Presenters:
Sandra Hofferth, University of Michigan
Jacqueline Eccles, University of Michigan
David Almeida, University of Arizona

Sandra Hofferth: Time Diary Measures in the PSID Child Development Supplement (CDS)

The Panel Study of Income Dynamics (PSID) began in 1968 as a representative sample of 5,000 American families. It follows all members when they leave home and form their own families. Therefore in 1997 it interviewed 7,000 families. Its purpose is to study income dynamics, but it lacks information on children per se.

The CDS’ objectives are to provide a nationally representative data base on children and families and to support studies of the effects of parental and community resources (time, money, social capital) on child health and development.

Various methodologies are used. Methodology 1 identifies families with children from birth to age 12. The families are called and materials, including the time diary, are mailed to them ahead of time. The child’s cognitive development is assessed in the home (Woodcock-Johnson, WISC). Interviews are conducted with the child’s primary care-giver about each child (up to two children). Diaries and SAQs are collected.

Methodology 2 obtains SAQs from the child’s other parent. Permission forms are obtained from their teacher/care-giver and administrator. Booklets are mailed. Phone numbers of the absent parent are obtained, and the interview is conducted.

In the Time Diary Methodology, diaries are mailed ahead. The interviewer collects two diaries per child within three days of the designated weekday and within one week of the designated weekend day. The interviewer reviews the diary with the respondent or administers it in the home to the respondent/child. Then the interviewer edits the diary, and data coding and entry are done.

The focus of time diaries in the CDS includes:

1. The type and extent of the interaction. Interaction with others (e.g., parents, siblings) distinguishes accessibility from interaction and involvement (M. Lamb et al.). Responsibility is a core question for parents. It gets at parental time with children.
2. Educational activities. A look at the school diary in teachers’ booklets covers classes and types of curriculum and instruction. Also a look at the home diary covers reading, studying, TV watching and computer use, specifically programs, videos, and computer games.
3. Detailed questions about each activity such as the type of activity, time spent on that activity, detail of the program, where the child was, who did the activity with the child, who was present but not involved, and what else the child was doing at that time.

There are several research issues still remaining. The national picture of children's time use needs to include both cross-time and cross-national comparisons. Causes of children's differential time use include differences in time use by number of parents, employment of the mother, poverty and the receipt of public assistance. Also there are differences in time use between the activities of fathers and mothers with their children. Consequences of differential time use include: time spent reading, studying, and watching TV which relate to test scores and grades; time spent with father and children's behavior; supervised time and its relation to children's grades and test scores; and the consequences of differential amounts of time spent on tasks in the classroom for test scores.

(Sources: F.T. Juster and F.P. Stafford. 1985. Time, Goods, and Well-Being. Ann Arbor, MI: Institute for Social Research.

M. Lamb, J. Pleck, E. Charnov, J. Levine. 1985. Paternal Behavior in Humans. American Zoologist 25:883-894)

Jacqueline Eccles began by talking about how she, as a developmental psychologist, thinks about time use, which is different from how economists view it. There are several reasons to collect time-use data. She has been involved in three types.

1. To generate estimates at the population level of the distribution of time use for various sub-groups (e.g., amount of time spent watching TV by different age groups or different national groups)
2. To assess individual differences in exposure to various types of activity settings in order to test specific hypotheses about the impact of such exposure on the course of development (e.g., do children who read more at home do better in school?)
3. To assess variations in the types of experiences provided in various activity settings and their influence on the course of development (e.g., does the quality time parents spend with their children influence positive developmental outcomes such as social adjustment and school achievement?)

She has been most interested in 1 and 2 above. In addition, her group has been interested in predicting individual differences in time-use patterns in order to test hypotheses about the determinants of individual behavioral choices and behavioral patterns.

She got her introduction to time use from the National Time Use Study. From this, she learned about some pros and cons of collecting time-use data. She summarized the main non-intuitive findings:

The more time kids spent on homework, the lower their grades;
The more time parents spent helping their kids with homework, the lower their grades.

The explanation for this is that parents are reactive, so those students with more problems are getting more homework and more help from their parents.

Also, she looked at the quality of parental time and how this related to maternal employment. They asked kids how much time they spend with mom or their primary care-giver. They aggregated the data to see if mom’s work status had any effect on time with children, but children were not aware of losing time with their working moms.

There is a clear link between TV watching and school performance after grade three, but not before. This is not taking time away from homework because there is so little homework given in this country.

Next she talked about other ways to study adolescent time use. Gender differences peak during adolescence. The main question is, does it matter how adolescents spend their time out of school?

Why would time matter? Because it is an opportunity to acquire skills and attitudes, to narrow or expand pathway options, to provide settings for engagement (more than flow). It provides opportunities for socio-emotional and identity development, structures immediate social networks and peer groups, and provides opportunities and risks for future social capital.

There is evidence that it does indeed matter. Using data from MSALT(Michigan Study of Adolescent Life Transitions), which asks children to report, using a checklist, all the things that they do; all their activities. This is measured in the 10th grade and used to predict outcomes in the 12th grade. (Children responded yes/no to each item, rather than by amount.)

They aggregated the list into four categories of activities:

Pro-social; involvement in church, etc.
Performance type
Team sports
School activities.

The findings show that gender and sex differences fall as you would think. The lack of pro-social activities are linked by 10th grade with problem behaviors, and even more so by 12th grade (i.e., the more pro-social activities, the less risky behaviors).

Predicting from 10th to 12th and controlling for 10th-grade level, there is more skipping school, getting drunk, and using alcohol and drugs for children not engaged in pro-social behaviors.

When looking at effects of engagement in sports, both good and bad effects are found. There is a clear link to engagement in problem behaviors such as alcohol use for males already by grade 10, but for girls, it takes until 12th grade.

Another positive finding is that GPA (grade point average) is positively affected by engagement in all four of the above categories, according to 11th-grade GPA scores. The likelihood of attending college at age 21, based on 10th-grade pro-social activities, is not significant. However, performance-type activities, school involvement and team sports are all related to an increase in college attendance.

Question: Do the number or range of activities matter?

Answer: The range seems to be more critical. We see it most in sports. With one sport vs. none, we will see a gain. More than one sport does not have an increased effect. The breadth of activities gives a linear effect. The more involvement in different activities, the greater the benefit in terms of GPA and college attendance.

David Almeida: Unfortunately, a computer failure prevented note-taking during Almeida’s session. The except below is from his paper, "National Study of Daily Experiences: Examining Quantity and Quality of Child-Related Experiences Through Daily Telephone Diaries."

The use of innovative research tools such as time diaries and event sampling methodology has permitted researchers to obtain detailed accounts of how people spend their time. The present daily telephone interview study aims to add to this literature by examining stable and time-varying correlates of how much time parents spend with their children and the types of interactions they have with them on a day-to-day basis.

The results provide the first steps toward understanding the role children play in the day-to-day well-being of parents. By mapping these child-related experiences, we can begin to understand the types of interactions parents have with their children and who is most likely to experience them. Consistent with other time-use research, women spend more time with their children and have more frequent and nurturing interactions with them than do men (Juster & Stafford, 1991; Robinson & Godbey, 1997; Stains & Pleck, 1983). The pattern of results also suggests that parental and familial characteristics are more predictive of the quantity of child interactions than the quality of these interactions. Apart from gender, sociodemographic variables were not associated with the frequency of emotional support and parent-tensions. Perhaps these interactions are more related to situational characteristics of the child. A few notable associations emerged for frequency and severity of child stressors. Parents with higher levels of education had more frequent child stressors, but those stressors were less severe. Having an adolescent in the household was related to fewer child stressors. However, when a parent-adolescent tension occurred, it tended to be of a more severe nature than tensions occurring with younger children.

Directions for future research include linking these child experiences to daily parental well-being. Are child tensions and stressors predictive of daily psychological distress and physical health symptoms? As researchers continue to recognize how major life transitions of children affect parents, we should also keep our eyes open to the everyday experiences of parenthood.

Discussant: P. Lindsay Chase-Lansdale

There is both danger and opportunity here. We are in danger if we do not define the purpose of our time-use surveys. She sees two broad purposes:

1. Painting a landscape of children’s lives. What are they doing? Who is around? This adds a perspective of the children. What are the key people in the child’s life doing if not with the child? We need a picture of the family landscape. This is relative to welfare reform, especially since many entry-level jobs are at odd hours.
2. What contributes to healthy child development? We do not just need a description, but an explanation. We know that the quality of relationships matter. Parenting is related to all sorts of outcomes. Also we know that there is not really a simple relationship between parental time and child outcomes, partly because of poor measurement. We need to focus more on structure, meaning and quality of time, and see how these link with the landscape of time. For example, we have not measured time well in developmental studies. Instead we use proxy measures for time such as full-time working, lots of time away from child. Capturing time well has been difficult for developmentalists.

Some examples:

• Maternal employment and young children’s attachment. In the NICHD child care study, from birth to 8 years old, shows that the number of hours in infant child care does not seem to be causing a problem. However the quality of mothering of infants during that time mediates, so if the family is poor, time in child care does matter.
• Adolescent time use: They spend little time with parents. We could conclude that that is why there are problem behaviors. But it is not simply if they are in activities or not. Those just hanging out are not at risk if they check in with parents and parents know where they are. Monitoring and supervision are key.
• Employment in low-income families: Many are engaged in shift work to balance work and child care without using outside child care. This misses that parents are overlapping and are very tired. The quality of interactions with the child goes down, affecting the child.
• Volunteer work: In regards to Eccles work, what is family life like for these children in different activities. Work done by Joe Allen finds that volunteer work is not all the same. How it is structured matters. A classroom setting for discussions with mentors is very important.

We have a need to look at both quality and quantity. We need to define the issues and not put all our eggs in either the quantity or the quality basket.

Unfortunately, another break in the note-taking occurred limiting the amount of discussion and comments recorded below.

Group Discussion and Comments:

Smeeding: What did it cost for the supplements?

Hofferth: It was about $800 direct cost per child for approximately 2500 families.

Almeida: It was $280,000 dollars for the add-on costs for his study.

Greg Duncan, Northwestern University: We heard about three studies and three methodologies. Why did each of you do what you did?

Almeida: They had money, but time diary was too intensive for their purposes. Their main focus was on daily stress, but it is important to get some estimates of time use. They asked questions on a daily basis, so the recall time was not too bad. The vast differences are explained because they are less restrictive.

Hofferth: They should add primary and secondary time use to their study. Then they would get closer numbers.

Almeida: They wanted to ease the response burden.

Eccles: They are very interested in individual differences and longitudinal, so they did not have time. Students filled in the diaries themselves, and they were interested in reliability at the individual level. Information that aggregated across time for each student needed a bigger window of time.

Session ended at the scheduled time.

back to top