Wow! Repeat After Me: Correlation N.E. Causation

Checkthis:

The relationship between frequency of family dinner and adolescent problem behaviors after adjusting for other family characteristics

Bisakha Sen, Journal of Adolescence, forthcoming

Objective: To examine the association between frequency of family dinners (FFD) and selected problem behaviors for adolescents after adjusting for family connectedness, parental awareness, other family activities, and other potentially confounding factors.

Methods: Data are drawn from the National Longitudinal Survey of Youth, 1997. The primary variable of interest is self-reported FFD in a typical week. Problem behaviors studied are substance-use, physical violence, property-destruction, stealing, running away from home, andgang membership. Multivariate logistic models are estimated for each behaviors. Linear regression models are estimated for behavior-frequency for the sub-samples engaging in them. Analysis is done separately by gender.

Results: FFD is negatively associated with substance-use and running away for females; drinking, physical violence, property-destruction, stealing and running away for males.

Conclusion: Family meals are negatively associated to certain problem behaviors for adolescents even after controlling rigorously for potentially confounding factors. Thus, programs that promote family meals are beneficial.

I am willing to believe that FFDs stand in for a host of unobservable features of family life, and so it is a good predictive variable. No worries; you can use it. But it is absurd to think that it is the actual family dinner that CAUSES the salubrious effects.

In fact, I'm pretty sure that being required to have LOTs more family sit downs with Jim and the little woman at the Angus ancestral home would have produced MORE of the "running away behavior" the study finds is here reduced. At a minimum, I know I was pretty big on "running away" from the FFD.