One of the biggest Freedom of Information stories from recent weeks has involved the number of children drinking.
The headlines shouted about the number of under 11s being rushed to hospital for being drunk.
The story originates from the BBC and was covered by almost every news outlet in the country.
However there are some major warnings about using figures from Freedom of Information requests and whether they show the real picture.
The BBC’s More or Less show, with statistical genius and economist Tim Harford, ripped the story apart.
The show points out the original FOI request asked:
“How many under 18s had been to an A&E department for an alcohol related condition during the last five years.”
They picked out the phrase ‘alcohol related condition’ for analysis.
Harford raises the point: “It’s not the same as drunk even though that’s how most of the media have reported these 11 year olds.”
The ambiguity is not uncommon from a lot of requests used. One of the problems with mass FOI requests comes from the way different public authorities store and collate information.
Human interpretation, and different systems of recording the information, will almost always result in a variety of responses from the same question.
NHS Grampian, one of the highest trusts for child alcohol numbers, were questioned by the More or Less show as to how they recorded and searched for the the information.
‘They add some descriptive phrases in what’s called a free text field, or notes… They simply search that free text field for words such as drunk, been drinking or intoxicated.’
This seriously calls into question the headline figure of 300 under 11s who were drunk – or may not have been. For example children who went to the the department with an injury and were taken by a drunk parent would have shown up in the FOI answer if the parent’s state was mentioned in the notes.
If one member of staff always adds detailed notes to files then there is going to be more of a chance of it being swept up as part of a response to a FOI request whereas a a more succinct colleague may input less information.
All requests asking for searches of databases run the same risks.
Requests like these should be treated with caution by reporters an in an ideal world there would be consideration to whether the figures which have been provided will be probed further.
It is impossible to say whether the original requesters has considered whether their response would have cases where children may not have been drunk in the results – although it is unlikely they would have wanted it to.
In this case the overall vagueness of the question has prompted the poor data set which has been reported on countless times. Undoubtedly the figures will be used and quoted in the near future in other discussions of children drinking.
It highlights the importance of how the request is written and the need for proper research before a request is made.
The full podcast and de-bunking can be heard here.
Photo credit: Greencolander/Flickr under Creative Commons Licence.
You must be logged in to post a comment.