Sunday, March 15, 2009

Why can’t newspapers just tell us things?

A deeply unsatisfying story in the Independent on Sunday. It begins:

More people are out of work in a third of constituencies across Britain than when the Conservatives left power, new figures revealed yesterday.

As with many statistics-based media reports, my immediate reaction was: ‘Is this claim all that significant?’ There are 646 constituencies and you’d expect a fair amount of variation not just in unemployment rates but in changes to those rates over time.

The Sindy does not make it easy to find out. The report tells us absolutely nothing about where these “new figures” come from – other than that it’s an “analysis of unemployment in parliamentary seats”. Conducted by whom? Based on which data? Published where?

What’s more: does this analysis compare numbers of people unemployed or the rate as a percentage of working-age people? Given the growing population, the former comparison will produce results that are both less favourable (i.e. more headline-grabbing) and less meaningful. And is it using the claimant count or the Labour Force Survey measure of unemployment?

A little time on the web leads me to suspect that these mystery researchers will have got their data from Nomis, an Office for National Statistics research centre that deals with local labour markets. But at that point, what with it being a bloody lovely day outside, my detective instincts die. Sorry. I tried.

No comments: