.
*The simulation is built and run using Python and updated daily with Excel.
Through this methodology, MBFC’s 2024 Electoral College Polling Map aims to provide an insightful and dynamic tool for understanding the electoral prospects in the upcoming presidential election.
Last Updated on October 15, 2024 by Media Bias Fact Check