If You're Happy and You Know It: How Do Mothers and Fathers in the US Really Feel about Caring for Their Children?

ABSTRACT

This paper considers the question posed by popular media: Do women like doing childcare more than men? Using contemporaneous subjective well-being data paired with 24-hour time diaries from the 2010 American Time Use Survey, the paper explores gender differences in how mothers and fathers feel when engaged in a set of common daily activities. We find that both mothers and fathers engaged in child caregiving enjoy their time spent in child caregiving; fathers as much, or even more so, than mothers as evidenced by their average values for happiness, meaningfulness, tiredness, and stress and an aggregated statistic, the unpleasantness index. Simulations provide evidence that the difference between mothers and fathers comes almost completely from differences in their subjective well-being rather than from differences in how they use their time.

KEYWORDS:

• Time use
• subjective well-being
• childcare
• gender wage gap
• happiness
• experience emotions

JEL codes:

• D13
• J13
• J16

NOTES ON CONTRIBUTORS

Rachel Connelly is Bion R. Cram Professor of Economics at Bowdoin College, Research Fellow at the Institute for the Study of Labor (IZA), and an Associate Editor of Feminist Economics. Overall her research focuses on the intersection of economic demography and labor markets, most often on the economics of mothers’ employment and childcare. In her recent US-based work, Connelly examines the time use of mothers and fathers. In her China-based work, Connelly studies the relationship between family structure, age, migration, and women's employment.

Jean Kimmel is Professor of Economics at Western Michigan University (WMU) and Research Fellow at the Institute for the Study of Labor (IZA). Prior to joining the WMU faculty in 2001, she was Senior Economist at the W.E. Upjohn Institute for Employment Research. Kimmel is a labor economist with a broad range of research interests, including the motherhood wage gap, mothers’ time use, childcare availability and choices, multiple job holding, the racial digital divide, and most recently, gender differences in subjective well-being.

ACKNOWLEDGMENTS

The authors would like to thank the four anonymous referees and the anonymous Associate Editor as well as participants at IAFFE session at the 2013 ASSA meeting in San Diego, California, and China Center for Human Capital and Labor Market Research seminar in March 2013 for their helpful comments and suggestions.

Notes

¹ This article was published three days later in the New York Times magazine with the title, “Diaper Changing Index” (March 25, 2012, p. 16).

² Specifically, the research was based on a survey of 184 academic couples with at least one child under the age 2, all on the tenure track at universities. Additionally, the authors discuss concerns with potential respondent bias inherent in the survey instrument design.

³ See, for example, Camille Ryan and Julie Siebens (2012) and Judith McDonald and Robert Thornton (2007).

⁴ Also see Joni Hersch and Leslie Stratton (1997) and Mark Bryan and Almudena Sevilla-Sanz (2011).

⁵ The U-index is drawn from Daniel Kahneman and Alan Krueger's (2006) work with similar data.

⁶ See Almudena Sevilla, José Gimenez-Nadal, and Jonathan Gershuny (2012) for a similar comparison of RSWB and CSWB measures.

⁷ F. Thomas Juster, Paul Courant, and Gregory Dow define process benefits as the

direct subjective consequences from engaging in some activities to the exclusion of others … For instance, how much an individual likes or dislikes the activity ‘painting one's house,’ in conjunction with the amount of time one spends in painting the house, is as important determinant of well-being independent of how satisfied one feels about having a freshly painted house.

For a presentation of an underlying theoretical model of utility, see Alan Krueger (1985: 120–1), Daniel Kahneman, David Schkade, Norbert Schwarz, and Arthur Stone (2009).

⁸ Alexander Szalai (1972), the pioneer of time-diary research, also experimented with collecting CSWB information together with yesterday's time diary.

⁹ “Positive” was the emotion included in the PATS and excluded from the ATUS. The ATUS added questions about the meaningfulness of the activity not found in the PATS.

¹⁰ José I. Gimenez-Nadal and Almudena Sevilla-Sanz (2011) consider the effect of gender and parenting on “leisure satisfaction.” Leisure satisfaction is somewhat different from GRSWB measures, as it comes from a question in the European Community Household Panel (ECHP) that asks “How satisfied are you with the amount of leisure time you have?” Thus, it is a measure of satisfaction with the amount of time one has to devote to leisure, rather than a measure of the satisfaction one gets out of the specific leisure activities in which one is engaged. Gimenez-Nadal and Sevilla-Sanz find that mothers of children 0 to 17 years of age have lower levels of satisfaction with the amount of leisure time they have compared with fathers of children the same age (as seen in the statistical significance of the coefficient of the interaction between number of children and the female indicator variable).

¹¹ Bertrand (2013) does not include tiredness as a negative emotion. Since tiredness is high for mothers, including tiredness would be expected to change some of her findings.

¹² Wang (2013) writes that women find more meaning in child caregiving too, but this difference is very small and unlikely to be statistically significant.

¹³ The ATUS also has a series of questions about “whether children were in your care” during the activity. These questions are asked after the time-diary data collection is completely finished. The interviewer returns to each activity reported and asks if, during that activity, there was a child under age 13 in your care. These responses to these questions are often called secondary child caregiving and were designed as a compromise between survey length and the desire to measure multitasking. However, previous research has shown that the secondary child caregiving questions in the ATUS produce much higher estimates of secondary child-caregiving time than surveys that allow all two uses of time at any moment (Suzanne Bianchi, Vanessa Wight, and Sara Raley 2005; Kimberly Fisher 2005) because some parents report that every activity that takes place at home when a young child is present in the home is an activity in which a child is under one's care. This is, of course, technically true, and gets at the problem of measuring child-caregiving time articulated so well by Nancy Folbre, Jayoung Yoon, Kade Finnoff, and Allison Fuligni (2005) and Nancy Folbre and Jayoung Yoon (2008). But given the comparative estimates of Bianchi, Wight, and Raley (2005) and Fisher (2005), we believe the alternative measure of “was there a child in the room during the activity?” is most likely to affect CSWB.

¹⁴ Because of lower response rates by men than women and the sample selection criterion that one must report at least one child-caregiving activity on diary day, our analysis sample comprises 33.8 percent men and 66.2 percent women. However, all results reported in the paper are weighted using BLS supplied weights to return the sample to the population proportions. The differential in sample inclusion rates of men versus women may lead to the concern that the fathers included in the analysis sample are self-selected to be those who enjoy child caregiving differently than mothers. This would be the case if fathers are allowed to opt out of child caregiving and mothers are not. It is difficult to think of how to “fix” this problem, since one cannot ask about experienced emotion for an activity that has not happened. What we can do is acknowledge that all reported emotions are conditional on having engaged in that activity the previous day, plus that the entire sample of fathers (and mothers) is selected on the criterion that they engaged in child caregiving minutes the previous day.

¹⁵ We include an activity if there are thirty or more CSWB responses for both mothers and fathers. See the Data Appendix for further detail on how the activities were chosen for inclusion.

¹⁶ In Connelly and Kimmel (forthcoming), developmental caregiving time includes time spent reading to children, helping with homework, playing with children, talking and listening to children, and school conferences. Maintenance child caregiving includes physical care of children, organization and planning related to children, attending children's events, watching children, waiting for children, and providing food and obtaining medical care for children.

¹⁷ See Gigi Foster and Charlene Kalenkoski (2012) for an experiment in multitasking. The child caregiving task in their experiment involved enough down time that respondents could also do the laundry task productively at the same time. However, in the experiment by Thomas Buser and Noemi Peter (2012), both tasks were problem-solving intensive; thus, the productivity advantage went to solo-tasking as compared to multitasking.

¹⁸ Daniel Hamermesh and Jungmin Lee (2007) note that women report a greater sense of feeling rushed, and this gender difference may be the result of women changing activities more often throughout the diary day.

¹⁹ Marie Connolly (2013) describes the relationship between weather and reported well-being, thus suggesting the importance of controlling the diary month in our regression.

²⁰ The full regression results (84 separate regressions; 21 activities × 4 emotions) are available from the authors upon request.

²¹ The parallel results for the remaining two emotions, sadness and pain, are available from the authors. These scores are much lower overall and few gender differences are observed.

²² Mothers' tiredness is also not simply a function of getting less sleep. In fact, the mothers in our sample report thirty minutes more sleep on their diary day than fathers (a statistically significant difference).

²³ Tie scores go to pleasant; that is, for an activity to be characterized as unpleasant the highest negative score must be strictly greater than the highest positive score.

²⁴ Bertrand (2013) excludes tiredness from her measure of an unpleasant activity. In addition, the U-index she calculates is very different from Kahneman and Krueger's, as it is based on just the three activities that each individual was asked about and she does not control for different probabilities of being asked about a specific activity based on differences in time use. Kahneman and Krueger's U-index (and ours) is based on average emotion scores by activity (averaged over all people in the relevant sample) and individual time-use proportions for all waking activities of the day (not just the three for which CSWB questions were asked).

²⁵ This is analogous to the Oaxaca–Blinder decomposition of the wage gap, which decomposes differences attributable to composition effects from behavioral effects. The composition effects in this case as differences in how time is used, while the behavioral effect is how unpleasant activities are judged.

²⁶ The counterfactual for mothers if they used their time the ways that fathers use their time is shown in Supplemental Table 2, available online on the publisher's website.

²⁷ Krueger (2007) compares the size of gender gaps he calculates (based on non-gendered activity level U-indices) to the size of the gap between weekday and weekend U-indices. The weekend/weekday gap he reports is about 3 percentage points, making a 9 percentage point gap quite large indeed.