How we show election results is misleading

If you were watching the 2019 Alberta general election results, you probably saw a map like this at some point.

If you were a UCP supporter at the time, you were probably excited when you saw this map. If you voted NDP, your heart probably sank, especially after how well the party did in 2015.

Here’s the thing though: this map is misleading.

This map shows the electoral districts based on their geographic boundaries, which themselves are heavily influenced by population size. That means areas of low population density are going to be huge geographically. Likewise, those with high population density will look tiny.

That’s why the NDP wins in 2019 look like tiny dots in this map. They won only urban ridings (Calgary, Edmonton, and Lethbridge). And it’s why many of the UCP ridings look gigantic: they’re rural.

But this distorts perception of how much the UCP won the election. It makes it look like the party absolutely destroyed the NDP; it most certainly doesn’t look like the NDP won the second highest seats of any unseated government in Alberta’s history. Nor does it look like the UCP won less than 55% of the popular vote.

What if there was a better way to map the results?

Well, there is. And it’s called cartograms.

Cartograms visualize geographic data in a more accurate way, one that isn’t skewed by traditional map representations.

Here’s one I did of the 2019 elections results, with each hexagon representing a seat won in the election, organized to roughly represent the shape of Alberta.

Now, this one is a lot less overwhelming and more accurately shows the split between the two parties.

Except we could go even farther.

See, one thing the new map doesn’t show is seats that were easily won and ones that were barely won. So, let’s do that now.

In this map, I used the same colours as the previous cartogram if the seat was won with 60% or more of the popular vote. I lightened the colour slightly for those won with 50–59.9% of the vote, and I lightened it even more for those under 50%.

What we see here is that most of the ridings (more than half) weren’t won with solid leads. In fact, 22 of the 87 ridings. were won with less than 50% of the vote, both NDP and UCP.

This paints a much different picture of the 2019 election results than the first map above did. The 2019 election was actually much closer than the UCP would have us believe.

So, what about now?

Well, for the rest of the cartograms, I’m going to use data from 338Canada, an election projection website.

They updated their projections late last month and project the NDP to win 47.7 seats, with about 41.6% of the vote, plus or minus 11.5 seats. So, as many as 59.2 and as few as 36.2.

The UCP, on the other hand, could win as few as 27.7 and as many as 50.9.

Anything over 44 would be a majority, which 338Canada gives the NDP a 77% chance of winning. (As an aside, they give the Wildrose Independence party as 7% chance of winning even just one seat, but that’s based on the most recent polling data.)

So, with all that information, here is what the cartogram looks like for the current 2023 projected seat counts.

Much different. Those NDP seats are in Banff, the Calgary area, the Edmonton area, Lethbridge, Lesser Slave Lake, and the Vegreville area. The UCP are projected to win as many as 8 seats in Calgary, none in Edmonton, and pretty much everywhere else.

Now, what about barely won seats.

In this case, rather than the colours representing a percentage of the popular vote, they will represent 338Canada’s labels of “safe” or “likely”, “leaning”, and “toss up”. (You can read more about those labels here.)

This time around, there only 18 ridings that may not be landslides—10 for the NDP and 8 for the UCP—which is less than the squeakers we saw in 2019. If you want to see how your preferred party is doing in your riding, check out the list of all ridings on the 338Canada’s website.

Either way, it should make for an interesting race.