It is a cliche that journalism classes tell reporters to ask the basic questions: Who? What? Where? and When? Sometimes they even pursue How? and Why? though usually these are asking too much. I remember being taught that list of questions. But apparently a requirement for being on the health beat is to have missed that day of class. Interestingly, epidemiology students are taught basically the same list of one-word questions.
The headline reads “Childless Couples Eat Healthier, Study Finds“.
Who and where? Reading the headlines and most of the story, you might think “all people everywhere”. But they do actually tell us the study was in Britain (of a sample of the population with unreported properties, but let us set aside the possibility that it is not even representative of the British). The American news sources that reported on it (which are the ones I looked at) seemed unconcerned by this fact.
Biological health information translates pretty well from one group to another. Economic health information (i.e., behavioral, social, etc. measures) often do not. Strangely, if a study of some biological health effect had been done only on wealthy white Americans, there probably would have been a disclaimer about the results possibly not translating to other populations. But the far more likely possibility that a study of behavioral choices might not translate beyond Britain (or, more likely, beyond a subset of the British population) receives no mention.
The British are about as much like Americans as anyone in the world, but I would hesitate to assume dietary behavior was similar. While English food is not as bad as it was only a few decades ago, their foodways remain rather different from Americans’. Also, the prevalence of vegetarian or near-vegetarian diets is higher there, the ethnic mix of the non-anglo population is very different, urbanization patterns are different, etc.
When? The data was from 2003. Eating habits are changing rapidly enough that the information might no longer be accurate, even for the people studied.
What? That is, how did they define healthier eating? It just means eating more vegetables. That was actually what they measured, not some more robust measure of healthful eating. These are not synonymous — e.g., many restaurant meals that include vegetables are higher in added fat and sugar and contain more total calories than is optimal (though this seems to be less true of Britain than America). It is certainly true that more vegetables is better, all else equal, though, so the conclusion about healthy eating is probably true. But why not just report what was observed? A more accurate headline would have been:
“Eight Years Ago, Childless British Couples Ate More Vegetables than those With Children”
Being accurate is just such a bother, though, with all those extra words. And everyone likes being healthier but vegetables is still a bit of a negative word.
Finally, there is the tricky question of how? Did having children reduce vegetable consumption, as the news articles about this would have us believe (harried parents, and such) or do those who are more likely to become parents just eat less vegetables (i.e., are people who are inclined to delay or avoid having children just more likely to be herbivorous?). We cannot tell from from the study design. But it makes such a good story to pretend this is about Happy Meals and over-scheduled lives, so the news reports went ahead and assumed a causal relationship.
In fairness to the researchers, it seems they were most interested in testing some statistical methods, since that is what they emphasized in their publication. This was pretty clearly a study that was meant to be used by other technical experts, not touted as having an everyday interpretation. On the other hand, they did mention the childless couples result in their abstract, and someone presumably touted the study to the media.
The lesson here is an easy one: Surely you would object to a headline that claimed that “childless couples are less physically fit” when the study was conducted in China in 2003 and defined “physically fit” as being more likely to buy a Nintendo game system. Obviously it would tell you little about who currently buys a Wii in in America, or even China, let alone about their fitness. So why would we believe basically the same leap when published in the health section?