In a recent study, we looked at whether the amount of time mothers spend with their children is related to children’s and adolescents’ well-being. The study garnered a lot of attention from the public and the media—including some criticism—in part because it had a perhaps surprising finding: We found that the quantity of time mothers spend with children – either engaged in activities with them or just present – doesn’t have a link to the emotional or behavioral health, or math and reading scores, of kids aged 3 to 11. We did find a relationship for teenagers, though, with more quantity time engaged with mothers in activities related to less delinquent behavior among teens and engaged time with mothers and fathers together linked to better behavioral health, math scores and less risky behavior.
In New York Times Upshot columns responding to our study, which was first covered in The Post, the economist Justin Wolfers questioned our findings, saying the measure of maternal time we used might not be a reliable indicator because it might capture atypical days. He uses the example of family vacation to Disney World – if a child was randomly sampled on a trip there, the diary might show that the child and mother were together for more hours than if the child were sampled on a more typical day. This could make the data “noisier” and thus harder to detect an effect of mothers’ time even if one did exist. He suggests that the diary data may have missed capturing a positive effect of maternal time on child outcomes.
But we don’t think the critique stands. Here are the reasons why.
Time diary data are high quality. To generate our data, we used time-diary data from the Panel Study of Income Dynamics, Child Development Supplement (PSID-CDS), a nationally representative sample of families with children. In the time diaries, children’s activities were recorded in sequence for a 24-hour period, including a randomly sampled weekend day and weekday. Time diaries are excellent sources of data on how people spend their time, what activities they do, with whom they do the activities, where the activities take place, and who else is present (Bianchi, Robinson and Milkie 2006; Hofferth 2006). Time diaries allow respondents to record their time in activities in sequence, a common recall task, and the activities must add up to 24 hours, so there is little room for embellishment. Because the diaries are not “about” parenting, issues of social desirability are muted compared with survey questions that ask parents to retrospectively report on time they “typically” spend with children, or time they spent with children in the past week or month. In these survey questions, parents are more apt to inflate time spent with their children and/or state they did fewer hours of less culturally valued activities like watching TV with kids because they want to be seen as good mothers or fathers.
Additional analysis shows neutral or “non-findings” are the same for children with “very typical” time use. The dataset has information as to how typical the diary days were for the child. As reported in our study, we took these “typical day” indicators into account in our analyses. Here, we isolate the sample to families who report that their time diary days reflected “very typical” days, the sample of children with the least “noise,” to illustrate that the “non-findings” are the same with this more select group of 3-11 year olds. Below is an example of kids’ behavior among just those whose data came from “very typical” days -- even as hours of time engaged with mom increase, the average level of behavior problems and emotional problems is the same. If more time in activities with mothers was better, we would see shorter bars (fewer problems) for those with the most hours with mom.
In the following chart, emotional problems are measured as a scale ranging from 0 to 13. Behavioral problems are measured as a scale ranging from 0 to 15. As you can see, there is actually little difference in the average levels of emotional problems or behavioral problems among children aged 3 to 11 who spend engaged time with their mothers for 5, 10, 15, or 20 hours per week.
Still, it is important to consider potential missed associations. Detecting associations between more general or common activities (like time with mother) and well-being, however, should be easier than detecting relationships between infrequent activities (like reading) and well-being. Notably, relationships between the total amount of maternal and parental engaged time are detectable in adolescence within our study, so relationships should be detectable in childhood too.
Other research supports the lack of a positive influence of quantity time at younger ages – indeed a recent study shows the sheer amount of maternal time has a small negative effect at that life stage, and provides nuance regarding quantity and quality time. A parallel recent study focused on maternal employment, by sociologist Amy Hsin and economist Christina Felfe published in Demography -- using the same time diary data from the PSID-CDS, but with a different sample, age groupings, aggregations of maternal time, and analysis -- finds that for children under age 6, more time with mothers is related to worse outcomes. This negative effect is likely driven by a lot of “unstructured” time which includes shared time in activities like watching TV or playing video games. Similar to our study, for older offspring, more mother time is positively correlated with positive behavior. Across all age groups, Hsin and Felfe find that more time with mothers doing educational activities (e.g. reading) or structured activities (e.g. talking, having meals together, attending events)—is positively correlated with cognitive and behavioral outcomes of children. In a nutshell, this study shows that more “quality” time with mom -- educational and structured activities -- is good for kids, but this can be offset by mother-child “non-quality” time (e.g., TV)—that, even with mother present -- is not good for (especially young) children. Yet how more maternal time matters is dependent on life stage, as our study also indicates.
Other studies show parents’ involvement might not matter like we think it does. Non-findings or counter findings may challenge our beliefs. Other recent studies by sociologists may counter popular beliefs about how parental involvement influences child outcomes. Keith Robinson and Angel Harris’s 2013 book, The Broken Compass, found that many measures of parental involvement in children's school activities, widely believed to be important, were not necessarily related to children's academic outcomes. Ann Meier and Kelly Musick’s 2014 study found that the link between family dinner and child well-being depends on the quality of the parent-child relationship.
Reactions to issues related to time with children. This research study has perhaps generated reactions not only because findings may challenge current beliefs about maternal time, but because feelings about time with children are intense for many parents. Our previous research shows that many U.S. parents do not feel they have enough time with their children (in part linked to their long work hours), and that these feelings are linked to parental work-family conflict, and lower psychological well-being.
Melissa Milkie is professor of sociology at the University of Toronto. Kei Komaguchi is an associate professor of sociology at Bowling Green State University. Kathleen Denny is a lecturer in the Sociology Department at the University of Maryland.