Are there really three kinds of lies: “lies, damned lies, and statistics”? No.

These days, British official statistics are amongst the most reliable in the world.

That’s since reforms that separated the Office of National Statistics (ONS) from Ministers, and established the independent UK Statistics Authority to admonish politicians who distort statistics.

But statistics must be used with great care.

Take obesity.

There are more overweight people, especially children, around today than there were 20 years ago.

I thought I’d Google some statistics to look at trends in eating over recent decades.

Up came a table on Cancer Research UK’s website. It was labelled, “Household food consumption”.

To my astonishment it suggested that the consumption of both fats and sugar had gone down dramatically – fats by 30% between 1992 and 2012, and sugar by 53%.

I didn’t believe the table.

I sent it off to the expert statisticians in the House of Commons’ Library.

It wasn’t the table itself that was wrong, but how Cancer Research had labelled it.

They were using information on the purchase of bags of sugar, and packets of lard, margarine and butter to show how much of these foods had, on average, been eaten.

But the big difference in the intervening decades is that much more of the food we eat is processed, bought as ready meals, from takeaways, and restaurants.

If you look instead at the figures on what households eat – “nutrient intake” – there’s a different story; though it’s still a surprise. Intake of both fats and sugars is down, by 16%.

These are accurate figures. To me, they raise a crucial policy question, at a time when we are being bombarded with calls to ban or tax one allegedly harmful foodstuff after another.

If the average intake of sugar and fats is going down, as it is, the key factor in the increase in obesity is in lifestyle. We need therefore to target the couch potatoes, the parents who drive their children to school when they could easily walk, and not just the food stuffs themselves.