Several years ago Jamie Oliver set out to improve school food for a group of British children. In part he was motivated by wanting to impart a love of good, healthy food in children by sharing his own joy in food, and in part he believed that healthier school meals would lead to less obesity and better academic performance. As Tim Harford noted in his Undercover Economist column in November:
Oliver’s mission to persuade schools to serve healthier lunches – and get children to eat them, and stubborn mothers not to stuff chips through the school railings – became a national phenomenon in 2005. Tony Blair and David Cameron fell over themselves to jump on the Naked Chef’s bandwagon, and soon everyone in the country had an opinion on the campaign.
Harford then discusses a working paper by Michèle Belot and Jonathan James that uses Jamie Oliver’s school lunch work as a field experiment; from Belot’s web site:
Children’s diet is a major source of preoccupation in many developed countries. The concerns have mainly been focused on the implications for obesity and health outcomes. However, the effects of children’s poor diet may extend beyond health; food is an obvious input in the “learning production function” and deficiencies in diet may result in important deficiencies in nutrients playing an essential role in cognitive development. This study exploits a unique experiment in the UK, the “Feed Me Better Campaign” where the meals served in the 81 schools of one area (Greenwich) were changed drastically by the British Chef Jamie Oliver. Because the campaign was literally designed as a large-scale experiment, it offers a unique opportunity to assess the causal effects of healthier food on educational outcomes. We find that educational outcomes did improve in Englsih [sic] and Science, although we cannot rule out small effects. We also find that the campaign reduced absenteeism by 15% .
A well-designed field experiment in economic policy is hard to achieve, particularly because to get a representative distribution among treatments you have to randomize who participates in what treatment, which is politically difficult (I can tell you stories about this in electricity, but it will require a cocktail). This experiment does not randomize, but it does provide a large-scale test with comparative demographics that minimize the selection bias problem. As Harford notes,
The chef had convinced Greenwich’s council and schools to change menus to fit his scheme; he mobilised resources, provided equipment and trained dinner ladies. Other London boroughs with similar demographics received none of these advantages – and indeed, because the programme wasn’t broadcast until after the project was well under way, probably knew little about it. The result was a credible pilot project. It wasn’t quite up to the gold standard of a randomised trial, but it wasn’t far off. …
… Surely what counts is that a new idea was tried out on a respectable scale, and now we have a chance to figure out whether it worked. What astonishes me is that it took a television company and a celebrity chef to carry out a proper policy experiment.
Oliver’s work has led to his receiving a TED Prize for 2010, drawing attention to his work to improve diet, food knowledge and understanding, and cooking for children, focusing on the UK and US. His TED address from a couple of weeks ago is well worth a listen; he does go a bit histrionic for my taste in some parts of it, but his passion for incorporating knowledge about food into education is obvious. And, he knows how to design a policy experiment.