Saturday, 20 September 2014

Scottish Independence, Lord Ashcroft Polls, BBC, and a mockery of numbers and graphs


Edit 21/09: The Guardian is now taking Lord Ashcroft's awful poll seriously with an infographic here. Appalling.

I've seen this graph a couple of times already. It annoyed me, because the age brackets are the same width, when they shouldn't be.
Pictures below innit

Just like when the BBC "Scotland Decides" results page gave constituencies the same widths, even when they had massively different numbers of voters.

Why not show the green and pink as the actual number of votes.
Seeing as it's a referendum and not an election.
They I went to the data source, and I got really annoyed...
http://lordashcroftpolls.com/wp-content/uploads/2014/09/Scotland-Post-Referendum-poll-Full-tables-1409191.pdf

The data comes from Lord Ashcroft Polls. Lord Ashcroft is a Conservative Party lord and openly employed by and linked to the Tories, who makes giant profits offshore and has been in trouble with the courts for dodgy financial dealings.

They polled 2047 people via landline phones and the internet - not mobiles.

And here are the numbers they asked in each age range:
16-17: 14
18-24: 84
25-34: 263
35-44: 384
45-54: 415
55-64: 399
65+: 488

Not only are these very different sample sizes. The 16-24 sample sizes aren't even big enough to justify a conclusion.

So there you go. This graph is drawn badly and based on insufficient data collected poorly by a biased source.

What a bunch of bollocks.


Why not show an age bracket of 2 years as 20% of age brackets of 10 years.
Just saying. Just saying.



No comments:

Post a Comment