Indicators of Child, Family, and Community Connections, Companion Volume of Related Papers. II. Existing Measures and Limitations


In recent years, there has been an increase in the availability of data sources that measure family time, including the 1997 Child Development Supplement to the Panel Study of Income Dynamics (PSID-CDS), the National Survey of Families and Households, and the American’s Use of Time Project. Examples of smaller scale studies include the Annual Meaningful Time Check-Up and California’s Children Activity Pattern Survey. Concomitant with this rise has been an increase in the number of studies that focus on family time (see Bianchi, 2000; Bianchi & Robinson, 1997; Bryant & Zick, 1996; Budig & Folbre, 2002; S. Hofferth & Sandberg, 2001; Yeung, Sandberg, Davis-Kean, &Hofferth, 2001). The field has been further aided by qualitative studies that delve into the quality, meaning, and nature of family time and interactions (Blum-Kulka, 1994). The Changing Workforce Study (Bond, Galinsky, &Swanberg, 1998) tracks the subjective experience of employment and family demands on workers, and is unique in collecting this information over time. Both the American Community Survey and the Current Population Survey collect data on commute time to work; however, neither of these surveys has made estimates of commute time for parents publicly available.

Together these surveys, and subsequent studies based on these data, have provided national estimates of the amount of time that families, parents and children in particular, spend together. However, outside of a few key activities, less is known about how families spend their time together or the nature of the family interactions. Moreover, while important advances have been made in the measurement of family time and in data collection methods, most notably in area of time use diaries, the field as a whole is still “underdeveloped (Smeeding & Marchand, 2003). Future studies designed to measure family time will have to balance the need to improve the reliability and accuracy of estimates (Robinson, 1997) with the costs associated with collecting data from multiple perspectives (i.e., parents versus children) and for a wider range of activities.

Review of studies and data on family time

Despite the entrance of mothers into the labor force and concerns about longer commute times and work days, analyses of available time use data indicate that the amount of family time, in particular, mother-child time, has remained stable since the 1920’s (Bianchi, 2000; Bryant & Zick, 1996). This stability is partly due to the fact that in the earlier part of the century large segments of the population were rural and engaged in agriculture—work that left little time for direct parent-child interaction (Bryant, 1996). Moreover, the gap between employed and non-employed mothers in mother-child time is relatively small. Including time spent directly engaged with children, as well as time spent in their presence, children of employed mothers spend approximately five and half hours less per week with their mothers than children of non-employed mothers (Sandberg & Hofferth, 1999). The gap in time spent with children between employed and non-employed mothers is minimized because non-employed mothers tend to allocate their “non-labor hours to household tasks, volunteer and leisure activities, and not to direct child care (see (Bianchi, 2000). In addition, employed mothers appear to take steps to protect their time with their children by minimizing time spent in household tasks, personal leisure activities and sleep (Bianchi, 2000).

The stability of parent-child time has also been made possible by the increased involvement of married fathers in childcare, in particular among families with employed mothers (Bianchi 2000; Sandberg & Hofferth 1999, 2001). For example, the amount of time fathers spent in direct child care more than doubled between 1965 and 1998 (Bianchi, 2000). Further, Bryant and Zick (1996) find that as mother’s employment hours increase, the amount of time parents (i.e., mothers and fathers) spend with their children doing household tasks and engaging in leisure activities together increases. Last, reduction in family size has resulted in small increases in parental time per child from the perspective of children (Bianchi, 2000).

While trend data and population estimates of parent-child time are useful, they often mask variations by family-structure, family stage, gender, socioeconomic status, and race/ethnicity. In general, children in single-parent families spend less time with their parents than children in two-parent families (Child Trends, 2002). Data collected from the PSID-CDS indicate that children in two-parent families spend approximately 2.21 hours with their mothers and 1.45 hours with their fathers on a daily basis compared to 1.16 hours with mothers and .25 hours with fathers among children in single-parent families (Child Trends 2002). In addition, the data also suggest differences in how parents and children distribute their time across activities by family structure. For example, children from single-mother families spend less time in educational activities but more time in organized sports (S. Hofferth & Sandberg, 2001). Other studies, however, indicate that family structure differences in the types of activities that children engage in are small and non-significant (Bianchi & Robinson, 1997). Additionally, recent studies suggest that children in stable single-parent families may fare better than children whose parents remarry since a new spouse often results in a reallocation of parent’s time (Thomson, Mosley, Hanson, & McLanahan, 2001).

The amount of time that children spend with their parents also varies by developmental stage. Young children, who require extensive direct care, spend more time with their parents than older children who require less direct supervision and who as a result of school, homework, and part-time jobs have less discretionary time (S. Hofferth & Sandberg, 2001; Robinson & Bianchi, 1997). Bryant and Zick (1996) find that mothers share a wider range of household and family maintenance tasks with their children, but the sharing of tasks is somewhat gendered. For example, mothers are more likely to share preparing meals and cleaning with their daughters than with their sons. The authors argue that by engaging in household tasks together, mothers are transmitting and teaching nurturing behavior (Bryant & Zick 1996).

Important differences by race/ethnicity and socioeconomic status have also been found in both the amount of family time and the types of activities in which families engage. Hofferth and Sandberg, for example, find that black children spend more time in church, Asian children spend more time studying, and Hispanic children spend more time engaged in eating meals together and in household tasks.

The body of research reviewed thus far points to several important aspects for the measurement of family time. First, the studies reviewed indicate that, in addition to measuring the amount of family time, it is also important to measure the kinds of activities that family members engage in. For example, Bryant and Zick’s research suggests that, while the amount of family time does not vary greatly by maternal employment status, the type and distribution of family activities do. In addition, it may be that what family members are doing together is as important for family and child well-being as the amount of time families spend together. Second, these studies suggest that it is important to collect data from multiple family members and from multiple perspectives. For example, measuring only mother-child time underestimates the amount of time that children spend with parents and ignores the important contributions that men make to family life. Third, the research of Hofferth and others speaks to the importance of including items that capture the meaning of and the impetus behind family activities. For example, the salience of the family unit among Hispanics may result in Hispanic children and parents sharing household tasks to a greater extent than other racial groups (see (Taylor, 1994).

Strengths and limitations of existing measures

Time use diaries: One of the most common sources of data on family time has been time use diaries. Time use diaries ask respondents to report how much time they have spent in last 24 hours in various activities, such as in child care, reading or playing, talking, eating meals together, etc. Time use diaries, while popular, have several limitations. Most notably are the costs associated with collecting extensive and detailed information (S. L. Hofferth & Sandberg 2001). Researchers have attempted to curtail costs by asking about activities for a randomly selected weekday or weekend(1) rather than for extended periods of time (e.g., a whole week or month).

Also of concern are issues surrounding recall (Bianchi & Robinson, 1997; Hill & Stafford, 1985; Robinson, 1985). Attempts to improve recall include providing respondents with beepers or timers that alert respondents to record the activities they are currently engaging in. Nevertheless, tasks and activities that occur less frequently have produced less reliable estimates (Hill & Stafford, 1985; Robinson, 1985). Consequently, studies that use time diaries for estimates of the amount of family time often limit their estimates to frequently occurring activities (see Hofferth & Sandberg 2001a). Unfortunately, while this step helps to improve the reliability of estimates, the diversity of family activities that are captured is limited. This may be especially problematic when studying family time among minority groups and non-middle class families whose activities may differ from white middle class families.

Also problematic, although more easily corrected, is the way in which time is allocated to tasks (see (Folbre, 2001). For example, most time diaries ask respondents to report the primary task in which they were engaged during that time. Thus, secondary and tertiary tasks are underestimated. This is particularly problematic for the study of parent-child time since child-related tasks "tend to spill over into many indirect as well as direct responsibilities" (Folbre, 2002, p.5). Moreover, only time that is spent directly interacting with children is counted as parent-child time (also known as direct-time). For example, time that fathers spend doing their children's laundry is counted as time spent on laundry. Similarly, time spent doing laundry while supervising their children's homework is not counted as time spent with children. Recent estimates of direct parent-child time are just under two hours for mother-child and one hour for father-child (Bianchi, 2000).

Clearly, measures of direct parent-child time paint an incomplete picture of parent-child time. Fortunately, recent time use diaries have collected data on secondary and tertiary activities; thereby, providing measures of indirect parent-child time. When secondary activities are included, estimates increase to 2.8 hours per day for mothers and 1.3 for fathers (Bianchi, 2000). Including all time spent with children in any activity, parent-child time increased to 5.5 hours per day for mothers and 3.8 hours for fathers (Bianchi, 2000). Including time spent in an activity, as well as time that parents are accessible to children, though not directly engaged in an activity with them, children are estimated to spend 29 hours with their mothers and 19 hours with their fathers on a weekly basis (Sandberg & Hofferth 2001). The importance of including indirect measures of family time may be particularly relevant for working and single-parents for whom time is more limited and, as noted above, for certain minority group members who are more likely to undertake households tasks together.

Moreover, measures that are based on direct and indirect family time miss the importance of parental accessibility (Bianchi, 2000; Budig & Folbre, 2002). Parental presence regardless of whether parents and children are engaging in activities together may be important for child well-being.

Finally, because time use diaries record data at the individual or dyad level, estimates of time spent together as a family unit are largely missing. That is, less is known about the amount of time that all family members spend talking together, engaging in leisure activities, or interacting as a family unit than is known about parent-child or couple time.

The Home Observation for the Measurement of the Environment (HOME): The HOME scale was developed as a tool for assessing children's home environments (Caldwell & Bradley, 1984). The intent of the scale is to identify children whose home environments place them at risk for unhealthy development. The scale is made up of both parental report items and interviewer observations, and is designed to be child-specific and age appropriate (e.g., items vary by age of the child). The scale has been used in a variety of ways from following the development of low-weight infants over time (Bradley, Caldwell, Rock, &Casey, 1987; Bradley et al., 1994), to researching the influence of work circumstances on children's home environments and their outcomes (Cooksey, Menaghan, &Jekielek, 1997; Parcel & Meneghan, 1994), to assessing the well-being of children who live in poverty (Moore, Glei, Driscoll, Zaslow, &Redd, 2002). Many large-scale studies include measures of the home environment, such as the National Longitudinal Survey of Youth 1979, the Panel Study of Income Dynamics, and the Early Childhood Longitudinal Study.

While the HOME scale was not developed to measure time use specifically, it does contain items that might be helpful in this regard. Several of the items tap into aspects of children's time use, such as how often parents read to their children, how often children perform chores, how often parents take their child to the grocery store or how often the parents take their child to a museum; the response options are typically in Likert-scale format, and do not capture actual time spent in each activity. Other items tap into the quality of interactions: Affective items tap into the quality of parent-child interactions by reflecting pleasant conversations, parental responsiveness to children's questions, kissing, hugging and caressing of the child by the parent (as observed by the interviewer). Questions about cognitive stimulation often probe the frequency of activities with a child, such as reading with the child, or helping the child with tasks, going to museums, etc. Questions from the scale also probe aspects of overall family time spent together, and not just parent-child time, although these questions are very limited. They tap into the frequency with which the child gets together with other family and friends, and the frequency with which the child eats dinner together with the family.

The HOME scale is useful but also too limited to address current critiques of family time use measures. While some items provide a sense of quality and others of quantity of time spent together, the items do not offer a way to judge whether quality time is spent within specific activities. In addition, although quantity of time in activities is reported, it is difficult to get an idea of the make-up of the family day: this would be a limitation of most any measure that is not collected with a time use diary. The items in this scale do not provide information on the interactions that other family members have with each other. While this scale has proven an important tool for understanding child development, it is limited in addressing current issues related to family time use.

Other measures: The most common survey research method for collecting time use information is to directly ask parents how much time they spend in specific activities with their children. While seemingly more simple than time use diaries, or more accurate in terms of actual time than items from measures like the HOME scale, this method is known to be biased: Parents will report more time spent on desirable activities than on less desirable ones, the reports are difficult to validate, and some research shows that times have been inaccurately reported with this method (S. L. Hofferth & Sandberg 2001).

Routines and rituals: We have suggested that current measures of time use are better at capturing activities between parents and children than they are at capturing interactions at the family level. We view family-level activities as important from the standpoint that one's experience in their family of origin has implications throughout the lifecourse, influencing the manner in which offspring form and interact with their own families. Additionally, Howe (2002) asserts that family-level activities shape a unique family identity, a sense of commitment to others in one's family that fosters networks of social support. Toward this end, there has been a resurgence in attention to the notion of family routines and family rituals as separate but important constructs.

The term 'routine' is used to identify habits of daily living - instrumental, brief, observable behavior that is repeated over time (Fiese et al., 2002). To a limited degree, data on some key family 'routines' are already collected (eating dinner together, going shopping together). Rituals are contrasted from routines as having an affective component, a symbolic meaning that conveys a sense of family identity that leaves an individual with a sense of belonging (Fiese et al., 2002, page 382). Distinctions between the two have rarely been made in empirical studies.

Fiese and Kline (1993) developed an instrument that actually distinguishes between the two concepts (Fiese et al., 2002, page 384). Their questions on routines go beyond typical measures that capture 'frequency of eating dinner together.' For example, they probe whether family members are expected to contribute to the chore of getting the dinner on the table. They identify rituals by collecting responses dealing with the "affect, symbolic significance, and commitment to continue the activity across generations" related to routines. These appear to be open-ended questions. They find that the distinction makes a difference in their research - family rituals were associated with adolescent identity and marital satisfaction, whereas family routines were not (Fiese, Hooker, Kotary, &Schwagler, 1993). While these instruments help identify family routines that are meaningful family activities, they have only been used on small samples. It would likely be costly to code responses on the meaningfulness of family routines in large scale surveys. Alternatively, Larson suggests that one way to identify meaningful family interactions (or in his conceptualization, activities that involve an emotional component) is to include a survey question probing the most significant and least significant activities of the day (Larson, 1997).


(1) It is important to collect information for both weekday and weekend activities because the type of and amount of time spent in activities varies by weekday and weekend.

View full report


"papers.pdf" (pdf, 639.68Kb)

Note: Documents in PDF format require the Adobe Acrobat Reader®. If you experience problems with PDF documents, please download the latest version of the Reader®