For what it's worth

Here is a graph that appeared on the BBC Ten O'clock news on Friday 10th September 2010:

Looks like a dramatic drop that is coming up. But check for a moment, and look at the vertical axis. Those bars don't start at zero, as one might expect. They start at about 80,000. Nice dramatic graphic - shame it distorts the truth that the BBC is charged with communicating. (The source is perhaps relevant: the Police Federation, who might be called an "interested party" when it comes to the number of Police Officers as it is their representative body. Of course this is the way they would want to show the numbers: the question is whether the BBC should go along with it).

 Let's take that data and show it with a proper zero:


Not quite so dramatic now, is it?

When such a graphic appears fleetingly on television, the viewer has no chance to reflect on what they have seen, or to check the presentation - the graph has gone and the story has moved on.

So it is vital that you are careful to make sure that the average (non-statistician) viewer takes away a fair understanding from the graphic that you show.

Black Mark BBC.



Please note: I have nothing against Police Officers. My interest is in honest and accurate presentation of information and my gripe is with the BBC, not the Police - it just happens that it is Police Officer numbers that were the subject of the example that caught my attention.


POSTSCRIPT 28 September 2010: Response from BBC

I write in response to your recent submission via the BBC Complaints website regarding the BBC News at Ten broadcast on 10 September 2001.

My name is David Larner and I am BBC Audience Services’ Complaints Co-ordinator with responsibility in this area therefore this matter has been escalated for my personal attention. I’m sorry for the delay replying.

You suggest that the BBC is guilty of “dishonest presentation of graphical data”, but what you call a "suppressed zero" is entirely commonplace therefore I’m struggling to see your point here as such graphs simply represent the relevant section of data.

Clearly there would never be any circumstance where the number of police officers would go down to zero or anywhere near it therefore it would be entirely irrelevant to make reference to zero on a graph.

Our data compared a decade of specific figures thus the graph simply encompassed the general range of those specific figures hence our graph and the axis values thereon were perfectly relevant and accurate.

The reporter also explained verbally that in 2000 the number of officers was around 120,000 therefore the middle value on our axis was 120,000. He went on to say that in 2010 the figure had risen to around 140,000 thus the upper value shown was 140,000. The final, lower, figure quoted was around 100,000 and so our lower value was 100,000 in order to give context to the figures given.

This did not “exaggerate the scale of change” as you suggest but merely presented the relevant data, and whilst the source was credited as the Police Federation it is the BBC which chooses how to present such information in our own programmes.

You ask “Surely the BBC must have some standards on the presentation of graphical information?” and indeed we do, in that we ensure that graphical information is accurate, appropriate and relevant as it was - demonstrably - in this case, even though it is your personal opinion that a zero should have been shown.

You refer to the graphic in question as one that “appears fleetingly on television” and extrapolate from this that “the viewer has no chance to reflect on what they have seen, or to check the presentation” because “the graph has gone and the story has moved on”.

In actual fact, the graph was on the screen for no less than twenty seconds, a very long time indeed in the context of a news report, and the context and the figures was clearly explained by the reporter verbally too. The values on the axis were perfectly clear and were displayed as static figures throughout.

Indeed, so clear was the data that you - as "the viewer" - have been able to both submit a complaint to the BBC and also create an entire blog site dedicated to the matter.

I can’t agree with your implied suggestion that viewers other than yourself were not able to compute the data being presented to them, and audience feedback suggests to us that your personal view on this matter is not shared by others.  

I hope this clarifies the BBC's position and in conclusion, please once again accept my apologies for the delayed response to your email.