Something's fishy. And it's not what's for dinner.

(Warning: disorganized draft version...)
Reading Judith Rich Harris's The Nurture Assumption has left me highly skeptical of claims that parents have control over their teenagers' behavior. So, it's not a surprise that after reading the recent New York Times article "The Family Meal Is What Counts, TV On or Off", about how family meals are the amazing panacea for healthy eating habits and health-risk-related teen behaviors like drug use, smoking, and eating disorders, I was left thinking, "Really?"

The Times article is a report of the latest published study from Project EAT at the University of Minnesota, which continues to find correlations between eating family meals and pretty much anything happy and healthy you can think of. The study itself isn't available online unless you cough up $15, and my local library won't have it on microfiche until about 2040, so you have to draw all your conclusions from the Times report and the article abstract.

Based on survey data of students in Minnesota in the late 1990s, students eating healthier diets also ate meals as a family more frequently. It's not clear from the story or the abstract whether this means their overall diet or the foods they ate during family meals were healthier, but presumably the former. It's also not clear whether they took into account such obvious confounding variables as the ability of the kids' families to afford healthier food (which is nearly always more expensive than unhealthy food on a price per calorie basis) or the ability of the families to afford having an adult at home to prepare and participate in a family meal (not easy if you are working multiple jobs or second shift or have only one adult in your household in the first place.) Nonetheless, the Times story concludes that "the best strategy for improving a child's diet is simply putting food on the table and sitting down together to eat it."

The Times story claims that the implied causal relationship between frequent family meals and a healthier teen diet is supported by the fact that in a separate study 'family connectedness', a proxy for the psychological health of a family, was shown to be "less important than whether they regularly dined together" for the incidence of these teen health behaviors. (This article on connectedness did take into account socioeconomic status and race/ethnicity, which leads me to believe that the Journal of Nutrition Education and Behavior article did too; but there's no way to tell without reading the latter.) How the discovery that one independent variable A is more strongly associated with a set of dependent variables than another B can lead you to conclude that the A must then be a cause of the patterns seen in the dependent variables is beyond me.

Once you look at this second study, however, things rapidly become even murkier. The authors' definition of the variable 'family connectedness' is actually an averaging of the answers to the questions, "How much do you feel your mother/father cares about you?" and "Do you feel you can talk to your mother/father about your problems?" They say, as stated in the Times article, that the correlations they've found between family meals and adolescent health behaviors hold even if the effect of family connectedness on their models was controlled for.

The author of the Times article repeatedly stresses the importance of the connectedness factor in showing that its truly family meals that are the beneficial phenomenon in her blog. In fact, the connectedness questions actually look like they in part measuring how open your communication is with your parents, not how functional your home life is. I suspect that very hierarchical, authoritarian parenting styles are not going to give you high connectedness, but may very well still correlate with frequent family meals and even happy, healthy teens. A teeny piece of evidence towards this is that in the abstract of another Project EAT study, "Family meal patterns: Associations with sociodemographic characteristics and improved dietary intake among adolescents," Asian-American ethnicity correlated with higher frequency of family meals. The Asian-American student population in Minnesota in the late 90s was likely to be children of first or second generation immigrants--households where non-mainstream American family practices are likely to still be strong before the immigrants are fully assimilated into the local culture. These are probably the exact kids who are not happily sharing all the secrets of their teen years with their parents but who actually participate in a lot of family meals. It would be nice to see the two types of connectedness questions separated out---sharing secrets with your parents vs. feeling like your parents care about you. The latter seems like it would be a far better way to estimate how functional your family life is. In any case, the authors of the study also admit that "[r]egular family meals may also be a proxy in this study for other elements of family connectedness that are not captured in the measures used here."

More importantly for my point here, although their primary conclusion is that family meals may enhance the health and well-being of adolescents and that public education efforts should stress the benefits of family meals, they also say explicitly that "the study's cross-sectional design meant the results do not imply a causal relationship between eating family meals together adolescent health behaviors." (Emphasis all mine!)

So, why should we think that family meals are this magical ingredient that keeps our kids healthy and happy? I admit that I am a strong proponent in my own household for family meals. My husband could actually eat three meals at work a day, and has on occasion suggested that this would be easier for him and us as a family. This inevitably leads to me jumping up and down and throwing pronouncements like "Hell no, you're not going to start eating dinner at work!" around. Eating dinner together provides a reason for me to cook an interesting nutritious meal on a daily basis. It gives us a chance to all hang out without any major distractions for at least one short stretch. And, I just plain like family meals. But is it the cure-all for the social and public health problems of the day? Will our daughter grow up with a protective vaccine against the trials and tribulations of adolescence because of it?....um, I doubt it.

Why am I such a doubter? This is where the Harris theory comes in. Her idea in short is that---based on many, many behavioral genetic studies of adult personality, behavior, and intelligence---there is little to no effect on these same factors of growing up together in the same household. Nearly every study that purports to find some benefits of a particular parenting style or particular feature of home life does not control for the fact that parents and children share their genes. Without taking heredity into account, you just cannot sort out whether the effects you are seeing are caused by genes or environment. And, on top of that, based on how dissimilar identical twins raised together are---just as dissimilar as twins raised apart---growing up in the same household does not appear to have more than a weak effect on behaviors like substance abuse/use, depression, academic performance, and self-esteem.

So, in light of all of Harris' evidence, the so-called "surprising power of family meals" appears to be just another case of causation-correlation confusion in parenting practices. Healthy, happy parents are more likely to have healthy, happy kids because those kids share their parents' genes. Happy families do things like eat family meals together, care for each other, spend time together, and eat relatively healthy diets. The kids get good grades, don't run around raising hell with smoking, drinking, and drug-taking kids, and don't think about suicide. Living in a wealthier neighborhood with more highly educated, wealthier parents makes all this stuff easier and more likely, but not having a lot of money or education does not make it impossible.

The number of studies telling us that eating family meals together 'leads' to happy, healthy teens, and thus adults, may continue to increase, and strong correlation can hint at causation.
But without studies that rule out other possible causes, the case for causation here remains weak. (Frequent family meals and happy, healthy teens may very well strongly correlate with how often a family needs to get their car repaired, the number of letters following the parents' names, and their use for high SPF sunscreen, but no one would mistake any of those associations for the causes of healthy, happy teens!) It remains perfectly plausible that there is some independent cause or set of causes of both frequent family meal eating and high-levels of adolescent health and happiness!

So, if family meals aren't a cause, does it matter? Yes! If family meals are not an actual cause of better health and more happiness, then devoting scarce public health resources to promoting them may take resources away from finding and promoting actual sources of health and happiness. And for all we know increasing the frequency of family meals in unhappy families may make things worse for those families!

Please, someone, somewhere do a study that compares happy, functional families that eat together with happy, functional families that don't eat together to see if there's anything important to conclude. A nice test demographic that comes to mind is day students at English boarding schools. Some schools may require day students to have dinner at school, and others might not. Is there really an important difference? (My husband and his siblings spent their teen years at school until late in the evening, rarely if ever eating a family meal because of required evening study sessions and school dinner hour. They are all happy, healthy, successful members of society today. Not eating family meals together may have had effects on them, but not ones that resulted in them or all their schoolmates being unusually drug-addled, eating-disorder ridden, suicidal basketcases.)