I received a message from Kester Takayama, which was a well written note in defence of the pie chart. Kester correctly notes that adding the third-place party to the first-place or second-place party doesn’t in itself create a government because they still may not have an overall majority. He also correctly pointed out that while the column chart made it easier to compare the “Conservative plus Liberal Democrat” and “Labour plus Liberal Democrat” values, it doesn’t greatly help to compare “Coalition versus all others”, which is the number that really matters.
The only part of Kester’s message that I don’t agree with is that a pie chart makes it easier to compare “Coalition verses all others”. His suggestion was that you put the coalition slices together and the non-coalition slices together and can therefore easily see if the coalition makes it past the half-way mark. This does work in some specific situations, but no more or less effectively than if you had two columns in a column chart similarly arranged. If you are going to arrange the values into groups, column charts give you an additional benefit, not only will one column be taller than the other – clearly showing the majority government if there is one, you are able to show the proportion of the column relating to each party and could also have a third column of parties not willing to align themselves with any coalition.
Here is a comparison of the pie chart verses the column chart for coalition option 1, Conservative and Liberal Democrats (C + LD).
And here is a comparison of the pie chart verses the column chart for coalition option 2, Labour and Liberal Democrats (L + LD).
Be honest about how long it takes to see whether the coalition is a majority government in either case using the pie chart instead of the column chart!
So the column chart can show the “Coalition verses all others” just as well as a pie chart, but in the event of a hung parliament, it will be even better.