AppId is over the quota
AppId is over the quota
29 September 2011 Last updated at 12:32 GMT 
By Michael Blastland GO FIGURE - Seeing stats in a different way 
The media constantly warns of an obesity epidemic Pick up a paper and you'll read that people are destined to get fatter and fatter, but is this right, asks Michael Blastland.
Ready to play a data detective game? The case is obesity.
Your task is to decide which way the trend is going. Isn't that obvious? Here's the first piece of evidence.
Exhibit A: What people say about obesity in the UK.
We're in the midst of a fat explosion, basically, would be my summary of the public argument and media coverage, with recent projections that half of UK men will be obese by 2030 - and women not far behind.
That's the top end of expectations, rounded upwards, of course. The bottom end is that about 40% of UK men will be obese by 2030.
Exhibit B: The source of the reports.
The figures came from a series of articles in the Lancet. The authors point out that they are extrapolations of current data and the past might not be a good guide to the future.
Exhibit C: What is the currently available data?
Here in a chart is the latest data for England - not the whole UK, note.
The lines show the percentage of adult men and women who are either overweight or obese, according to the Health Survey for England, as measured by body mass index.
Where do you reckon these wobbly lines are going? Still soaring upwards or turning flat-ish? And if flat-ish, what does that mean? Is it a temporary break in a long-term rise, or a real and lasting change of direction?
Some commentators suggest that this is not a rising tide but flat water.
What's your judgement? Time for more evidence.
Exhibit D: The data for obesity only, excluding the merely overweight.
What do you reckon now? Not such an evident flattening in the last decade but still a slower rate of increase, perhaps. Or still too soon to tell? And then there's the figure for 2009, the latest available, dipping like that. What do we make of it, if anything? It is only one year, after all. Mass weight-loss during the recession?
Whatever the explanation, if even the best-case recent projections are to prove accurate we'll need about a 90% rise in obesity in 21 years, after something a little above flat in the preceding nine years.
Exhibit E: Predictions for obesity made in the past.
These have tended to be on the high side. The one reported here suggested the figure for men would reach 33% by 2010 and 28% for women. That now looks way too pessimistic. But does the past, including past projection errors, provide any guide?
Exhibit F: How the latest projections look compared with the latest data.
Is it plausible that we will go from here to there? I wouldn't rule it out. But I'd also say that it looks as if the 2030 projections assume the recent decade is an aberration. But what do you think?
All right, enough with the exhibits. But you might still say you want more evidence. Good. Like, what's happening among children?
Or you might want international comparisons, or methodological evidence about whether the measurement changed in any way, by altering who's included in the survey or adjusting for non-response, which these surveys used not to do.
And how about the most serious cases of obesity, the morbidly obese or the numbers operated on for obesity by the NHS? Both sharply up in the last few years.
Then of course there's the rest of the UK. Scotland continued to go up a bit throughout the 2000s for adults, not so for children.
But perhaps most interesting of all, since these are average rates of obesity, are there differences of class?
Then it's decision time.
So, what do you think, Sherlock? We're all capable of a bit of sleuthing here, once presented with some data.
So you might say there's a lot of hype in obesity stats. You might say the lines are still mostly on the way up, even if not so fast, and that's bad enough. We're much fatter than we used to be a generation ago, no doubt about that.
But current trends are less clear, despite the headlines, let alone what will happen in the longer term. Or maybe you don't think so.
Go Figure has trouble understanding why data like this can't simply be presented to the public so that we can listen to the experts, look at their figures, then make up our own minds.
The data for England is here.
The current figures for Wales are here, and similar to England's.
And here for Scotland.
Interpreting ups and downs is seldom easy. Even when numbers go clearly in one direction, they tend to go up and down along the way. And whenever there are significant ups and downs and flats and wobbles, the question is how long you need to look before you know which way things are really going.
And this is only the "what" question, as in "what's happening?" We haven't touched the "why?".
The next few years will be fascinating. But watch the evidence, not the headlines. Fat could be in the balance.
View the original article here
0 komentar:
Posting Komentar