Reporting Statistics - when 25% isn't 25 out of 100.

Statistics often raise problems for journalists. Clearly some journalists don't get numbers, or find the concepts involved difficult to communicate.

Drive on BBC Radio Five Live today ran an interview with Jillian Satin on this story linking depression with reduced cancer survival rates. It's on the iPlayer here for the next week or so (about 2 hours 52 minutes in).

At the end of the interview I didn't feel much more informed than before. Some of that has to be down to Anita Anand who did the interview, but the contributor didn't do much to help either. For instance, when asked about the significance of the increased risk, she said "When we hear 25% we think of 25 out of 100, and that's not the right way to interpret it."

That made me stop and think for a bit, and put me off the rest of the piece. What I was hoping for was a clear explanation of how I should think of this 25% increase in probability. Unfortunately that wasn't forthcoming. Part of this is a problem with the format - Five Live tends to favour brevity, but on this occasion the item, and contributor, deserved more time.

As a listener I felt let down - the story is interesting and deserved a proper explanation, but Jillian Satin's answers didn't help explain her work and Anita Anand's questioning didn't give her the opportunity to tell the story.