Thursday, October 9, 2008

Misleading Graph: BBC on house prices

Map design is a subset of informational design and so techniques on how to mislead with statistics in graphs should interest those of us involved in maps.

Today I was surfing the BBC site and came across this:

It comes from an article discussing house prices and shows the percentage change of house prices with time according to Halifax and Nationwide. I added an inset (see arrow) which cuts a section of the y axis at the bottom of the graph and aligns it with the graph further up. As you can see, -10 lines up with 0 but the spacing of the intervals at the bottom of the graph is less that in the middle of the graph so -11, -12 and -13 don't line up with -1, -2 or -3. This has the effect of making the graph look less steep than it actually should be over the last 3 months.

So was this just some BBC junior being told to fit the graph into a smaller space and making a genuine mistake or is there a reason for this error?

Recently I've noticed the BBC trying to put a positive spin on the financial crisis gripping the world economy. I broadly welcome this, since part of the problem is people's trust, the media talking negatively certainly don't help the situation. In the text that goes alongside the graph at one point it says:

"It joined the Nationwide in claiming that the rate of decline [of house prices] was starting to stabilize when looking at three-month comparison figures."

So it could be someone trying to back up that statement by making the graph look like it agrees. However, the title of the article "Further decline in house prices" is pretty negative so I'm inclined not to believe its a mistake rather than an underhand attempt to mislead.

Whichever it is, this is not what I expect from the usually excellent news coverage from the BBC.

Later (20:40 9th Oct): With impressive speed its been corrected.


Ollie said...

Nice spot.
Have you told the BBC about this?

Rich Treves said...

Yes, they said they'd correct it and (checks site), they have.