RESEARCH NOTE
The Family Dinner Lesson: What Happened When a Family Routine Was Tested Hard
When family dinner was tested strictly in 17,977 teenagers, about half or more of its apparent benefit turned out to be the family already being close, and nothing lasted into adulthood. Family play has not had a test like it.
Why dinner belongs on a page about play
Family dinner is the best-studied shared family routine. It is also the routine that has been tested with strict methods. Family play and family leisure have not had an equivalent test. So dinner is the closest guide we have to what a hard test of play might find.
For years, studies found that young people who ate with their families more often did better. A review of 14 studies found that frequent family meals went with less disordered eating, less substance use, less violent behavior and fewer depressive feelings (Canadian Family Physician, 2015). Seven of those studies followed young people over time. Seven were surveys taken once.
The review carried its own warning. Of the one-time surveys the authors wrote that "we can infer associations but not causality." They added that "there is the potential that other unmeasured confounders (eg, family structure) could explain the positive results."
The strict test
Musick and Meier took the question to a national US study of 17,977 students in grades 7 to 12 (Journal of Marriage and Family, 2012). They did three things the simple surveys had not.
- They took account of how close the family already was. For depressive symptoms, "including controls for family relationship quality in Model 3 reduced the bivariate association by a total of 61%". For substance use the reduction was 56%. For delinquency it was an estimated 49%. In plain words, about half or more of the link between dinner and a good outcome was the family already getting along.
- They compared each teenager with the same teenager at a later point. This asks whether a change in dinners goes with a change in the teenager. It sets aside anything about the family that stays the same. On this test, a change in family dinners still went with a change in depressive symptoms. It went with substance use only before other changes in the family were taken into account. It did not go with delinquency at all. The authors wrote: "We found no evidence of a causal effect of family dinners on delinquency, irrespective of how we modeled the process."
- They looked again at ages 18 to 26. Their finding: "family dinners associations did not persist into adulthood".
How much was left
What survived was small. The paper reports that "increasing family dinners from 2 to 7 per week was associated with a 0.07 decline in depressive symptoms". By the paper's account that is about 15% of a standard deviation, which is a common yardstick for how spread out people's scores are. In the stricter model it was under 10%.
Even that may be too generous. The authors say so themselves: "Our estimates of family dinner effects may be upwardly biased by reverse causation". Reverse causation means the arrow may run the other way. A teenager who is doing well may be the one who shows up for dinner.
Their summary is that the large effects reported in earlier one-time surveys "are due mostly to unobserved factors".
Two more checks
Dinner did little where the relationship was weak. Using the same data, Meier and Musick found that family dinners "have little benefit when parent–child relationships are weak but contribute to fewer depressive symptoms and less delinquency among adolescents when family relationships are strong." (Journal of Marriage and Family, 2014). For teenagers with the best relationship with their parents, dinner every night went with a depression score 0.13 points lower than no dinners, about 28% of a standard deviation. For teenagers with very poor relationships the two scores were 0.74 and 0.72, which is essentially no difference.
Younger children showed the same shrinking. A second team followed about 21,400 US children from kindergarten in 1998-99 to eighth grade (Child Development, 2012). In models "which controlled for unchanging aspects of children and their families," they found no significant relation between how often the family ate together and either school results or behavior. They add a caution of their own: "we are not able to be certain that we have eliminated all possible confounders and estimated true causal effects."
What this means for play
- Play research is where dinner research was before these tests. Across four leisure journals from 1990 to 2012, 12 of 181 articles on family leisure followed families over time, and one used an experiment (Journal of Leisure Research, 2015). Nearly everything else is a survey taken once.
- The dinner result is not a result about play. Nobody has run these tests on game nights or family outings. The link might shrink the same way. It might not. We do not know.
- The outcomes were different. The dinner studies measured how young people were doing. They did not measure whether the family stayed close or stayed together.
- The lesson is about method. Most of the raw link between a shared family routine and a good outcome was the family already being close. A one-time survey of family play cannot rule out the same thing. For more on that problem, see which way does the arrow point.
None of this says family dinner is pointless, or that play is. It says a strong-looking link from surveys can be mostly something else.
The full picture
This page covers one question. The whole verdict on the saying, with its history and every source, is in our research paper: A family that plays together stays together: is it true?