Skip to content
← Blog

Election graphics 2020: Land (still) doesn’t vote

A collage of various electoral maps of the United States, illustrating red and blue regions in different states and counties.

Every four years, a similar debate re-emerges around the iconic red and blue electoral map. For the last week we’ve watched as the states changed color from TV broadcasts, social alerts and website homepages. It’s a familiar picture: blue states on the coasts with a red center (and a few other blue states scattered about). But it paints a highly confusing picture of the electoral college because land is colored without regards to the actual voting power of each state (without even getting into the wisdom of the electoral college).

This is not a new concern or argument. Indeed, it’s been one the news graphics industry has recognized for at least twenty years. Yet, news organizations continue to default to this misleading (and ultimately, unhelpful) data visualization. For the election this year, we confronted the same challenge in our work with PBS NewsHour — and found an effective and simple system to solve it. But first, let’s look deeper at the problem and what’s been tried before.

The problem with the Electoral College map

This map paradigm, while familiar, masks the relative importance of each state. Montana has three electoral votes while New Jersey has fifteen. So is New Jersey 5x larger than Montana? No, just the opposite. Montana appears 19x larger than New Jersey. The gap is even worse comparing apples-to-apples: Montana has the same number of electoral votes as Washington, DC. However, on most maps it’s over 2500x times as large as DC (if DC appears at all).

A comparison graphic shows Montana, New Jersey, and Washington D.C. Montana has a large red map, area of 55,470px², and 3 electoral votes. New Jersey has a small blue map, area of 2,875px², and 15 electoral votes. Washington, D.C. has a tiny blue map, area of 21px², and 3 electoral votes.
What’s the relationship between a state’s landmass and electoral votes? Nothing.

The act of constantly showing these maps during an election (and aftermath) gives readers/viewers a distorted picture of who’s leading or won (and by how much).

Let’s look at 2004 (Bush re-election) versus 2008 (Obama election). 2004 was razor-thin (both in the electoral college and popular vote) while 2008 was a landslide. Yet, the area of the maps would make you think the opposite. The 2004 map’s red area uses about 75% of colored space (exact number is dependent on how Alaska is included). In 2008 Obama’s blue area occupies only about 42% of the colored space.

Two maps of the United States display electoral results with each state being either blue or red. The left map shows the 2004 election: Bush (Republican) defeats Kerry (Democrat) with 53.2% of electoral votes and ~75% red area. The right map shows the 2008 election: Obama (Democrat) defeats McCain (Republican) with 67.8% electoral votes and ~42% blue area.
Despite winning the electoral college (and popular vote) by a much wider margin, Obama’s 2008 blue area is about half the size of Bush’s 2004 red area.

The standard map, in other words, doesn’t successfully show who’s won or how much they’ve won by. What good does it do? It reminds you that Montana has more land than New Jersey. But land doesn’t vote.

Alternative approaches

So why do networks, websites and newspapers keep using this? It’s recognizable, but also the alternatives aren’t obvious (at first). One of the simplest solutions is to pair the map with a simple shared bar chart :

A map of the United States showing projected and apparent winners for each state in the 2020 presidential election. States are color-coded between Biden and Trump, and there is a bar above the map with the same colors and the number of electoral votes.
ABC News & FiveThirtyEight

Fivethirtyeight created an expanded “snake map” incarnation allowing them to label and sort the segments for each state.

A chart titled "The winding path to victory" depicts states' electoral vote margins from the 2020 election. Each state is on a winding line with one end being Trump and the other Biden. The closer the state is to one end, the more they are forecasted to being won by a big margin by the corresponding candidate.
FiveThirtyEight’s “Snake Map” visualization of the electoral college

While this is great for keeping score, it strips geography from the picture. How did New Jersey or Montana vote? I have to scan through every state name until I find them. It turns out a map is actually very helpful to locate a state or recognize regional patterns.

So if it’s a map you want, one of the popular approaches is a “cartogram.” These redisplay the states into blocks and arrange them in rough relationship to the geographic map.

A United States map styled as a grid with states occupying a number of squares corresponding to their number of electoral votes and each square being colored blue, red, or gray.
The New York Times provided a cartographic view for many of its electoral maps.

Of course, the problem with these visualizations is: “what the fuck is this?” The political die-hard might get it, but what about the other 95% of the population? A successful visualization shouldn’t require a narrator to explain why you’re looking at a map of the USA on LSD.

Six maps of the United States, showing election results in various ways. The first three use a standard geographical representation of the United States while the last 3 use a deformed representation where the size of each state or county is scaled proportionally to its number of electoral votes.
An infographic from The Washington Post on how a cartogram shows proportions of the Electoral College. It’s better in some ways, but is hard to parse.

Even if you solve for the electoral college, the problem intensifies when carrying this system through to county-by-county or other population-level results. The reason The Washington Post has to label major cities like NY, Chicago and LA on this map is because their location is so skewed. And if you have trouble finding Los Angeles, you can forget about finding Peoria.

A better Electoral College map

We approached the challenge this year when developing election graphics for PBS NewsHour’s 2020 coverage. With consideration for NewsHour’s non-partisan audience and ethos: how could we make a visualization that accurately conveyed the electoral college results and was instantly understandable to a PBS viewer? The solution we arrived at was familiar, scalable and simple. Dots.

Map showing 2020 U.S. Presidential Election Electoral College results, with Biden in blue and Trump in red. Each electoral vote is represented as a dot colored with blue or red inside of their state. Some states are small in size and have many dots while other are big in size and have only a few dots. Above the map, there is a bar chart representing number of votes between Biden and Trump.
PBS NewsHour’s electoral map uses colored dots against a background of states to illustrate electoral vote counts.

The dot system takes advantage of what both the other approaches do right. It uses the familiar geographic positions and areas from the “normal” map — and marries them with the proper weighting seen in a cartographic map. At a glance, you can not only see that New Jersey has more electoral votes than Montana; you can easily locate both states and follow as an election unfolds.

For 2020, we built a suite of election graphics tools for NewsHour to deploy this system to the web, social and TV.

A webpage displays various graphics related to general election results from "PBS NewsHour." It includes tables, maps, cards, and statistical data for different states and candidates.
The Election Graphics app dashboard we created for NewsHour. This lets them generate different graphics and easily bring them over to Twitter, Facebook, TV or WordPress.

Dots have another advantage: they provide a tool to represent both absolute values (like an electoral college vote), relative values (vote differentials) or comparative values (Trump versus Biden within certain counties/areas).

A map of Pennsylvania shows the 2020 US presidential election results. Biden has 50% and is marked as the winner, while Trump has 49%. Blue and red dots represent votes, with blue for Biden and red for Trump. Key cities and a vote scale are noted.
Pennsylvania’s county-by-county helps show that no city or county is “all red” or “all blue” — each is a different balance between the two.

They work great for maps and the same dots carry over into more abstract concepts like the House or Senate balance of power:

Two reporters are standing in front of a large screen displaying a graphic of the 2020 U.S. House elections. The screen shows 203 Democrat seats and 188 Republican seats, with 218 seats needed for control. Both presenters are discussing the results.
During the live broadcast, NewsHour anchors interacted with the same visualizations available on the PBS website and social media accounts.

The app allows NewsHour producers to browse from a visualization gallery and then select the particular race, style and other customizations they need. It builds upon a comprehensive system we started with them for the 2018 midterms and 2020 primaries. NewsHour now has a suite of tools to cover every major type of federal election with consistent styling across their distribution platforms — all of which build upon the comprehensive Design System and WordPress Editorial Website we originally designed and architected for them in 2017.

The dot was the simple mechanism that “broke it open” for us. It provides not only a mechanism for PBS NewsHour, but a methodology other news organizations can use or build upon to ensure the results of the 2024 election are better conveyed.