This chartbook has been prepared for the Office of the Assistant Secretary for Planning and Evaluation at the U.S. Department of Health and Human Services and with the assistance of a panel of experts. Its purpose is to present examples of indicators of the social context of families that can be developed from currently available data, as well as to help identify critical gaps where such data are meager or do not yet exist. The chartbook does not seek to provide a comprehensive or exhaustive list of all available indicators, nor do these examples imply a judgment as to which are the most critical indicators to describe the family, instead, it is part of an exploratory effort to characterize families in their social context. This exploratory effort includes (1) synthesizing research on the multiple dimensions of the social context of families (see the Conceptual Framework in this volume); (2) identifying data sources and indicators to describe and monitor these dimensions (as summarized in this chartbook); and (3) identifying critical gaps in knowledge and data, as well as future directions for measuring and monitoring these dimensions (see discussion at end of this introduction, including references to the four papers prepared under this project).
The indicators in this chartbook expand the traditional set of indicators used to describe families, characterizing both the situation within families and how families relate to the community at large. A representative set of key indicators from the various social contexts of families are provided in this chartbook to illustrate the range of indicators as well as the value of the information that a broader effort might provide.
This chartbook differs in several ways from America’s Children and other recurring indicator volumes. This project is an exploratory effort, and an important goal was to uncover data gaps. Thus, consideration of indicators was not limited strictly to measures from nationally representative data sets, or data that were recent or recurring, but rather included other measures of interest for which the data may be less than perfect. In addition, the list of measures presented in this chartbook does not represent a committee consensus as to the best measures; rather it is an illustrative list of the types of indicators that would be important for the domains listed below.
The indicators were selected through a process that involved multiple steps, including a thorough review of research and data sources, development of a conceptual framework to guide the selection of indicators, and input from a panel of experts. Selection criteria were then applied, resulting in 110 potential indicators, of which 25 are presented in this chartbook. The list focused primarily on measures that were readily available, due the relatively small scale of this project. The steps in the process and the criteria applied are detailed below.
A review of the literature was conducted on the social context of families, including reviews of research in the domains of family structure, labor market participation, family functioning, volunteerism and civic/neighborhood involvement, youth development, religiosity, and social connections. A review of data sources was also conducted which identified sources in each of these areas, the periodicity of data collections, the availability of data for population groups of children, and family background characteristics. From these reviews of research and data, a conceptual framework document was developed which outlined and described the most salient research pertaining to developing indicators of the social context of families. The development of the conceptual framework and a preliminary list of potential indicators was the first step in exploring what would ideally be included in the final chartbook.
The second step was to select and assemble a panel of experts to provide additional expertise and a variety of viewpoints to help inform the decision of which indicators would be presented. The primary objective of this review was to gather a broad range of perspectives, rather than to reach complete consensus. The expert panel reviewed the conceptual framework and the recommended list of indicators (see Acknowledgements), as well as recommended additional topics of potential measurement. Based on the panel’s suggestions, additional potential indicators and data sources were identified and located. Subsequently, 25 indicators were selected based on the following criteria:
- Adequate coverage of each domain of the conceptual framework and maintaining a balance across the domains;
- Strength of the research on the indicator’s relationship to child and family well-being;
- Representation of both parent and child perspectives;
- Preference for family-based rather than individual-based indicators;
- Inclusion of both attitudes and behaviors;
- Variability in the indicator;
- Data quality and currency, with preferences for data collections using nationally representative samples, periodic versus one time collections, recent data, and for data sets allowing analysis by parental status;
- Policy interest or relevance;
- Importance to the expert panel; and
- Whether the indicator would make a unique contribution to portraying how families connect to each other and to the world around them.
The final list of indicators were organized into six broad areas:
Family Structure. Indicators in this area include a traditional measure of living arrangements, as well as more complex measures capturing an array of familial relationships:
- Children’s living arrangements, Family structure change, Families with grandparents who live nearby, Births to unmarried teens
Family Functioning. Specific measures examine amount of family time together and quality of relationships:
- Parental warmth and affection, Positive parent-adolescent relationships, Parental awareness of adolescents friends and activities, Time spent with parents, Contact with nonresident parents
Family, Work, and Child Care. This broad area includes traditional measures of employment status and hours of work for both parents, as well as measures pertaining to the impact of job stress:
- Parental employment by family structure, Work-family stress, Family income, Patterns of child care
School Involvement and Civic Engagement. These measures include parental and student engagement with child's school, and family and student civic engagement:
- Parental involvement in school, Volunteering as a family, Student participation in community service, Parental voting, Youth connection to school peers, School supportiveness
Religiosity. Indicators of religiosity include a measure of participation in religious services, as well as a measure of participation in a broader group of religious activities as a family.
- Parental religious service attendance, Adolescent participation in religious activities with their families
Social Connections. These measures describe the extent to which families have a sense of community in their neighborhoods and among friends:
- Neighborhood community, Community of friends, Concern for safety, Residential mobility
Each indicator includes a figure that highlights the data for the total population as well as for one subgroup. Subgroups were chosen based upon the availability of subgroup data, the salience of the subgroup to the indicator, and upon a review of the data so that interesting differences across population groups were highlighted. A data table accompanies each indicator, typically presenting several subgroups. The indicator text describes patterns in the data, and all differences mentioned are statistically significant, except where noted.
The data that have been chosen for each indicator were carefully selected for quality and currency. However, it is not possible to present each indicator systematically for the same years, or for the same subgroups, since the availability of the data varies by data set. This indicator volume is intended only to represent examples of indicators that are possible given currently available data, rather than a complete and comparable set of indicators.
As noted above, the data presented in this chartbook come from many different sources. These include both well-known data sets such as the Current Population Survey (CPS) and the Panel Study of Income Dynamics (PSID) as well as lesser-known sources such as Social Capital Community Benchmark Survey. The data sources used for each indicator are listed below and described in more detail in Appendix A:
- Current Population Survey
- Giving and Volunteering in the United States
- National Household Education Survey Programs
- National Longitudinal Study of Adolescent Health
- National Longitudinal Survey of Youth-1997
- National Study of the Changing Workforce
- National Survey of America's Families
- National Survey of Families and Households
- Panel Study of Income Dynamics
- National Vital Statistics System
- Social Capital Community Benchmark Survey
In addition, there are many other data sets that have the potential to inform our view of the social context of families that are not represented here due to the limited scale of the project. These include, among others:
- National Survey of Family Growth
- Survey of Income and Program Participation
- National Health Interview Survey
- American Community Survey
- Early Childhood Longitudinal Survey-Kindergarten and Birth cohorts
- Fragile Families Survey
- Monitoring the Future
- General Social Survey
- National Compensation Survey
- The Civic and Political Health of the Nation
- National Family Violence Survey
- National Crime Victimization Survey
While the indicators in this chartbook make important strides in describing the social context of families, gaps remain in our ability to measure and report on the domains listed above. In some cases currently available data are insufficient to measure an important concept. In other cases data may be available but additional conceptual work is needed to define an appropriate measure. In addition, for some important constructs such as family structure, measures are widely available but defined inconsistently across data sets. There are also gaps that reach across all of the areas of investigation, such as our ability to present data on trends over a consistent time period, across a consistent set of population groups and across cultures, or by stage of family development. The following section identifies gaps within each domain by comparing the critical measurement areas discussed in the conceptual framework or suggested by the expert panel with available data and measures. Additional gaps became apparent while working with the data for this chartbook.
Examples of Gaps Within the Domain of Family Structure
- A basic indicator to accurately portray the complexity of family composition in America today is lacking. There is no current source of data available that adequately combines information on whether the parents are married or cohabitating, whether they are biological, step, or adoptive parents of the children in their household, and whether other relatives are living with the family in the same household. The National Health Interview Survey is developing such a cross-sectional measure. Furthermore, additional measures are needed to reflect the complexity of not only measuring trends over time, but also tracking families longitudinally. Some currently available data sets have some of these pieces, for some years, but no one data set can yet present this complete portrait on a regular basis over time.
- It is not currently possible to define family structure consistently across data sets that address the social context of families. This makes it impossible to accurately compare family types across indicators.
- Many indicators in this report rely on data for which it is difficult, and in some cases impossible, to analyze with families or parents as the unit of analysis. Surveys often use the household head as the respondent and reference person for household relationships, but this procedure does not always accurately identify whether other members of the household are parents of children in the household. Therefore, special analyses were needed requiring different assumptions across data sets in order to create estimates for parents. In some data sets, such as the Social Capital Community Benchmark Survey, it was not possible to portray parents at all, but only adults.
- We lack both data and measures to fully reflect the process of couple formation and related changes in patterns of courtship and dating, attitudes toward marriage and cohabitation, sexual relationships prior to marriage, and barriers to marriage. (See the paper written for this project by Steven Nock, The New Chronology of Union Formation: Strategies for Measuring Changing Pathways for a strategy to develop such measures). Similarly, the indicators in this volume do not address the growing proportion of families that are stepfamilies resulting from remarriages.
- Family structure transitions are known to be stressful on families, yet the measure available in the Panel Survey of Income Dynamics which is included in this chartbook, combines divorce and remarriage along with births and adoptions over a two-year period in a family's life. A more finely-tuned measure of transitions is needed that would separate entrances and departures from unions, such as marriage and divorce, and the entrances of new family members (including births/adoptions and immigration) as well as departures from the family (including deaths, and those leaving home) over a longer period of time in the life of a family.
- Data that capture marriage and divorce events at the subnational level are currently not readily available. Counting Couples, a forum sponsored by the Federal Inter-Agency Forum on Child and Family Statistics has identified several targets of opportunity for improving these data, but significant changes are several years off.
Examples of Gaps Within the Domain of Family Functioning
- Marital quality is key to healthy family functioning, yet measures of marital quality are just now being developed and have not yet been fielded in national surveys.
- Available measures of family conflict, including punishment, child abuse and domestic violence lack rigor and currency. Better measures need to be developed and fielded in such a way that biases are minimized in order to adequately monitor this critical area of family functioning
- There is anecdotal evidence that parental stress is increasing, and that mothers are particularly stressed. New ways of incorporating biological measures of stress within surveys are being explored, and could be extremely useful for the study of stress among parents in various social contexts in the future.
- Research demonstrates that children who are exposed to parental risk behaviors such as smoking, drug and alcohol abuse are at higher risk of developing these habits themselves. Creating indicators of parental risk behaviors in the home for children from existing data can be done, but it requires complex and time-consuming analyses.
- An index of turbulence in residence, school, and family structure would be an important contribution to this study, yet it is not possible to create from one existing data source.
- Family routines, rituals, and time together are key components of family functioning (see the paper written for this project by Lina Guzman and Susan Jekeliek, Family Time) but there are few such measures fielded in national surveys. Furthermore, measures are needed at the family level rather than at the individual level in order to capture interactions between family members.
- Although the chartbook contains indicators on parenting characteristics that have been related to positive outcomes for children, such as warmth and awareness, cultural variation in effective parenting is not captured in currently available measures.
- While a general measure of parent-child communication quality is available, measures of specific types of communication are needed.
- Parents provide gate keeping, resource management, and networking functions for their family, yet these functions are not captured in available national surveys.
Examples of Gaps Within the Domain of Family, Work and Child Care
- While an attitudinal measure of work-family stress is included in this volume, consistent trend data on the number of hours spent at work and the corresponding effect on the number of hours spent with family, for both mothers and fathers, are not available. In addition, commuting times to work are increasing for adults, but this information is not available by parental status, which limits the ability to analyze the extent to which commuting infringes upon family time.
- More measures are needed of the various ways in which parents arrange their work and child care schedules, and the extent to which the diverse arrangements made by parents reflect parental preferences or economic necessity. For example, while there are data showing the widespread use of care by relatives, there are not good data on the extent to which use of this care is influenced by availability of relatives outside the household, cultural values, personal preferences, high costs of formal care, lack of access to subsidies, or other factors. Similar questions can be asked about the use of multiple arrangements, after-school care, work during non-traditional hours, part-time work, etc.
- More generally, indicators of parental satisfaction with child care arrangements have been developed for small scale studies, and have been incorporated in the National Household Education Survey of 2001, but they fail to correct for parental biases toward their current child care provider, so that a true national assessment of parental satisfaction with child care remains elusive.
- There are non-economic costs associated with nonparental child care that may impact family strengths, such as time family members spend together, for which indicators need to be developed.
Examples of Gaps Within the Domain of School Involvement and Civic Engagement
- Detailed data on youth civic engagement are not available after 1999, the last time that the Youth Supplement was administered of the National Household Education Survey. Data collection in this area is needed in the future in order to monitor trends over time. New studies on civic engagement that incorporate promising and broader measures of civic engagement, such as The Civic and Political Health of the Nation: A Generational Portrait, can only be analyzed at the individual level, and parents and youth are not identifiable separately, nor can families be analyzed as a unit.
- More specific data are needed on families volunteering together, including the number and ages of family members involved, and whether the volunteering is initiated by the family or by an organization to which they belong, such as a school, church or community service organization. The stages in a family life cycle during which families are likely to volunteer is also important to know. For these reasons, data need to be collected on volunteering with families as the unit of analysis.
- Family structure variables differ between the November Current Population Survey, the data source used for the voting behavior indicator, and the March Current Population Survey, which is typically used to portray family structure. Thus, it is not possible to portray voting behavior with the same family structure definitions across months of the same survey.
Examples of Gaps Within the Domain of Religiosity
- There is no current source of trend data on youth participation in religious-oriented youth groups. Monitoring the Future used to ask the question, but stopped including the question after 1996.
- Data and measures are needed on the prevalence of couples that do not share religious affiliation, and the affiliation of their children. Indicators are needed on how family religiosity changes over the life cycle of the family, but data are rarely collected this way.
- Current measures of religiosity are largely limited to attendance at religious services, affiliation, and importance of religion. A much more diverse set of measures is needed to accurately portray current family religious practices (see the paper written for this project, The Measurement of Family Religiosity and Spirituality, by Laura Lippman, Erik Michelsen, and Eugene Roehlkepartain).
Examples of Gaps Within the Domain of Social Connections
- For three of the four key indicators within the domain of social connections, including neighborhood community, community of friends, and concern for safety, data are only available for individuals. Data are not available for parents, youth, or families.
- Although a measure of residential mobility in the last year is included in the chartbook, data sources do not allow analyses of mobility over a longer time period.
- An indicator of residential segregation by socioeconomic status needs to be developed.
- Better measures of social networks and community resources need to be fielded in national surveys, including those that are valid for various cultural and immigrant groups.
Across all of the areas of study, it is possible to develop some indicators of trends, either from published data or by conducting new analyses. Included in this chartbook is a table that identifies the availability of trend data for each indicator. For the majority of the indicators, some trend measures could be developed through further analyses, though trends could not be monitored over a consistent time period across indicators. For a few indicators, trend data are not currently available at all.
Just as important, but even less available than trend data, is detail for each indicator by the family life cycle stage, as pictured in Chart B of the Conceptual Framework. In order to understand how and when families interact with their environments and how these interactions affect children in families, the age of children in the family needs to be known. It is also important to track changes over the life course of a family by developing longitudinal measures of key constructs that are already measured in cross-sectional surveys.
A number of important issues and potential avenues for further development are discussed in a series of papers written by noted researchers in the field of family indicators. These papers are available in a separate volume and include: