Monday 2 June 2014

Junk food junk science

More quackery from the cesspool of public health research, as reported in the Observer...

McDonald's adverts for healthier Happy Meals fail to put children off their fries

Television adverts for McDonald's healthier Happy Meal deals fail to encourage children to choose food that is better for them and instead promote a general preference for fast food, academic research suggests.

This 'academic research' has not been published or submitted for publication. Indeed, it hasn't even been written up yet, but that doesn't bother the relentlessly statist Observer.

The gist of the article is that McDonald's should be banned from advertising all its products, including its healthier options, because the mere sight of the golden arches makes impressionable young minds hunger for french fries.

Dr Emma Boyland, of Liverpool University, who carried out the study, said the research into the choices of 59 children, six of whom were overweight and five obese, raised questions over whether the big fast food brands should be allowed to advertise to children.

Who, you may ask, is Emma Boyland? An impartial and rigorous scientist? A disinterested observer? A gallant seeker of truth who follows the evidence wherever it may lead?

Perhaps not.

Here are the details of Boyland's research according to the Observer...

In Boyland's study, half of the children watched 10 programmes with adverts for McDonald's Happy Meals, while the other half watched the same number of programmes with toy adverts. Of those children who had watched the McDonald's adverts, 84% subsequently said that they had a liking for fast food. This compared to 76% of those who watched the toy adverts.

In other words, a large majority of the kids in both groups like fast food. The difference between 84% and 76% is trivial. It is, literally and statistically, insignificant. Bear in mind that the total sample size is just 59 (if I was going to divide a sample into two I would have picked an even number, but then I'm not a public health scientist).

So we have two pitifully small comparison groups, one of 30 and one of 29. About one classroom each. In one of them, 22 kids say they like fast food. In the other, 24 say they like fast food. That is random variation. It is nowhere near statistically significant. It is a meaningless finding and the research is worthless.

This sort of utter bilge is sadly typical in policy-based pseudo-health research.


Ben said...

This begs the question: what is fast food? Fried potatoes, ground beef, a leaf of salad a piece of bread. My mother used to serve this kind of food quite often. What is wrong with that?

Ivan D said...

I expect the Observer to be statist and untrustworthy but universities re publicly funded and ought to be judged on the quality of their academic output. Increasingly, they are being used as vehicles for those with political axes to grind and no idea what academic rigour entails. The result is that fewer and fewer people believe anything that comes out of them, which is unfortunate for the genuinely talented but perhaps they should try speaking up against garbage such as this.

Chris said...

"59 children, six of whom were overweight and five obese"

Only 8.5% were obese? Where's this epidemic they keep screaming about?

Dr Evil said...

She had less than 30 kids watching the TV with Micky D ads. Her sample size is ridiculously low for any statistical analysis to be meaningful. She would need 300-500 in each group to get anywhere and even then the batch size is low and her screening is rather poor mixing normal weighted kids with fat and obese ones. If this was a drug study she would be laughed at by every journal editor and his statistician.

Jonathan Bagley said...

I agree with Dr Evil. Since it is unpublished, we don't know what exactly she did, but I don't see how her study could have been statistically significant with just 30 in each group. With her results it should have needed more like 200.

Anonymous said...

Smokers buy less food to fund habit
Irish Health
The survey was carried out in six European countries, including Ireland, where 1,000 people were questioned about their smoking spending habits.

So High taxes don’t deter smoking but instead cause hunger and blackmarkets for tobacco!