I blogged earlier today about how the Mirror bigged up a poll showing Tory support unchanged (within the margin of error) into a story of how their support was plunging.
But looking again at the story, I realise just how badly wrong their graph is.
The two key pieces of information about Labour’s rating in the poll are that:
(a) It was 30%
(b) It was 32% in the previous poll
Now look at the graph:
See what’s happened? What should be a downward Labour line has become a flat line and the end which should be at 30% is actually put at about 31%. Both the trend and the end point are wrong – and both in ways that flatter Labour.
Not good at all.
(You may notice that the Tory point also looks like it may be a little on the low side too, but having deployed ruler and calculator the overall impact on the graph is to make the Tory-Labour gap 25% smaller than it should be. Plus the Labour line is flat rather than down.)
Keep up with the latest news and analysis
about the Liberal Democrats with my
free monthly email newsletter.
I scour hundreds of blogs and dozens of media outlets for the best news and analysis - so you don't have to. It's completely free and you can leave the list at any time. So why not give it a try today?
You might also be interested in...



Goodness me, iffy graphs/bar charts showing a distortion of polling preferences.
It’ll never catch on
Rob