Quick exercise to calculate the Marginal Votes for Electoral College Victory for US presidential elections. All work done in the Jupyter Notebook.
I was inspired to create this notebook by some random headline statistic thrown out after the 2016 US election like "The Election Came Down to 77,744 Votes in Pennsylvania, Wisconsin, and Michigan". Due to the Electoral College (EC) method of electing the US president, there can be a large discrepancy between Electoral College percentage win and total voting percentage win (or loss, in some elections). So in the 2016 election, Trump won the EC by only 77,744 votes across three states. If you define this metric as the main one to determine how "close" a US presidential election actually is, I was curious what this metric was for all other modern US elections. Unfortunately I couldn't find any article that detailed, calculated, or even mentioned any other election in the same way they did for 2016.
So I thought I would calculate this metric for other US elections. I am going to officially call this metric the "Marginal Votes for Electoral College Victory" or MVECV. This is described as the minimum number of popular votes needed to add to the Electoral College loser across all states such that the Electoral College loser becomes the Electoral College winner.
Here are the figures I made that summarizes my calculations. If you want to know more information about where the data came from, how the calculations were done, why I am displaying data only from 1976 and onwards, or why I have a Figure 1-alt, you can look through the Notebook for that information. Additionally, the Discussion section in the Notebook contains more information about what you can learn from these figures.
It turns out that US presidential elections are usually closer than the Electoral College victory margin makes it seem. Go to vote.gov to register to vote.