Number A-21 in Series, "New Federalism: Issues and Options for States"
The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders.
Social indicators for tracking the well-being of children and their families are important tools of government. Uses include simple description; monitoring to inform planning and resource allocation; goal-setting to guide broad policy and coordinate efforts across agencies and between levels of government; accountability efforts to hold agencies and even whole levels of government responsible for making progress toward specific social goals; and, under limited circumstances, evaluation of comprehensive government initiatives.
1
As the new federalism has proceeded through the 1990s, states have assumed increasing responsibility for the design and administration of social policy and for the ultimate success or failure of the resulting programs. They have also been given greater flexibility in determining which social goals to focus on and how to best meet them. States vary in the degree to which these new powers are further devolved to the local level, with some states giving a great deal of flexibility to local governments.2 As a result, both the need for social indicator data at the state and local levels and the need for expertise to use the data effectively have risen substantially under the new federalism. An increasing number of states are adopting goals-driven benchmarking programs and accountability-based budgeting policies that rely heavily on social indicator data. Indeed, the ultimate success or failure of the new federalism paradigm may depend in part on states' abilities to develop better sources of social indicator data and use them in an informed and effective manner.
The federal government has long played a dominant role in producing social indicator data at the state and local levels through such vehicles as the decennial census, the vital statistics system, various disease surveillance systems, and mandatory reporting of administrative data for selected social programs. Social indicator data on children are available for a variety of domains including education, health, economic security, social development, and family structure, though fewer data are available at the state and local levels than at the national level.3
In response to the evolving needs of states and to the changing federal-state relationship as embodied in the new federalism, federal agencies have been working to both expand their efforts and redefine their role as producers of social indicator data. The purpose of this issue brief is to review recent federal efforts in this area, outline future directions for those efforts, and identify factors that may limit or promote progress.4
Social Indicators: The Evolving Federal Role under the New Federalism
Since 1990, federal agencies have been involved individually and collectively in a number of data development efforts to increase the amount of state and local data on children and their families. Several strategies have been followed, including the design and fielding of new state- and county-level surveys; the development of more detailed child population estimates at the local level; the redesign of existing national surveys to facilitate the production of state-level estimates in the larger states; the development of flexible survey mechanisms that would allow states to field state surveys on various topics economically; and the expansion of outreach and technical assistance to the states for the development and use of social indicator data.
I. Surveys and Assessment at the State and Local Levels
During the 1990s, several federal agencies have developed programs that regularly measure child health and educational achievement at the state and local levels. In addition, the Bureau of the Census is currently field-testing a large sociodemographic survey called the American Com munity Survey (ACS), which will, when operational, provide basic demographic and economic data at the state and local levels on an annual basis.
Health Surveys
Since 1990, the Centers for Disease Control and Prevention (CDC) has developed several programs designed to monitor the health of children at the state level through regular surveys. Topics covered include youth risk behaviors, birth outcomes, and child immunization.
Youth Risk Behavior Surveillance System (YRBS). The YRBS is a survey of health and risk-related behaviors among students in grades 9 through 12. Topics covered correspond to the Healthy People 20005 goals and include tobacco and alcohol use, sexual behavior, behaviors contributing to intentional and unintentional injuries, and diet and physical activity. The survey, which is voluntary, is repeated every two years. The YRBS questionnaire was designed in partnership with the education departments of all 50 states and 16 of the nation's largest metropolitan areas, over 20 federal agencies, and experts in each of the categorical areas measured by the system. In 1997, 38 states (including the District of Columbia), 17 cities, and 4 U.S. territories participated.6 A national survey is also fielded regularly. Each participating state and city may omit or add questions at its discretion. The surveys are fielded by the education departments in each participating state and city, with the CDC providing technical assistance, data processing services, and funds.
Pregnancy Risk Assessment Monitoring System (PRAMS). The PRAMS is an ongoing population surveillance system designed to survey mothers who have recently given birth. PRAMS is designed to oversample women at increased risk of poor birth outcomes. As of September 1997, 16 states are participating in PRAMS, which is funded through federal cooperative agreements with participating states.7 Through state health departments, data are collected on various maternal behaviors and experiences, including whether the pregnancy was planned, access to and use of prenatal and infant care, smoking, violence during pregnancy, and pregnancy-related morbidity. In addition, individual states ask questions on priority topics including maternal mental health, social support, employment, and receipt of social services.
National Immunization Survey (NIS). The NIS, a joint project of the National Center for Health Statistics and the National Immunization Program of the Centers for Disease Control and Prevention, provides annual state and local estimates of immunization rates among two-year-olds. Telephone surveys are fielded in all 50 states and in 28 urban areas. The survey, which began in 1994, is one component of the federal Childhood Immuni za tion Initia tive,8 and has demonstrated increases in overall immunization rates in most states and urban areas over the last several years.
Education Assessments
Since 1990, the U.S. Department of Education has fielded state-level assessments of student academic achievement in most states. In addition, it is now developing standard national tests that might be fielded in state and local areas interested in using them.
The National Assessment of Educational Progress (NAEP). This is a biennial survey intended to monitor the knowledge, skills, and academic performance of our nation's children and youth in reading, writing, math, science, and geography. State assessments have been administered to fourth- and eighth-grade students in math, reading, and science, with writing scheduled for 1998. The assessments are identical across all jurisdictions, allowing for direct cross-state comparisons. A national survey has been fielded since 1969. Since 1990, NAEP has been fielded on a voluntary basis in all states wishing to participate. In 1996, 44 states, the District of Columbia, and Guam participated.
Development of the assessments is overseen by a board comprised pri marily of state and local politicians and educators, providing for a maximum level of state and local control over content and approach. Members of the National Assessment Gov erning Board (NAGB) include governors, state legislators, chief state school officers, state and local board members, teachers, principals, and business leaders.
National Tests in Reading and Mathematics. President Clinton recently signed a bill that allows for the development of voluntary national tests in reading for fourth-grade students and mathematics for eighth-grade students. The design of the tests are being overseen by the NAGB, an independent and bipartisan panel comprised largely of state and local educators and policymakers. The tests are being designed to provide individual student assessment results, but may also yield valid scores at the school district and state levels when they are fielded to all students. The Department of Edu cation plans to pilot the tests in 1998 and begin full administration in the spring of 1999. Further congressional action will be required before the tests can be field-tested or made available for general use, however.
Sociodemographic Surveys
The American Community Survey (ACS). Data from the decennial census provide valuable basic demographic and socioeconomic information on all population groups for states and localities down to the neighborhood level. State and local governments rely heavily on these data for basic planning and resource allocation decisions. The major drawback is that these data are collected only once every 10 years. In response to the expressed needs of state and local governments for such data on a more timely basis, the U.S. Bureau of the Census is developing the ACS. National and state-level estimates from the ACS will be made beginning in the year 2000, and when fully implemented in the year 2003 will provide accurate population estimates on an annual basis for all population areas in excess of 65,000 persons. Estimates for census tracts will be produced every five years. The survey, which is virtually identical to the census long-form instrument, is currently being field-tested. When fully operational it will survey three million housing units per year. If successful, the survey may replace the census long form by 2010.
II. Estimates of the Child Population at the State and Local Levels
Access to accurate and detailed estimates of the child population is very valuable to states, particularly those that lack their own population estimates offices. Such estimates are very useful for county-level planning and provide the population denominators needed to estimate and track rates for all sorts of child well-being measures (e.g., the percentage of females ages 15–19 who gave birth in a particular county).
County Estimates of the Child Population. For many years the Bureau of the Census has worked in close cooperation with state governments to produce annual population estimates for states and counties. While annual child population estimates have been fairly detailed at the state level, with separate estimates by sex, race, and Hispanic origin for five-year age groups (0–4, 5–9, 10–14, and 15–19), such estimates have been produced at the county level only sporadically under contract to the National Cancer Institute. Beginning in late 1997, however, the Census Bureau began to produce more detailed child population estimates at the county level as part of its standard set of annual population estimates. These include the five-year breaks described above as well as single-year age estimates.
Small Area Income and Poverty Estimates (SAIPE). As part of a more general effort to provide local income and poverty estimates between censuses, the Bureau of the Census has begun producing biennial county-level estimates of child poverty for related children ages 5–17 and all children under age 18. Additional poverty and income estimates are being made for all persons. The first estimates were released in the fall of 1997, and a revised set of estimates became available in January 1998. County poverty and income estimates for 1995 are expected to become available in the summer of 1998, and poverty estimates for school districts the following fall.
These estimates are produced using an experimental methodology that uses data from the 1990 census, tax returns, and the Food Stamp and SSI programs. Once the estimation procedure has been adequately tested, these estimates will be used by both federal and state agencies in determining the allocation of program funds to counties and for a host of other planning purposes. Estimates must be lagged by three years (i.e., 1995 estimates will become available in 1998) because of the time it takes for tax and other necessary data to become available.
III. Redesigning National Surveys to Yield State Estimates.
National Health Interview Survey (NHIS). The NHIS, which has been conducted annually since 1957, is the nation's largest survey of the general health of the U.S. population, and includes a substantial amount of data on child health. In 1995, the sampling procedure was redesigned to allow for the calculation of state-level estimates. Due to sample size limitations, however, annual estimates using a single year's worth of data can be produced for only the largest states. Estimates for many additional states can be produced by combining several years' worth of data. Even after this is done, though, estimates for most states will have relatively large standard errors, making it difficult to detect significant change over time.
Current Population Survey Annual Demographic File (CPS). The CPS was redesigned in the 1970s to allow for the production of state estimates. For all but the largest states, this requires combining several years' worth of data in order to produce stable estimates. The Census Bureau has been publishing state estimates on selected population characteristics since 1989. The major sponsor for the Assessing the New Federalism project, the Annie E. Casey Foundation, has been producing such estimates for its annual Kids Count Databook since 1990.
IV. New Flexible Survey Mechanisms
State and Local Area Integrated Telephone Survey (SLAITS). The National Center for Health Statistics (NCHS) is developing the capacity to field telephone-based surveys tailored to the needs and interests of individual states at a reasonable price. A survey module featuring health questions, which were taken primarily from the National Health Inter view Survey, was pilot-tested in two states during the summer of 1997. A special welfare module containing questions about employment, service use, and child well-being will be tested in two to three additional states in early 1998.
If and when SLAITS becomes fully operational, it may target particular population subgroups, such as families with children, and will allow states to contract for surveys of their populations. States may choose from the existing survey modules. Health and welfare modules have already been designed, and others may be constructed. Standardized modules have the advantage of providing comparable estimates across participating states and will be more economical to field. Adding new questions of particular interest to the state is also being considered.
The federal government is paying the initial development costs for SLAITS. Once it becomes operational, costs will be shared by the federal government and the states. SLAITS can be fielded economically because it is piggy-backing on an existing telephone survey, the NIS, which is already contacting and screening over 800,000 households each year.
The Bureau of the Census is considering a similar arrangement using the ACS as the backbone for other state and local survey efforts. The ACS will be fully implemented in 2003.
V. Outreach and Technical Assistance Efforts
Federal agencies have recently initiated several efforts to work with states, identifying needs and offering technical assistance regarding the production and use of social indicators of children's well-being. Most of these efforts are still in their early stages, but may offer some insight into how the federal role will develop in this area.
Interagency Forum on Child and Family Statistics. This forum, which began in 1994 and was formally established through a Presidential Executive Order in 1997, includes 17 major federal statistical agencies that collect data on children and their families. The heads of these agencies meet every four months to plan and coordinate their activities to improve the efficiency and utility of the federal statistical system and better serve the evolving needs of its customers. The initial focus of the forum has been on the production and dissemination of social indicators, culminating in an annual report entitled "America's Children: Key National Indicators of Well-Being," the first of which was published in July 1997. A second edition is forthcoming in the summer of 1998.
While the focus of the report is on national trends, the forum is also interested in responding to increased state and local needs for such data and has formed a committee for this purpose. In May 1997, the committee and Child Trends, Inc., co-sponsored a conference that brought together representatives of states making substantial use of social indicators for government planning, goal-setting, and accountability; representatives of federal statistical agencies that produce state and local data; and national research organizations that work with states and localities on the development and effective use of social indicator data. The activities and emerging data needs of states were presented and discussed, as were possible roles that federal agencies could play in meeting those needs.
This forum is potentially an important conduit through which states can work with the federal statistical system to make it more responsive to their needs. As many states move toward a more comprehensive, cross-agency approach to planning, budgeting, and goal-setting, federal-state dialogue around issues of data and measurement will undoubtedly benefit from the existence of such multi-agency efforts as the forum.
Technical Assistance to the States Using Social Indicators of Child Well-Being. Through the Office of the Assistant Secretary for Planning and Evaluation, the Department of Health and Human Services has recently launched several projects to provide technical assistance to states regarding the selection, development, and use of social indicators of child well-being. The most recent project, led by Harold Richman of the Chapin Hall Center for Children at the University of Chicago, will develop activities with states that are interested in improving their capacities to use child indicators in policy efforts. In addition, the Chapin Hall group will work with other organizations (both private and public) that have expertise in developing and using social indicators in ways that serve the needs of interested states.
Child Trends, Inc., has organized a second project in the form of one-day meetings in three states—Oregon, Vermont, and Florida—whose governments are making substantial use of social indicators in state and local planning efforts. In these meetings state and local officials are brought together to discuss their work, their goals, and how federal agencies could help them reach those goals.9
The third project provides funds and technical assistance to five states that have been awarded grants to incorporate child well-being measures into their own welfare evaluation designs. In addition to the program evaluations that each has developed, the states are considering ways to develop and track selected indicators of child well-being to monitor how children in general are faring as additional changes occur in welfare, health, and other policy areas. The National Institute of Child Health and Human Development (NICHD) Family and Child Well-Being Net work is providing technical assistance with Child Trends, Inc., as the lead organization.
The National Performance Review and the Oregon Option. The White House initiated the National Per formance Review in 1993 to reinvent how the federal government performs its tasks to make it more efficient and more responsive to the needs of those whom it serves. In 1994, it entered into an agreement with the state of Oregon to participate in a multi-year experiment to "develop an outcomes-oriented approach to intergovernmental services." Oregon had for some years been reorienting its approach to government toward measurable goals as embodied in Oregon Benchmarks. Benchmarks currently includes over 90 measurable goals for improving the well-being of Oregon's population, including 20 that are directly related to the well-being of children.
The Oregon Option builds work clusters of federal, state, and local staff around particular issues, facilitating cross-agency and cross-level planning. The federal role in these clusters includes coordinating the federal response to specific state and local needs, providing waivers to federal requirements on a selective basis, and providing technical assistance to state and local staff on a variety of data and measurement issues.
Prospects for the Future
During the 1990s, federal agencies have put substantial effort into expanding the amount of state and, to a lesser extent, local data available for tracking the well-being of children. Furthermore, they have begun to enter into serious dialogue with states to determine what states need where social indicators are concerned and what role federal agencies can usefully play in meeting those needs.
Such conversations, if consistently pursued, can produce the outlines of a new role for federal statistical agencies under the new federalism. There are several reasons to believe that there is a legitimate role for federal agencies and that it may a substantial one. First, they must produce substantial amounts of social indicator data for their own purposes, including estimates at the state and local levels. While the particular measures of social well-being of interest to individual states and federal agencies may not be identical, states can take advantage of economies of scale by participating in federally orchestrated survey efforts such as the YRBS and SLAITS, particularly when they are given substantial input into the original survey design. Second, federal agencies have human resources that many states and most local governments cannot match.
While no one can predict the ultimate outcome of conversations between the federal and state agencies regarding the production and use of social indicator data, there are a number of areas in which federal agencies might usefully develop their work beyond the efforts already described. These include
- Filling in existing data and measurement gaps. While federal agencies have increased the amount of data on children and their families available at the state and local levels, there remain substantial gaps in areas of great concern to states. For example, there are no federal sources of data on school readiness among young children below the national level, and the measures themselves are inadequate to state needs. Measures and data that reflect the positive developmental assets of children rather than their shortcomings represent another important gap. These are areas where federal agencies can contribute by working with states to develop good measures that can be fielded by states in a cost-effective manner.
- Generating county-level data. To date, most of the subnational data generated by federal statistical agencies have been limited to the state level. States, on the other hand, often need county-level social indicator data. While it is often impractical for federal agencies to collect county data directly, they can assist states that wish to generate county estimates. For example, Oregon and Vermont have paid the marginal additional costs to field representative samples of the YRBS in most school districts in each state.
- Facilitating cooperative networks among states. Much of the expertise regarding the development and use of social indicator data at the state and local levels resides in the states themselves. The federal government can help states organize and share best practices, develop solutions to common problems, and promote interstate cooperation in data collection. Participants in a recent federally organized conference, which included many state representatives, strongly recommended the formation of such networks and asked that the federal government work to facilitate their organization.10
Though the efforts of federal statistical agencies to redefine their role under the new federalism are promising, they are also fragile. There are a number of forces that may limit or even reverse the initiatives described above. Tight federal budgets have already threatened and/or delayed many of the programs discussed in this issue brief. In 1996, Congress considered legislation that would have eliminated the NIS, a move which would have also effectively killed the innovative SLAITS survey program, which is dependent on the NIS survey sample. The Bureau of the Census has repeatedly had to push back the date for full implementation of the ACS for lack of funds. These funding issues may become even more serious as we approach the 2000 census, and the ACS may yet be further delayed or even eliminated.
Concerns about family privacy issues may also have a substantial limiting impact on the role that federal agencies can play in dealing with important but potentially sensitive issues at any level. Elements of the Family Privacy Act of 1996, which very nearly passed, would have vastly increased the cost of collecting data from youth by requiring active written permission from parents. Passage would have threatened the viability of the YRBS, currently fielded in 39 states, and greatly restricted future options for data development.
Finally, there are the challenges resulting from competing needs and interests within the bureaucracies of federal statistical agencies. When budgets are not increasing—and they are not likely to do so in the foreseeable future—attempts to expand activities to support data collection and use at the state and local levels will have to compete with existing program needs and new federal data needs. Strong and consistent leadership within the federal government and by the states themselves will be required if federal statistical agencies are to continue expanding efforts to provide child indicator data at the state and local levels and to support states in the development and use of such data.
Notes
1. Brown, B., and T. Corbett (forthcoming). "Social Indicators and Public Policy in the Age of Devolution." In Trends in the Well-Being of Children and Youth," R. Weissberg, L. Weiss, O. Reyes, and H. Walberg (eds.). Sage Publications.
2. The state of Colorado, for example, has given its counties substantial flexibility in the design, implementation, and monitoring of their own welfare reforms related to TANF.
3. For a recent review of child indicator data from federal sources available at the national, state, and local levels, see B. Brown, 1997. "Indicators of Children's Well-Being: A Review of Current Indicators Based on Data from the Federal Statistical System." In Indicators of Children's Well-Being, R. Hauser, B. Brown, and W. Prosser. New York: Russell Sage Foundation.
4. States gather a great deal of their own data on children independently of the federal government. States that are particularly active in this regard include Oregon, Vermont, and Minnesota. For a review of social indicator data collection activities for a selected group of states, see B. Brown, G. Kirby, and C. Botsko, 1997. "Social Indicators of Child and Family Well-Being: A Profile of Six State Systems." IRP Special Report Series #72. Madison, WI: Institute for Research on Poverty.
5. Healthy People 2000 is a federal initiative to improve the health of the U.S. population. There are nearly 300 specific objectives, which were jointly developed by the U.S. Public Health Service, 50 state health departments, and over 270 national organizations.
6. Participating states in 1997 include Alabama, Arkansas, California, Colorado, Connecticut, Delaware, Florida, Georgia, Hawaii, Idaho, Iowa, Kansas, Kentucky, Louisiana, Maine, Mass achusetts, Michigan, Mississippi, Missouri, Montana, Nebraska, Nevada, New Hampshire, New Jersey, New York, North Carolina, North Dakota, Ohio, Oregon, Rhode Island, South Carolina, South Dakota, Tennessee, Utah, Vermont, West Virginia, Wisconsin, and Wyoming. Participating cities in 1997 include Baltimore, Boston, Chicago, Dallas, Detroit, the District of Columbia, Fort Lauderdale, Houston, Jersey City, Los Angeles, Miami, New Orleans, Newark, New York City, Philadelphia, San Diego, and San Francisco. Participating territories include American Samoa, Guam, the Northern Mariana Islands, and the Virgin Islands.
7. Participating states include Alabama, Alaska, Arkansas, Colorado, Florida, Georgia, Illinois, Louisiana, Maine, New Mexico, New York (except New York City), North Carolina, Oklahoma, South Carolina, West Virginia, and Washington.
8. This initiative is a comprehensive effort to increase vaccination coverage levels among two-year-olds throughout the U.S.
9. A report will be available in the summer of 1998.
10. See Kirby, G. 1997. "Summary of the Workshop on Social Indicators of Child and Family Well-Being in the Age of Devolution: Defining Next Steps." May 29–30, 1997. Washington, DC: Child Trends, Inc.Other Briefs in the Series
About the Authors
Brett V. Brown is a senior research associate at Child Trends, Inc. His work focuses on issues related to the production, use, and dissemination of indicators of child well-being. For the past several years he has directed the production of a comprehensive annual report titled Trends in the Well-Being of America's Children and Youth, released by the Assistant Secretary for Plan ning and Eval uation, U.S. Depart ment of Health and Human Services. His other research interests include fatherhood and the determinants of successful transitions from youth to adulthood.
Child Trends, Inc. Child Trends, Inc. is a nonprofit, nonpartisan research organization dedicated to studying children, youth, and families through research data collection, and data analysis. In Assessing the New Federalism, Child Trends has responsibility for conceptualizing and designing ways to measure changes in children‘s well-being, and for assessing the implications of policy and programmatic changes for children.
