1 Introduction

How much time do parents devote to the care of their children, and how is the quantity of parental time affected by factors such as family income, education, and paid employment? The advent of new large-scale time-diary surveys administered to representative samples of national populations makes it possible to answer such questions with some precision. But quantitative precision can conceal conceptual ambiguity. In this paper, we explore several inconsistencies in the temporal categorization of child care and explain their relevance to the measurement of differences and trends in parental child care time.

We begin with a general overview of the conceptual issues at stake, illustrated by a comparison of coding categories used in the recent time-diary surveys of four major English-speaking countries: Canada (1998), Australia (1997), the United Kingdom (2000), and the United States (2003). Lack of consensus over the appropriate boundary between child care as a primary activity (“what were you doing?”) or a secondary activity (“were you doing anything else at the same time”) has spilled over into inconsistencies in the measurement of child care as a primary activity. Inconsistencies are even greater in broader measures, whether these are based on reports of a secondary activity or stylized questions regarding time “looking after children” or time in which children were “in your care.” Our comparison of the Canadian and US surveys, which both used stylized questions designed to capture supervisory responsibilities, shows that small differences in wording led to significant differences in the reported results.

The second section of the paper examines characteristics of primary child care activity and supervisory child care in the American Time Use Survey 2003. Focusing on married or cohabiting adults living in households with at least one child under the age of six but no child over age 12, we explore the distribution of these types of care, their overlaps, and their density (defined as the ratio of children to adults present during a reported care activity or responsibility). We show that weekly and seasonal variations in primary care activities differ from those in supervisory care. The relationships between maternal work hours and these two types of child care time also differ. These results suggest that analysis of primary child care activities alone provides an incomplete picture of the temporal demands that children impose.

2 Defining child care

2.1 The definitional dilemma

Most time-use surveys categorize child care in terms of “activities.” Unfortunately, humans are “multitasking beings,” whose activities often elude clear categorization (Harvey & Royal, 2000: 8). Most surveys define “primary activities” in response to a question such as “What were you doing during this time period?” The recent Australian and United Kingdom surveys designated “secondary” activities in response to questions such as “Were you doing anything else at the time?” The list of activities listed as “secondary” often includes leisure activities such as “listening to the radio” or “talking with friends.” But as other time-use researchers have pointed out, child care frequently shows up as a secondary activity (Ironmonger, 2004). Such child care is conventionally termed “secondary child care.”

Primary child care activities should not be conflated with “time devoted to children” or “time that parents spend with children,” although researchers aware of these distinctions sometimes use these terms synonymously (Gauthier, Furstenberg Jr., & Smeeding, 2004). Even measures of secondary childcare fail to capture passive or supervisory care that does not take the form of an activity. Adults are often constrained by the need to supervise or assume responsibility for young children, whether or not they are engaged in an explicit activity with them (Budig & Folbre, 2004; Folbre, Yoon, Finnoff, & Fuligni, 2005).

Neither the Canadian nor US national surveys question respondents concerning secondary activities. Both surveys, however, acknowledge the diffuse nature of child care by including a special child care module designed to ascertain if individuals were “looking after children” (the Canadian wording) or if “children were in their care” (the US wording) (See Table 1). Answers to these questions are typically reported as “secondary” child care activity (Fedick, Pacholok, & Gauthier, 2005; ATUS published Tables). Yet in the US case, the wording was explicitly designed to capture supervisory responsibility that did not necessarily take the form of an “activity” (Horrigan & Herz, 2004).

Table 1 Child care measures in time use surveys of four major English-speaking countries
Full size table

All four surveys considered here asked questions regarding whom respondents were with, or “who else was there” while they were engaged in activity. The American Time Use Survey (ATUS) specifically asked if there was another person “in the same room” if the activity was being conducted at home or “accompanying them” if not at home; multiple individuals could be listed. International comparisons are troubled by some inconsistencies in spatial specification—unless restricted to “in the same room” respondents may interpret the question more broadly, such as “in the same house” or “within shouting distance” (Harvey & Royal, 2000). Still, the “with whom” question provides another vantage point for assessing care responsibilities, one that has often been described as similar to the “in your care” designation because it, too, reaches beyond “activity.” Footnote 1

But it is important to emphasize that the mere presence of a child is conceptually distinct from having a child “in your care.” A child can be “in the care” of an adult even if napping in another room or playing in the backyard. Indeed, such spatial separation is a typical feature of passive or supervisory care. On the other hand, the “with whom” variable could overstate child care responsibilities by extending their definition to include social activities in which many adults are present, sharing responsibility for a small child. Many activities reported as leisure fall into this category (Mattingly & Bianchi, 2003; Bittman & Wajcman, 2004).

Few time-use surveys ask “for whom” an activity is conducted, although this contextual variable could add important information (Harvey & Royal, 2000). Children increase the burden of domestic labor in their households: their food must be prepared, their clothes must be washed, their shoes must be purchased, and their toys must be picked up. Yet many domestic activities benefit all household members at once: food is prepared for the entire family; laundry is mixed, et cetera. Multivariate analysis of the impact of number and age of children on the level and distribution of domestic work probably offers a better way of quantifying this effect (Craig, 2005).

A closer look at coding categories used by most recent time-use surveys of the major English-speaking countries illustrates some of the specific difficulties of defining and measuring child care.

2.2 Child care as primary activity

Measures of child care as a primary activity are more consistent across surveys than other measures. Even here, however, important discrepancies emerge. Footnote 2 Both the US and Australian surveys—unlike those conducted in Canada and the United Kingdom—include a coding category that confounds the otherwise rather tidy distinction between primary and non-primary child care time by including supervision of children—so called “passive care”—as a separate subcategory of the primary care activity. The ATUS uses the phrase “looking after children” while the Australian time-use survey uses the phrase “minding children.”

Appendix A lists the actual activity codes used for primary child care in the four major English-speaking countries. There is no exact equivalent to ATUS code 03-01-09 or to Australian code 54 in the other two surveys, although the UK survey adds “supervision of a child” onto a residual category, (3819-Other specified physical care and supervision of a child). The US and Australian primary child care activity codes represent a more concerted effort to capture more passive forms of care. According to the published ATUS Tables, about 5% of women’s and about 7% of men’s total child care time in households with the youngest child under 6 was devoted to the very general activity of “looking after children.” Footnote 3

The Australian and US surveys also made stronger efforts to measure the time devoted to communication with others on behalf of children, which might be termed the “transactions” dimension of primary child care. As designers of the US survey explain,

The Bureau of Labor Statistics (BLS) coding team conceptually defined primary childcare as any activity done with a child that is interactive in nature—such as reading, playing, and talking—and correctly coding such activities posed few difficulties. However, other activities were considered primary childcare as well, but were not limited to this restrictive definition requiring interaction with a child. For example, an activity could be coded as child care if a child was not present but the activity (such as “talking to my child’s teacher”) was clearly done in the child’s interest or on the child’s behalf” (Shelley, 2005:5). Footnote 4

In other words, the ATUS implicitly resorted to a “for whom” rather than an “activity with” definition of child care where activities related to health care and educational needs were concerned. Similarly, Australian codes 55 and 57 called attention to communication with others on behalf of children. Neither the Canadian nor UK time-use surveys make explicit mention of such activities. (All the surveys include a separate travel category for travel related to child care).

Australian and US efforts to capture the time devoted to managing the provision of health care and education services are admirable, since management represents an important component of parental responsibilities. But the inclusion of some activities conducted on behalf of children—whether or not they are physically present—is inconsistent with the exclusion of additional housework resulting from children. It might even introduce a class bias: High-education and high-income households may be more likely to spend time negotiating with doctors, teachers, coaches, and nannies (counted as primary care activities) on behalf of their children or driving them to soccer games and piano lessons. Low-education and low-income households may be more likely to spend time cooking, cleaning, and doing laundry for their children (not counted as primary care activities).

Inconsistencies in coding of primary activities should not discourage international comparisons but redirect them toward more specific comparisons. Researchers could devote more attention to specific categories such as “developmental” care (e.g. reading to children or helping them with homework), “physical” care (e.g. feeding and bathing), “low-intensity” or supervisory care, and “managerial” care. Multivariate analysis can be used to explore the impact of household characteristics on specific activities rather than primary child care activity in general (Bittman & Wajcman, 2004; Craig, 2005; Zuzanek, 2001).

2.3 Secondary child care activity and supervisory responsibility

More careful analysis of both secondary activities and measures of care responsibility could help build a stronger consensus on how child care should be defined. Unlike many other non-market work activities such as “cooking dinner” or “doing laundry,” child care often involves complex multi-tasking. Both Australian and British surveys collected information on secondary activities, including child care. Unfortunately, differences in specification of activity codes described above may have affected the results. Further, the Australian survey primed respondents by including instructions on the written form they were asked to fill out that called attention to specific activities. Under the second column of the diary form, headed by the question “What else were you doing at the same time?” three specific examples were listed: “e.g. childminding, watching television, listening to the radio” (Ironmonger, 2004: 95).

This priming of respondents almost certainly helps explain why secondary child care time represented more than 20% of all secondary activity time reported in Australia in 1997, but only slightly more than 10% of all secondary activity time reported in the United Kingdom in 2000 (Ironmonger, 2004: 98–99; UK, 2003: 32). Footnote 5 The impact of activity codes and survey instructions on measurement of child care as a secondary activity could also help explain why national surveys that list only a few child care activities, like the most recent Korean survey, do not yield very high estimates of secondary child care (Yoon, Unpublished).

As aforementioned, neither the Canadian nor US time surveys attempted to measure secondary activities, but relied instead on a special question aimed to capture care responsibility (in addition to collecting information about who else was present). The US survey design was influenced by the 1992 Canadian precedent (Frederick, 1993) as well as by cognitive studies of the impact of alternative wording on measures of time devoted to non-primary care (Schwartz, 2001).

A recent analysis of the 1998 Canadian survey provides a fascinating window into the relationship between three different measures of child care time: time that child care activities were reported as the primary activity, time that adults were “looking after” children, and time in which a child was listed as present while an adult was engaged in an activity (Fedick et al., 2005:17). Footnote 6 Results were tabulated for employed males, non-employed males, employed females, and non-employed females who were married or cohabiting, with at least one child under the age of five in the household (See Table 2).

Table 2 Estimates of mean time devoted to child care by married or cohabiting canadians with at least one child under the age of five in the household in 1998, by Fedick et al.a,b,c (hours per day)
Full size table

For all groups except employed males, the time spent “looking after children” in the special module was highest, total time spent in presence of children next, and time spent in child care as a primary activity was lowest. Fedick et al. emphasize that two broader measures, total time spent in presence of children and total time spent “looking after children” are relatively similar. Still, the difference between them amounts to more than an hour a day for employed women, or more than 7 h per week (a difference of about 18%). While the two measures are clearly related, Physical proximity and “looking after” children are related but distinctly different measures of child care.

2.4 Comparisons between Canadian and US measures of child care

In order to illustrate the impact of different survey definitions of child care we compare the estimates of child care time provided by Fedick et al. for the 1998 Canadian Social Survey with parallel results from the 2003 ATUS. We impose similar restrictions on the universe, which consists only of married or cohabiting adults in households with at least one child under five. In both cases, observations for weekdays, Saturdays, and Sundays are weighted to derive estimates for an average day.

While the 2003 ATUS took an approach similar to the Canadian survey of 1998, it used the term “in your care” rather than “looking after” children, a difference that the cognitive studies conducted by the U.S. Bureau of Labor Statistics showed would likely be significant (Schwartz, 2001). The ATUS restricted the time that children could be in an adult’s care to time between when the first child woke up and the last child went to bed. Footnote 7 Excluded from this category was all time that an adult was engaged in child care as a primary activity. The comparable measure for the Canadian survey is all time spent with children minus time devoted to primary care with children (see Row D of Table 2).

Sample sizes were larger for the US survey, and the alternative measures apply to exactly the same group of people, which was not the case for the Canadian survey. As a result of these minor differences, as well as the more recent date of the US survey, the two surveys are not perfectly comparable. But cross-national differences in the amount of time devoted to primary child care activities are relatively small (less than 15% for every demographic category), and display a relatively consistent pattern (the US estimates are lower for three out of the four categories). Similarly, comparisons of “time with” children are similar across the two countries. For the US the averages are consistently higher than for Canada, but never by more than 15%.

By contrast, differences in the stylized measures of child care time that are not activity-based loom large. Employed women in the US reported spending 30% more time with “children in their care” than their Canadian counterparts reported “looking after children.” Non-employed women reported 53% more. Footnote 8 Employed and non-employed men also reported more time fulfilling supervisory responsibilities in the US than in Canada.

The contrast is heightened by comparison of time devoted to primary child care activities as a percentage of total time devoted to children (See Row F of Table 2 and Row E of Table 3). The cross-country differences are particularly striking for non-employed men and women. In Canada, the primary child care activity time these two groups reported amounted to about 40 and 38% of the total, respectively. In the US, they amounted to only about 18 and 28% of the total, respectively. Both country surveys measured supervisory responsibilities using similar methods, but the differences in wording yielded very different results.

Table 3 Estimates of mean time devoted to child care by married or cohabiting US residents with at least one child under the age of five in the household in 2003 (hours per day, n = 3,080)
Full size table

In sum, international measures of the larger temporal demands of child care (beyond primary activities of care) are even less comparable than measures of aggregated primary child care activities. Differences in survey design probably affected the measures of secondary child care activity collected by Australia and Great Britain, and differences in wording had an even larger effect on measures of supervisory responsibility collected by Canada and the United States. A closer look at the relationship between time devoted to primary child care activities and time devoted to supervisory responsibilities in the ATUS provides some insights relevant to future survey design.

3 Activities versus responsibilities in the ATUS

Most empirical studies of child care time focus on the aggregate amount of time in primary care activities, examining differences based on factors such as gender, education, and hours of employment. The detailed structure and relatively large sample size of the ATUS offer a powerful lens for magnifying this analysis and examining disaggregated primary activities, the temporal structure of supervisory responsibility, and the social characteristics of care time.

In order to explore the relationship among different measures of child care we select a subset of the entire ATUS universe, married or cohabiting individuals living in households with a child under the age of 13, with at least one child under the age of 6. We eliminate all individuals living in households with a child between the ages of 13 and 17, because the “in your care” question was asked only regarding children ages 12 or under. Comparisons of care activity and “in your care” would be confounded by the presence of older children.

3.1 Distributions of different types of care

Few people engage in all of the activities coded in a time-diary survey on the actual survey day. As a result, reports of time devoted to many activities, including child care activities, are characterized by a high number of zeros. Frequency distributions show observations are heavily skewed to the left with large numbers reporting small quantities of time reported, as well as zeros. A small number of high values typically create a long tail. Of all the time devoted to different types of activity, the only one that resembles a normal distribution is time devoted to sleep. Among activities that average less than an hour a day, the standard deviation is typically higher than the mean. Even among married or cohabiting individuals living in a household with a child under the age of six but no child age 13 or over, more than 40% of men and 27% of women report zero time devoted to child care on the survey day (See Appendix B).

The distribution of supervisory responsibility time among these individuals takes a very different shape, for both men and women. While the proportion of zeros reported is much smaller, the distribution appears almost rectangular for men, but somewhat more bell-shaped for women (See Appendix B). These distinctly non-normal distributions have important implications for statistical inference.

3.2 Overlaps among different types of care

Temporal overlaps among different measures of child care offer some important insights into their similarities and differences. In order to summarize the overlaps of time in which adults reported a child care activity, a child “in their care” or a child physically present, we group primary child care activities into four categories: physical care, developmental care, managerial/logistical care, and low-intensity care (See Table 4). Physical care represents the least discretionary activities that are necessary to meet children’s basic needs—feeding, bathing, dressing, and attending to medical needs.

Educational/developmental care represents activities that directly contribute to children’s physical, emotional, and cognitive development. Managerial/logistical care represents activities performed on behalf of children, such as making arrangements on their behalf or transporting them to school or sports activities. Low-intensity care represents activities that require relatively little engagement with the child. As noted above, the latter two categories lie between the active engagement usually associated with primary care activities and the more passive constraints of supervisory care.

Among all married or cohabiting adults living in households with a child under the age of 6 but no child over the age of 12, the average amount of time devoted to primary child care activities was about 2.2 h per day. Physical care accounted for almost half this time (.96 h or about 44%). Educational/developmental care accounted for .75 h, or about 35% of the total). Logistical/managerial care accounted for about .31 h, or about 15% of the total. The average ratio of children to adults present during periods of active care was about 1.5. Footnote 9

Most of the time that adults reported a child care activity they also reported that a child was “in their care” (92.6% of time) and it was even more likely that a child would be present (94.8% of time). Within some sub-categories, however, there is surprisingly little overlap. For instance, among activities of managerial/logistical care, the overlap with children “in your care” was only 78.5% and a child was present only 79.5% of the time. Specifically, in care-related travel for a household child, a child was present only about 75% of the time, probably because this travel includes time that an adult is returning alone from a trip to drop off a child at school or at another activity. This is the most important example of an activity coded as primary child care that does not always involve direct interaction with a child.

3.3 Characteristics of time with a child “in your care”

Adults could report that a child was “in their care” during any activity except sleep. On average, married and cohabiting adults with at least one child under six and no child older than 12 reported that children were in their care about 26% of the entire sample day or about 6.3 h (not counting time that they were also engaged in a primary child care activity). This represents almost three times as much as the 2.2 h average amount of time devoted to primary care activities. The total average amount of time devoted to or constrained by children was about 8.5 h, more than twice the average amount of time devoted to paid work.

Supervisory responsibilities extended well beyond physical proximity. A child was listed as present only about 68% of the time that an adult reported a child “in their care” (See Tables 4, 5). A child that was not physically present could be playing in another room or in the backyard or perhaps taking a daytime nap. Some kinds of adult activities were more likely than others to be combined with “in your care” time. Adults were most likely to report children “in their care” while engaged in eating/drinking/socializing/leisure activities and religious/community activities (over 60% of the time) (See Table 5). In both these categories, children were often listed as present (with an 80% average overlap). Household work and logistical/managerial activities were next in importance, overlapped with “in your care” more than 40% of the time. But children were less likely to be in the same room for household work activities (with a 59% overlap) than for logistical/managerial activities (with a 93% overlap).

Table 4 Overlaps among primary child care activities, children "in your care" and "child present" for married and cohabiting persons in a household with youngest child under 6 but no child over 12, 2003 ATUS (n = 3080)
Full size table
Table 5 Overlaps among adult activities and child “in your care” and child present for married and cohabiting persons in a household with the youngest child under 6 but no child over 12, 2003 ATUS (n = 3080)
Full size table

In general, women were more likely to report taking supervisory responsibility for children than men were. Since only one person per household was surveyed, it is unclear to what extent adults may share supervisory responsibilities. Both a mother and a father could report simultaneous supervisory responsibilities. Taking advantage of information in the survey about “who else was present,” we calculated what percentage of the time that both a spouse and a child were present a male or female reported a child was “in their care.” Among the two activities above in which supervisory responsibilities were most likely to be reported, the incidence was high for both women and men, but slightly higher for women. For instance, women who were engaged in eating/drinking/socializing/leisure with both a spouse and a child present reported a child was in their care about 95% of the time, men about 83% of the time. Women seem more likely to assume responsibility for children, and the ratio of children to adults tends to be higher for their supervisory responsibility time. Still, supervisory responsibility may often be diffused across more than one adult.

3.4 Effects of age and number of children

Child care responsibilities are directly affected by the number and age of children in the household. These factors have different effects on care activities and on supervisory responsibility. Activities of child care decline more steeply than supervisory time as children age; neither form of time utilization increases much with number of children. Table 6 cross-tabulates age of youngest child and number of children for both active and supervisory care, normalizing both measures to 1 for households with only one child under the age of 1. This normalization highlights the percentage increase in time associated with changes in demographic composition. In households in which the youngest child is under the age of 1, for instance, the additional activity time associated with an additional child is 1.11 or 11%. The addition of a third child to a household in which the youngest is under the age of 1 is associated with a reduction in time on child care activity, probably due to child spacing and sibling care effects. Footnote 10 In general, the addition of a child to the household tends to have relatively small effects on both types of child care time except when the addition of the fourth child is associated with the peaks in supervisory time. The decline in child care activity as age of youngest child increases is far more noticeable. Supervisory time does not decline as steeply.

Table 6 Cross tabulation of child care activities and supervisory responsibilities for children by age of youngest child and number of children in household, normalized values (married and cohabiting persons in a household with the youngest child under 6 but no child over 12, 2003 ATUS (n = 3080))
Full size table

3.5 Tradeoffs between care activity and care responsibility

However large the differences between the absolute amounts of time devoted to care activities and time constrained by supervisory responsibilities, these would be less important if the two types of care time were linked. For instance, time devoted to care activities could serve as an indicator of the time devoted to supervisory responsibility. Interestingly, however, the two are not significantly correlated, perhaps because men tend to provide more supervisory than active care time. Disaggregation by gender reveals no significant correlation between the two types of time use for men, but for women there is a negative and significant Pearson correlation of −.17, and a Spearman’s rho of −.18. For women, at least, active care time and supervisory time appear to be substitutes.

This observation is consistent with the pattern of variation in both types of time over days of the week. Both men and women provide more supervisory time for children on weekends than on weekdays, and women reduce the time they devote to active care on weekends (See Table 6). This may partly represent substitution among parents—men provide slightly more active care on weekends. However, the relative magnitudes suggest that the reduction in time children spend in school or in non-parental care on the weekend leads to a substitution away from care activities toward more relaxed supervisory care. A similar pattern is evident over months of the year—time devoted to primary care activities is lower in July than any other month; time that men, in particular, devote to supervisory care is highest in July. Gender differences are smallest in supervisory responsibility on weekends, with men’s time reaching 79% of women’s time (See Table 6). Gender differences are greatest with primary child care activities on weekdays: men’s time amounts to only 39% of women’s in this category.

3.6 Maternal work hours and child care time

Mothers in the United States tend to buffer the temporal effects of wage employment on their children. An additional hour of market work is associated with a relatively small reduction in primary child care activity time (Bianchi, 2000). Since there are only 24 h in a day, and allocation of time to different activities is simultaneously determined, no causal inferences can be drawn from such associations. We believe it is useful, however, to examine the relationships between hours of maternal employment and the different measures of child care discussed above.

Treating five categories of child care time on a diary day as dependent variables, we specify a Tobit model in which hours of maternal employment represent the primary independent variable of interest for the subsample of females living with a child under the age 5, but no child over 13. The control variables include dummies for full-time work, marital status, age, college education, number of children, age of youngest child, household income, spouse’s usual hours of work, weekday versus Saturday or Sunday, and race/ethnicity. The estimates presented in Table 8 show that maternal hours of paid employment have a greater negative association with supervisory responsibility than with hours of active care.

Table 7 Variations in child care activities and supervisory responsibility for children by day of week (married and cohabiting persons in a household with the youngest child under 6 but no child over 12), hours per day, 2003 ATUS
Full size table
Table 8 Tobit marginal effects estimates of the impact of maternal employment on measures of child care for married and cohabiting females in households with youngest child under 6 but no child over 12, 2003 ATUS (n = 1523)
Full size table

The Tobit estimates presented in Table 8 show that maternal hours of paid employment are negatively related to all categories of child care time except managerial/logistical time. The negative association with “in your care” time is by far the largest, followed by that with low-intensity care. The negative association with developmental care is the smallest of the significant coefficients. These results are consistent with the hypothesis that mothers protect the most “high quality” time with their children, and are perhaps unable to reduce logistical/managerial time. Driving children to day care, for instance, may represent an important part of the shift in their time allocation associated with employment. College education has a particularly large and significant positive effect on developmental care, with no significant effects on other categories of care. As mothers age, development care and physical care increases up until 35 and 31 years old respectively then decreases. Dummies for Black and Hispanic show negative effects.

The Tobit results put the effects of number of children and age of youngest child in statistical perspective. While number of children has a positive and significant effect on two categories of primary care, the coefficient is largest on low-intensity care activities. The coefficient is even larger on “in your care” time. Time devoted to activities of physical care for children declines more sharply and significantly with age than any other category. As suggested by the descriptive analysis, weekend days have a positive and significant impact on “in your care” time but a negative and significant effect on overall care activity time. Even controlling for weekend effects, care activity and “in your care” appear to be substitutes rather than complements: when “in your care” time is added to the list of independent variables in column 1 of Table 8, it is negatively associated with time devoted to care activities. Footnote 11

4 Conclusion

What is child care? Despite much progress in developing empirical measures, the surveys of the major English-speaking countries use rather different definitions for both primary child care activities and other measures. The distinction between primary and secondary child care is not as tidy as it initially seems, since some primary activities are relatively “passive” (such as looking after or minding children). Others, such as logistical/managerial activities (transporting children or dealing with doctors or teachers on behalf of children) may not involve much direct interaction with children.

The ATUS provides an opportunity to compare activities of care with the supervisory responsibilities captured by the “in your care” question. Analysis of overlaps between care activities, “in your care” and “who else was present” reveals intriguing differences in timing and intensity. Logistical /managerial tasks are often performed without children present. Adults engaging in eating, drinking, social and leisure activities, as well as religious and community activities, are particularly likely to report having children in their care. The age of youngest child and the number of children in a household affect time use in different ways. Overall, the distribution of time devoted to supervision of children is more equally balanced between men and women than the distribution of primary child care activity time, especially on weekends.

Tobit estimates reveal a strong negative relationship between hours of maternal employment and low-intensity care and supervisory care. The relationship between hours of maternal employment and logistical/managerial care and developmental care is weaker. Further detective work on these issues could shed further light on issues of comparability across surveys of the major English-speaking countries. It could also spur efforts to define child care in more careful and consistent terms, a necessary step in the development of more accurate estimates of costs to parents and outcomes for children.