Post-Election Campaign Narrative

In every election, expert predictions and district forecasts do not always align with the true results that we see. There could be numerous reasons for these discrepancies and in examining these types of districts, we can learn a lot about the campaigns run. I will be looking at Maine’s 2nd Congressional District in hopes of understanding some important influences on this election and speculate where the forecasts may have gotten it wrong.

Post-Election Reflection

On Tuesday, November 8th, eligible Americans voted in the midterm elections for candidates who will make up the 118th Congress. Prior to the election, we had created forecasting models in an attempt to determine who would win the House of Representatives, at the district-level, national-level, or both. In this post I will reflect on the predictions made, my model’s performance, and comment on the election overall.

Final Prediction

Over the past nine weeks, we have explored a multitude of variables that may impact election outcomes in an attempt to forecast the 2022 midterm election. We have learned about economic forces, polling, expert predictions, incumbency, advertising, campaigning, and more to find if they hold any predictive power, and how together in a model they may foreshadow what is to come in the results next week. In this blog, I will create a final prediction model for the House of Representatives and share its results, with the intention of reflecting on it after the election.

Shocks

This week we learned about how the things we cannot predict happening can end up affecting election outcomes. Shocks capture everything from natural disasters and Supreme Court decisions to e-mail scandals and sports outcomes. While no one expects these events to happen, voters sometimes blame candidate losses on these shocks. We will look to explore the true effect of these political shocks, or lack thereof, on elections.

The Ground Game

This week we learned about how election turnout can affect the outcome of races, and how on-the-ground campaigning efforts aim to mobilize and persuade voters. We can look at this data on the district-level to uncover patterns, and compare it to last week’s data on ad spending to see how they interact.

The Air War

This week we learned about the effect of advertising on election outcomes at the district level. There is some conflicting literature about the effects of advertising on informing, mobilizing, and persuading voters. In this blog, I will look at trends between advertising data and vote shares in previous elections, then add to my model predicting the 2022 House election.

Incumbency and Expert Predictions

This week we looked at the effect of incumbency on models’ predictions. We compared and contrasted various pollsters and models to see which generated the most accurate predictions, which I will continue to look at through visualizations in this blog.

Polling

This week, we studied various polling methods and tried to explain the variance among pollsters. Why are polls inaccurate? In this blog, I will add polling data to the economic data we looked at last week in an attempt to improve my model and generate better predictions for the upcoming midterm election.

Economic Forces

This week, we looked at using various economic variables as predictors for election outcomes. In this blog, I aim to predict Republican seat share and incumbent party vote share using variables such as GDP growth percentage, CPI percent change, disposable income percent change, and the national unemployment rate.

Using Voteshare Margin and Seat Share to Analyze Gerrymandering

From now until November 3, I will be updating this weekly blog series with my 2022 US midterm election prediction model. In this first blog, I’ll be comparing the voteshare margin by party to the seat share by party in each state to hopefully explore the effects of gerrymandering.