I also live in an inner-city area, and there are middle class obese children here (and I've even seen some at swimming lessons), though the majority of the obese children probably are as frogs describes. So there are evidently (we've found out from a sample of just 3) regional and social variations. It's not a simple matter. Improving school meals would seem pretty obvious, but I understand that has backfired as many children who would previously have had school meals are now eating even more unhealthy packed lunches.
To deal with a problem of this complexity I believe an evidence-based approach is needed. It is necessary to know how much of a problem there is, where it is and whether particular groups are worse affected than others, and to track the situation over time. Majorca, you ask where the figures come from. I see this weighing as an attempt to obtain more accurate and consistent figures rather than the hotch-potch of partial information we have now.
desperateScouseWife, thanks for raising the point that boys have problems too, but I'd take issue with your dismissal of the collection of statistics as a pointless exercise. It was through the analysis of statistics that the link between smoking and lung cancer was established. It might seem bleedin' obvious now, but it wasn't at the time.
frogs, I wouldn't appreciate being weighed by my place of work (if I had one), but largely because of fears over the use to which my employer might put the information. If, say, there were a nationwide study of women over 40 carried out at workplaces for convenience and I felt my participation would be for the common good (and my employer didn't see the results) I'd not be remotely bothered.