RESEARCH NOTE
Do Close Families Play More, or Does Play Make Families Close?
Nobody has shown which came first. Close families choosing to play more fits the survey evidence as well as play making families close does, and the researchers say so in their own papers.
Two stories, one set of numbers
Surveys find that families who do more together rate their family life higher. The saying reads that one way: play first, closeness second.
There is a second way to read it. A family that already gets along will choose to spend time together and will rate that time highly. A family in conflict will avoid it. A survey taken once cannot tell these two stories apart. Both produce the same numbers.
Researchers have two names for the problem.
- Reverse direction. The closeness may be producing the play.
- Selection. Families who play a lot may differ from other families in ways that have nothing to do with play, and those differences may explain the result.
The oldest review already said it
A 1984 review of the research on family and leisure concluded: "It is concluded that leisure behavior appears to be both affected by, and to affect, marriage and family factors. Research to date, however, has been largely exploratory and descriptive." (Journal of Leisure Research, 1984). We read the summary of this review, not the full paper.
Note the order of the words. The phrase "affected by" comes first. The oldest review we found already said that family life shapes leisure.
The researchers' own words
The people who produced the main findings say plainly what their studies cannot show. Here are five statements from five papers, all of which we read in full.
- "interpretations cannot assume causality" (Family Relations, 2001)
- "this directionality has not been empirically tested" (Journal of Leisure Research, 2003)
- "structural equation modeling is still a correlational method" (Journal of Leisure Research, 2010)
- "Causal inferences cannot be made with confidence in the absence of experimentation." (Leisure Sciences, 2016)
- "data are cross-sectional and analysis techniques were correlational, hence no assumptions of causality can be made" (Journal of Leisure Research, 2018)
Some of those terms need a plain version. Causality means one thing making another happen. Cross-sectional means a survey taken once. Correlational means the study shows two things moving together and no more than that.
A 2017 review of the main framework in this field made the same call. It said "the need for future study design and analysis that contributes to the determination of directionality is vital" (Leisure Sciences, 2017).
What reverse direction looks like
Two stronger studies show the reverse story in action. Neither is about family play.
Couples. A Dutch study followed 4,255 couples and found that those who shared more of their lives, including free time, were less likely to split up (Demographic Research, 2018). The authors do not claim that shared time protected them. They write: "The present study is not aimed at strong causal inferences". They also give the reverse reading themselves: "Couples at high risk of separation might feel unsatisfied with their relationship and therefore withdraw from spending leisure time together."
Family dinner. In a study of 17,977 US teenagers, taking account of how close the family already was removed 49% to 61% of the link between family dinners and how the teenagers were doing (Journal of Marriage and Family, 2012). The authors name the problem outright: "Our estimates of family dinner effects may be upwardly biased by reverse causation". We cover that study on its own page, the family dinner lesson.
What selection looks like
Families who do a lot together differ from other families in other ways too.
Religion. One study asked 121 parents and 99 young people from various religious groups about family leisure, religious life and how the family worked (Marriage and Family Review, 2007). Both were linked with how the family worked. For parents, religion came out ahead: "Family religiosity was the strongest predictor of family functioning for parents."
Education. A study of 11 Western countries from 1965 to 2012 found that "For both mothers and fathers, results indicated a widespread educational gradient and an increase in child-care time." (Journal of Marriage and Family, 2016). An educational gradient means that parents with more schooling spent more time on their children. A US study of 6,640 mothers found that "differences across education groups in basic care and play time are largest among mothers with infants or toddlers" (Demography, 2012). We read the summaries of these two studies, not the full papers.
So time with children tracks schooling. A survey that links family time to a good family life has to deal with that.
What would settle it
Two kinds of study could point the arrow.
- An experiment that assigns some families to more shared play and others to carry on as usual.
- A study that follows the same families over years and measures play before closeness changes.
Both are rare. 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).
Where that leaves the saying
The link between playing together and being close is real in the surveys. The direction is not known. It is fair to say that close families tend to be the ones who play together. It is not yet fair to say that playing together is what made them close.
Both stories may be partly true. The 1984 review said as much. No study we found has measured how much of the link belongs to each.
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?