Barth_paper

Current Projects

DRAFT

Administrative Data on the Well-Being of Children On and Off Welfare

Richard Barth and Eleanor Locklin
University of North Carolina

Stephanie Cuccaro-Alamin and Barbara Needell
University of California

Richard P. Barth, Ph.D. is Frank A. Daniels Professor of Human Services Information Policy and Eleanor Locklin, MPH, MSW is a doctoral student at the School of Social Work, University of North Carolina, Chapel Hill. Stephanie Cuccaro-Alamin, MSW is a doctoral student and Barbara Needell, Ph.D. is a Senior Research Associate at the School of Social Welfare, University of California at Berkeley. The corresponding author can be reached at rbarth@email.unc.edu.

Introduction

The Personal Responsibility Work Opportunity Act of 1996 significantly altered the way the United States provides assistance to its neediest citizens. The Act eliminated the federal entitlement to services which existed under the Aid to Families with Dependent Children (AFDC) program and replaced it with the block grant program Temporary Assistance to Needy Families (TANF). TANF provides temporary financial assistance while recipients make the mandatory transition from welfare to work. These efforts to move adult recipients toward self-sufficiency undoubtedly have consequences for their children (Collins & Aber, 1996; Zaslow, Tout, Botsko, & Moore, 1998). Potential implications include changes in children’s health, safety, education, and social competence.

PRWORA is currently being praised for moving nearly 1.7 million recipients from welfare to work (U.S. Department of Health and Human Services, 1999). However, until the impact of these reforms on child well being is known, such celebrations are premature. For, even if reforms succeed in moving mothers from welfare to work, if this in turn has negative consequences for children, its effectiveness will need to be re-evaluated in light of these costs.

Prior to passage of welfare to work legislation more than 40 states received waivers to experiment with welfare to work programs. Experimental evaluations of these initiatives now underway, will provide valuable information regarding possible effects of PRWORA. Most, however, focus primarily on adult outcomes such as changes in income, employment, family formation and attitude changes and cover only a limited number of child outcomes (Research Forum on Children, Families, and New Federalism, 1999; Yaffe, 1998). Additionally, those child outcomes which are included "typically lack depth and uniformity" (Yaffe, 1998). Recently, the Administration for Children and Families (ACF), Office of the Assistant Secretary for Planning and Evaluation (ASPE) has implemented The Project on State-Level Child Outcomes to assist waiver states in using administrative data to expand these child outcome measures and make them comparable across states. While this uniformity will allow for the assessment of different state models, their utility is limited by their small sample sizes. In particular, small sample sizes make subgroup comparisons difficult and prohibit evaluation of rare events such as foster care placement or child mortality. Current evaluations of state welfare to work initiatives under PRWORA suffer from similar limitations.

Given these limitations, researchers are increasingly turning to administrative data. Administrative or program participation data collected by state agencies can be configured to provide both in-depth and uniform measures of child well being. The purpose of this paper is to assist researchers in defining such indicators and illustrate how they can be used to evaluate the impact of welfare reform. The paper is divided into four sections corresponding to each of four domains of child well-being¾health, child welfare (safety), education, and social competence (juvenile justice). ¾For each domain we will identify information that directly describes or reflects on child well-being.. Each section also includes a description of key data that are available to inform us about children’s outcomes. For each domain we also discuss exemplary efforts to use these data for evaluation of welfare reform. Finally, we conclude the paper by discussing some of the scientific sensibilities that should be respected in the use of such data during research on welfare reform, including a discussion of linkages between population surveys and administrative data.

Domains of Child Well Being

Health

Children’s health is a central consideration in the assessment of the implementation of welfare reform because these reforms changed the relationships between employment, public assistance, and insurance of health care services for poor families and children (Darnell & Rosenbaum, 1997; Moffett & Slade, 1997; Schorr, 1997). These changes have the potential to impact access to health care. In addition, welfare reform has the potential to change, positively or negatively, the family environment where health behaviors and health decisions are carried out (Willis & Kleigman, 1997; O’Campo & Rojas-Smith, 1998; Brauner & Loprest, 1999). These health behaviors and decisions include prenatal care, care for low birth-weight infants, immunizations, substance abuse, and teen pregnancy--all important health status measures for children. Therefore, child health indicators used to evaluate welfare reform should measure impact on access and impact on health status. One of the most basic indicators of child health is access to health care (Gortmaker & Walker, 1984; Margolis, Cole, and Kotch, 1997; Andulis, 1998a).

Most children’s health care is paid for through employer sponsored private insurance, or publicly funded programs such as Medicaid or the new Child Health Insurance Programs (CHIP). However, in July 1999 there remained an estimated 4.7 million uninsured children eligible for Medicaid but not enrolled (Families USA, 1999). These children lack access to care because they are not part of a health care payment plan. Furthermore, children who are covered by a health care payment plan may lack access to needed care due to limitations of the plan itself, --that is, referral procedures or physician availability (Farel, 1997). For all children, limited access to health care translates into: (1) lack of preventive services such as immunizations or dental care; (2) lack of diagnostic services such as vision and hearing screening, or weight for height measures (3) lack of treatment for chronic conditions and disability, with corresponding decrease in functioning, and (4) increased risk of secondary disability.

There are several indicators that measure children’s access to health services. First is a measure of the number of children with health insurance and the source of that coverage over time. In a recent study of CHIP and Medicaid enrollment data, researchers found that fewer children were covered in 1999 by these federally funded programs, both Medicaid and CHIP, than were covered by Medicaid alone in 1996. Yet, there was no simultaneous decrease in the number of children without health care coverage. Thus the gap between 1996 and 1999 was not being covered by private insurance coverage (Families USA, 1999).

Many states are beginning to track children’s enrollment in Medicaid and CHIP, implement outreach efforts to increase CHIP enrollment, and expand Medicaid and CHIP income level guidelines (Families USA, 1999; CDF, 1998). These changes over time can be studied at the population level and also at the individual child level. For example, a child who was automatically eligible for Medicaid due to AFDC enrollment will no longer receive the new TANF assistance after two years. The child would be eligible for Medicaid for one more year at the prior AFDC eligibility status for Medicaid. Three years after welfare reform the child would either be covered by an employer sponsored plan, Medicaid, CHIP, or have no coverage at all.

Another access measure that should be included is the type of health care coverage a child receives (Moffit & Slade, 1997). While children on Medicaid often receive more comprehensive services than many private plans, states must evaluate the exact benefits package of services offered by Medicaid, CHIP and private insurance. For example, if a child previously covered by Medicaid is then enrolled in an employer sponsored plan, is the new plan as comprehensive as Medicaid and is that child at greater health risk due to the new plan? If important services such as physical therapy, or specialists such as hematologists, are not easily accessed in the new plan, the change could be a negative one in terms of the child’s health.

In addition to measuring children’s access to health care, it is also necessary to examine measures of child health status such as birthweight, immunization, or child injury. Recent discussions about welfare reform and health suggest a range of indicators to measure child health status. Children in poverty are more likely to be under-nourished, iron deficient, or lead exposed (Geltman, Meyers, Greenberg, & Zuckerman, 1996). Several measures such as infant mortality, injury, and preventive services are all excellent indicators of child health (Pappas, 1998). Other more sensitive indicators have been developed such as Starfield’s Child Health and Illness Profile (Starfield, 1993). However, while these more sensitive measures provide an in-depth assessment of the child’s biopsychosocial development and can be used effectively in survey research, they are typically not found in administrative data sets.

Healthy People 2000, an initiative begun in 1990 by the U.S. Department of Health and Human Services, set health objectives for the nation, including child health status objectives (National Center for Health Statistics, 1996). Over the years the initiative has prompted state and local communities to develop their own similar objectives and indicators of progress towards achieving them. As a result, the Healthy People 2000 effort has created a set of fairly common measurements of child health across the country and across a range of public and private health programs. For example, one of the Healthy People objectives is to reduce infant mortality. This supports the inclusion of infant mortality reduction as part of most state health objectives, and as part of many state and local programs targeted towards women and children. At the federal level, the Maternal and Child Health Bureau (MCHB) identified eighteen of the Healthy People 2000 objectives that specifically relate to women and children. Of these, fifteen are child health status indicators that can be used to measure impact of welfare reform (Maternal and Child Health Bureau, 1996).

Our purpose here is to identify a reasonably comprehensive set of child health indicators that are appropriate to administrative data, relevant to changes in welfare policy, and address both the access and health status needs of children. The indicators described in Exhibit 1. below are derived from our discussion of both access measures and health status measures. For each indicator we describe whether data is generally available at the individual level or aggregated to some larger population. In many instances, however, the aggregation of data is at the community, county or state level, based on local health department records, disease registries, or program evaluation data. While these aggregate data may not be compatible for linkage to welfare tracking systems by individual, measures of income or welfare status may be included in the data sets and prove useful for our purposes. For individual level data, issues of linkages between data sets and of confidentiality remain. However the development of such linkages is progressing (National Conference of State Legislators, 1999; Farel Locklin, & Peoples-Sheps, 1999).

Exhibit 1. Suggested child health indicators.

INDICATOR	LEVEL	DATA SOURCES
Medicaid enrollment/services	Individual	Medicaid data files
CHIP enrollment/services Number/percent uninsured	Individual Population	CHIP data files State Dept. of Insurance
SSI benefits	Individual	SSA Data
Infant mortality	Individual	Vital Statistics
Low birthweight	Individual	Vital Statistics
HIV infection among women w/live births	Individual	Vital Statistics
Prenatal care	Individual	Vital Statistics
Newborns screened for genetic disorders	Individual	Vital statistics, clinic records
EPSDT	Population	Disease registries
Identification of hearing impairments	Individual Population	Clinic records, EPSTDT Program evaluation data
Immunizations	Individual Population	Medicaid, clinic record Program evaluation data
Blood lead levels	Individual Population	EDSDT, clinic record Program evaluation data
Dental Caries	Individual	Medicaid, clinic record
Unintentional injuries	Individual	Vital statistics, hospital discharge, School Based Health Clinics
Homicide	Individual	Vital statistics
Adolescent suicides	Individual	Vital statistics
Adolescent substance use	Individual Population	Vital statistics Hospital discharge, SBHC, program evaluation data
Youth STD's	Population	Hospital discharge, program evaluation data, SBHC
Teen pregnancy	Individual	Vital statistics

Data sets that include indicators of child health at the individual or population level are generated in several ways. Clinical health services, as well as initiatives to improve children’s health, are carried out by a wide variety of organizations -- local, state, or federal public health agencies, private doctors, nurses, social workers, hospitals, community clinics, research centers and social services. These organizations often work together in collaboration, sometimes due to funding mandates or evaluation requirements. Because public health as a profession is specifically charged with carrying out core functions of assessment, policy development and assurance (IOM, 1988), the public health role of monitoring health status has created an ongoing infrastructure of administrative data (Morris et al., 1996; Sondik, 1996).

There are several individual level data sets that can be linked to AFDC and TANF welfare data sets. Data from Medicaid eligibility, enrollment, and claims, and the new state Child Health Insurance Programs (CHIP) can be linked to provide longitudinal tracking of a child or family’s health care coverage or lack of coverage. For example, a state could follow a child from AFDC enrollment with Medicaid, to TANF with Medicaid, to Medicaid only, to CHIP only. Children who do not follow this pattern, or a similar pattern that includes CHIP or Medicaid expansion coverage, could be at risk of no health care coverage. It is unlikely that many families leaving TANF will go to jobs with sufficient health benefits or wages above the Medicaid and CHIP guidelines, that is, 200% of the federal poverty level in most states but 350% of the federal poverty level in some Medicaid expansion programs. Thus, if a child leaves TANF and Medicaid abruptly they may be at risk of no health care coverage. If Early and Periodic Screening and Diagnostic Testing (EPSDT) services are also recorded, similar linkages with welfare data will support evaluation of preventive services for these low-income children. Beyond data on eligibility and enrollment, the actual Medicaid or CHIP benefits within a state should also be considered part of the evaluation. State CHIP programs can vary by age, geographic area, disability status, or calculation of income.

Examination of claims data under Medicaid or CHIP must be concerned with issues of confidentiality. However, many states such as California, Maryland, Kentucky and Tennessee, are already addressing these concerns through data sharing and data warehousing projects (National Conference of State Legislators, 1999). In some states, such as North Carolina, a common health identifying number is used across a range of data sets from vital statistics to disease registries (North Carolina State Center for Health Statistics, 1997). Where the health identifying number and social services number can be linked, one can evaluate a child’s experiences and outcomes across both. California uses data files on Medicaid recipients as the core of their data sharing/data integration initiatives (National Conference of State Legislators, 1999). This can allow evaluators to track over time and across programs those services provided to low income children and families. Issues of data linkage, common data elements, confidentiality, and data protocols are all part of developing such a state system (Farel, Locklin, Peoples-Sheps, 1999). Using Medicaid as the foundation for such a system promotes linkage between welfare/income data and health services data.

Most of the current effort to evaluate welfare reform and the impact on children’s health focuses on insuring health care services for children. Families USA recently published a report about Medicaid and CHIP enrollment from 1996 to 1999. Monthly enrollment data from twelve states were analyzed to evaluate changes in traditional Medicaid enrollment, enrollment in Medicaid expansion programs and enrollment in State CHIP programs. The twelve states included were those with the largest numbers of uninsured children: Arizona, California, Florida, Georgia, Illinois, Louisiana, New Jersey, New York, North Carolina, Ohio, Pennsylvania, and Texas. Findings included a decrease in Medicaid enrollment, and gradually increasing CHIP enrollment. However, the numbers of uninsured children had not declined. The findings and questions raised in this report are useful to states that want to examine their own enrollment trends. Methodologies used to evaluate monthly enrollment data are also covered.

Supplemental Security Income (SSI) records can also be linked to welfare data sets and identify where children may be losing needed coverage. Due to recent changes in SSI eligibility, it is possible that many children will lose coverage, severely limiting health care for children with special health care needs unless needed services are provided elsewhere (Perrin, 1997; Doolittle, 1998).

Vital statistics systems should also be linked to welfare data sets to provide a range of child health data. Birth records carry information of at least three kinds: about the timing and nature of the birth (e.g., family size, birth spacing), about services and the payment source for the birth (e.g., prenatal care used and whether the birth was covered by private pay, Medicaid, or medically indigent funds), about the family (e.g., marital status); and about the well-being of the child at the time of birth (e.g., birthweight, length of hospital stay, APGAR 5 and 10 minute scores, and the presence of congenital abnormalities).

Although welfare agencies would indicate that preventing child death is not their responsibility alone, it is also important to understand how receipt of their services is related to variations in child mortality. A variety of relevant outcomes can be captured using death records, including adolescent suicide, adolescent homicide, accidental deaths, deaths caused by injuries, and deaths casused by substance abuse (if overdose is the cause of death). Death records are in the public domain and are available at the state level as well as from the National Death Index (NDI)--a central computerized index of death record information on file in the State vital statistics offices. Investigators can also obtain data at the state level and make arrangements with the appropriate State offices to obtain copies of death certificates or specific statistical information such as cause of death.

Several public welfare agencies have matched child deaths against their welfare caseloads to better understand the vulnerability of their populations. In Children’s Deaths in Maine, the authors (Maine Department of Human Services, 1983) concluded that "during the period of 1976-1980 children of Maine’s low income families, as defined by their participation in AFDC, Medicaid or Food Stamp Programs experienced an overall death rate approximately three times as great as that for all other children. A follow-up case control study (n=912), also included interviews of parents whose children had died as well as case controls selected from the general population. Children in Maine whose parents were on welfare did not have mortality that was significantly higher than other children in poverty, although there were some types of mortality that were high among welfare recipients. For example, the risk ratio for children whose parents were on welfare of experiencing a death from non motor vehicle accidents was 5.2. In a more recent study (Phillips et al., 1999), mortality related to homicide, suicide, and automobile accidents (when substance abuse was mentioned on the death certificates) was shown to be substantially higher in the first week of the month--probably related to the greater availability of discretionary income following the arrival of government assistance checks and pay checks. Thus, evaluations of the relationship between deaths and welfare changes need to assess the type and timing of the deaths. Because child mortality is relatively rare-even among high risk populations--studies of welfare populations may need to combine these mortality data with injury data and incarceration data (discussed later in this chapter) to obtain an overall assessment of significant threats to well-being.

Other data sets may not be as easily linked to welfare data sets, yet they should be considered. . Twenty-one states have comprehensive databases on hospital discharges (Pappas, 1998). These data can provide information about a wide variety of health concerns such as child injuries, acute illnesses, and emergency room visits. These data sets may include measures of income, payment authorization, or actual welfare status. They could also be linked to welfare data to obtain some of this information. Again, one can evaluate trends in method of payment for hospital services, from Medicaid to CHIP to private insurance to no coverage.

Data from school based health centers could also be useful. They are often under the umbrella of a local hospital, and can serve as Medicaid providers under managed care contracts. These data will be most useful when they cover a large proportion of all youth in the area under study and when they provide additional information not available in the Medicaid data. This is the case in Colorado and Connecticut which have extensive school based health center networks (Kopelman & Lear, 1998).

Another source of data are programs funded under Title V, the Maternal and Child Health Block Grant, that are using performance measurement for contracting and evaluation. State welfare reform evaluators should collaborate with Title V program staff to explore data linkage, inclusion of common data elements of welfare status and health across data sets, and other ways to share data and evaluate child health in the era of reform. For example, several states, including Kansas and Arizona, are implementing performance measurement systems in their Title V maternal and child health programs (Gabor, 1997; Grason, 1997). As a pilot across seven states, sponsored by the Maternal and Child Health Bureau in 1998, core performance measurements are monitored. These measurements include: needs assessments, percentage of Medicaid eligible children enrolled, standards of care for women and children, health insurance coverage, and cooperative agreements between state Medicaid, WIC, and other Human Service agencies. An emphasis on information systems development is also part of these pilot programs and should be explored for linkage with welfare reform evaluation. In another example, the Institute for Child Health Policy at the University of Florida Gainsville is currently evaluating enrollment in their Healthy Kids programs of outreach to uninsured children, as well as the qualtiy of services in the program for children with special health care needs (Reiss, 1999; Shenkman, 1999).

Efforts to promote and monitor state health objectives should include indicators of children’s health according to welfare, employment, and/or income status. As state and local communities plan for future Healthy People 2010 objectives, the impact of continuing welfare reform should be part of future health objectives. Where monitoring systems exist or are planned for, they should include either linkage to state and local welfare data sets, or common data elements that would provide for evaluation. For example, child health status measures could be regularly monitored according to the following categories: employed families with private health coverage, employed families with Medicaid or CHIP coverage, employed families with no coverage, unemployed families with Medicaid or CHIP, unemployed families with no coverage. These categories could be applied across a range of child health measures: prenatal care, infant mortality, low birthweight, immunizations, hearing and vision screening, specialist care for children with special health care needs, injuries, or teen pregnancy.

Child Abuse and Neglect

The considerable overlap between welfare and child welfare service populations is well documented. It is estimated that children from welfare families account for as much as 45 percent of those served by the child welfare system (American Humane Association, 1984). Considerable evidence suggests that the strong association between welfare and child maltreatment may be due to a number of factors including: the stresses associated with poverty, the existence of concurrent risk factors such as mental illness and illicit drugs, and welfare recipients more frequent contact with public authorities (Coulton, Korbin, Su, & Chow 1995; Gelles, 1992; Gil, 1971; Giovanni & Billingsley, 1970;

olock & Magura, 1996; Zuravin & DiBlasio, 1996).

Given the documented association between welfare and child maltreatment, a number of authors have reflected on the possible impacts of welfare reform on child welfare (Aber, Brooks-Gunn & Maynard, 1995; Haskins, 1995; Meezan & Giovannoni, 1995; Wilson, Ellwood, & Brooks-Gunn, 1995; Zaslow, Moore, Morrison, Coiro, 1995). Essentially all conclude that efforts to induce welfare mothers to self-sufficiency may impact rates of child maltreatment. Whether this impact is positive or negative depends in part upon what effect reforms have on family income and parental stress (Collins & Aber, 1997a & b). For instance, economic hardship related to loss of benefits or supports may strain family's abilities to provide basic necessities such as food and shelter causing increased neglect and homelessness even abandonment (Collins, 1997; Knitzer & Bernard, 1997; Shook, 1998). Additionally, changes in SSI and other program eligibility legislated by PRWORA might result in increased foster care entries (Knitzer & Bernard, 1997). Increased parental stress related to economic, employment, or childcare difficulties may also lead to increased rates of abuse (Knitzer & Bernard, 1997; Meezan & Giovannoni, 1995). In contrast, positive changes in these areas may be favorable to children and families. For instance, rates of abuse and neglect may decline if reforms reduce family’s economic hardship. Additionally, gainful employment might improve the mental health of single mothers thereby decreasing the risk of child maltreatment (Garfinkel & McLanahan, 1986). Better access to mental health and drug services might also have similar effects. Whether positive or negative, the these changes will likely be reflected in the number and types of maltreatment reports, the number of case investigations and substantiations, as well as the number of children placed in foster care.

In addition to impacting rates of child maltreatment, welfare reform may also effect the experiences of the children served by the child welfare system. With the passage of PRWORA, a family’s economic circumstances become a critical component of the child welfare decision making process. In particular, a mother’s TANF status could influence the decision to remove, and if removed the TANF status of potential kin caregivers might alter placement decisions (Zeller, 1998). For instance, the proportion of kin placements might decline, because kin caregivers may not be exempt from TANF requirements (Berrick, Minkler, and Needell, 1999; Geen & Waters, 1997, Boots & Geen, 1999). Economic factors might also influence children’s length of stay, placement stability, as well as their rates and types of exits from the system. Specifically, parental TANF status might facilitate or stall reunification efforts impacting the duration of children’s out of home placements. Children placed with kin might experience placement disruptions if their TANF status changes. While the impact of TANF noncompliance on reunification efforts is clear, compliance might also be problematic, with work making it difficult for parents to meet child welfare timelines such as visitation and court appearances (Knitzer & Bernard, 1997). In addition to these potential impacts on exit rates, changes in family’s TANF status following reunification might lead to an increased likelihood of re-abuse and child welfare system recidivism.

Unlike the domain of child health, in child welfare, there are relatively few sources of administrative or program participation data, which can be configured to provide both in-depth and uniform outcome indicators. Specifically, relevant data is collected only by state child protective service and foster care service departments. In some states (e.g., California) all child welfare administrative data are entered into one data system. The following section provides an overview of different configurations these data sources used to construct child welfare indicators. Additionally it discusses a variety of research designs that can be utilized to assess the impact of welfare reform on child maltreatment rates and children’s experiences in and exits from the child welfare system. Access and confidentiality issues loom large when using such data to study vulnerable children. Readers should consult Brady, et. al. (1999) for an in-depth review of these important topics.

Data Configuration

Most administrative data in the child welfare domain is comprised of service event types and dates which can be configured to construct a variety of outcome indicators. The two most common configurations are caseload and longitudinal data. In addition to program participation data, demographic information for the children and families under study (e.g., birthdate, ethnicity, home address or location) are also common elements found in these databases. When combined with this demographic data, caseload and longitudinal indicators can provide a comprehensive representation of both system performance and client status.

Caseload Indicators

Caseload data provides a snapshot of child welfare at a specific point in time. This configuration is usually used for program management purposes and is instrumental in assessing system impacts. For instance, understanding how many children are currently in care and their characteristics can be critical to system administration. Recently the Administration for Children and Families’ Child Welfare Outcomes and Measures Project developed a set of outcome measures using point in time data from the Adoption and Foster Care Analysis and Reporting System (AFCARS) to assess state performance in operating child welfare programs. Outcomes include annual incidence of child maltreatment, types of exits from the child welfare system, timing of exits, and placement stability. Although point in time data can also be used to measure case status outcomes such as foster care length of stay; the resulting statistics tend to be biased and provide a skewed representation of outcomes. Thus, while point in time estimates are the easiest and least expensive configuration of administrative data, this inherent bias limits their usefulness. Despite these limitations, caseload data can be configured to provide effective measurements of system performance.

Longitudinal Indicators

While point in time data provides useful information regarding system performance, an accurate representation of all children’s experiences can only be captured through the use of longitudinally configured data. Administrative data can typically be reconfigured into event level files that record program participation histories. Depending on the scope of available data, these events may be restricted to foster care spells or placements, or may more broadly include child abuse reports, investigations, and services provided in the home. Using data that can be subset into entry cohorts captures the dynamics of both system entries and exits, and therefore provide a more accurate assessment of outcomes than caseload. Although free from the biases of point in time data, it often requires considerable programming to reconfigure data into an event level, longitudinal format, and therefore it is both more complex and expensive.

The Multistate Foster Care Data Archive provides an illustration of both the complexity as well as promise that longitudinal data offers researchers. The archive is a collaborative initiative by the Children’s Bureau of the U.S. Department of Health and Human Services designed to foster increased collaboration between states with regards to administrative data collection in the child welfare services arena. Administered by the Chapin Hall Center for Children at the University of Chicago the archive currently includes data from child welfare agencies in 11 states. The archive reprocesses state data to make them comparable across state systems. To ensure data comparability the project focuses on “a limited set of characteristics and events that have clear meaning in all jurisdictions” (Wulczyn, Brunner, Goerge, 1999, p. 1). The core of the archive is two databases one consisting of child records including unique identifiers and demographic information and a second event level field which stores information on child welfare events of interest. This structure allows researchers to use the data in a longitudinal format to children’s spells in child welfare as well as other experiences. Additionally, data can be configured to provide traditional point in time estimates of caseload flow over time. A sufficiently comprehensive set of outcome indicators is shown in Exhibit 2 (note that some indicators have a clearer theoretical relationship to welfare reform than others).

Exhibit 2: Minimum Child Welfare Services Indicators

Child Maltreatment Reports (with Reason for Report Case Investigations (with Reason for Not Investigating) Case Substantiations (with Reasons for Providing Services or Not) In-Home Services (Duration And Frequency Of Provision) Foster Care Placements (with Placement Dates & Type of Placement) Placement Moves Foster Care Exits (with Type of Exit) Reentry to Foster Care (with Reason for Reentry)

Depending on the purpose of the analysis indicators can be derived from either point in time or longitudinal data. Indicators can often be expressed as rates based on state, county, and even zip code underlying populations, such as the foster care incidence (entry) rates and prevalence (caseload) rates by age and ethnicity. Benchmarks can be set for both caseload and longitudinal indicators, such as prevalence rate over time, or number and proportion of children who experience reabuse with one year of being reunified from foster care.

Measuring the Impact of Welfare Reform

A variety of designs can be used to evaluate the impact of welfare reform on child welfare. In a time series design, outcomes such as incidence rates, caseload size, or annual foster care entries are tracked each year to provide an aggregate measure of the impact of welfare reform. For instance, one could establish a baseline for the outcomes of interest for several years prior to welfare reform, and then follow-up measures after TANF implementation within specific states or counties. Changes in these performance indicators can provide important information about the child welfare system’s response to welfare reform. Although one can also build models that control for other factors related to welfare reform such as changes in the economy or grant levels, because welfare and child welfare data are not linked at the child level, such designs remain limited. Specifically, they can only provide information regarding the direction of changes but not identify the characteristics associated with children and families most likely to be affected (Magura & Moses, 1986).

To make stronger inferences regarding the impact of welfare reform on child welfare researchers must be able to follow individual children and families. This can only be done with longitudinal linked administrative data. In its simplest form, welfare careers for children can be linked to their child welfare careers. For example, a cohort of children new to aid can be followed to determine the extent to which they experience child welfare contact over time. As with time series analysis, other factors can be controlled for in models that look at the likelihood of child welfare involvement before and after TANF. In a more complex model, parental welfare paths can be followed and linked to child welfare events experienced by their children.

In anticipation of the effects of welfare reform, researchers in several states have undertaken projects using linked longitudinal welfare and child welfare data, to better understand the overlap between these two programs. These projects serve as exemplary models of what will be possible with post TANF data. The Child Welfare Research Center at the University of California Berkeley undertook an analysis to identify the characteristics of poor families at risk of child maltreatment. Using data from the California Children’s Services Archive, researchers constructed a longitudinal database of children entering AFDC between 1988 and 1995 using MediCal data in 10 counties. Probability matching software was employed to link AFDC histories for these children with birth records, statewide foster care data, and child maltreatment reporting data. Results revealed substantial overlap between the welfare and child welfare populations with approximately 27 percent of all 1990 child AFDC entrants having child welfare contact within 5 years and 3 percent entering foster care. Both total time on aid as well as the number of spells on aid were associated with child welfare contact. Children who transitioned to the child welfare system were more likely to have certain family and case characteristics including: single parent family, larger family size, low birthweight and late or no prenatal care (Needell, Cuccaro-Alamin, Brookhart, & Lee, 1999).

A similar analysis was undertaken in Illinois at The University of Chicago’s Chapin Hall Center for Children. Using linked longitudinal data from the state Department of Children and Family Services and the Division of Financial Support Services, Shook (1998) set out to identify baseline rates of maltreatment among children in Illinois AFDC program between 1990 and 1995. She also identified risk factors for child welfare contact among this population. Risks were higher for children on non-parent cases, children from single parent families, and white children. Of particular interest were the findings that transitions were more likely among children with sanctioned family grants, with removals for the neglect categories lack of supervision or risk of harm more likely among sanctioned cases. In addition to helping identify possible implications TANF sanctions, the research highlights the use of administrative data in assessing the impact of welfare reform on child welfare.

Despite the benefits of using linked longitudinal administrative data it does not come without its challenges. Linking is accomplished by matching unique identifying information such as Social Security numbers across data system of interest. In the absence of unique identifiers in the data sources to be linked, probabilistic matching software can be employed to form an "educated guess" as to which records should be linked across systems. Readers should consult Goerge and Joo Lee (1999) for an in-depth review of the benefits and limits of linked administrative data.

In addition to the complex logistics of linking files, new issues are posed by TANF reforms themselves. A model that thoroughly investigates the relationship between parental welfare paths and child welfare involvement requires not only data on the timing of welfare receipt, but also an indication of the reason that aid ceased. Without an explanation of the reason for termination, it is difficult to distinguish between parents who left aid for gainful employment and those who were dropped from the rolls due to sanction and/or failure to comply with regulations. In many cases, this information is lacking. Therefore, researchers may try and link up welfare and child welfare data to parental employment data in an attempt to understand which families are leaving welfare for "positive" reasons. In some states, like California, sanctions and time limits will result in a decrease in only the parental portion of the welfare grant, with the child portion maintained. Identifying and successfully tracking these parents and children may involve painstaking record linkage and incorporate case flow dynamics that are quite complicated.

Despite these hurdles, given the established association between poverty and maltreatment, it is incumbent upon child welfare advocates and policy makers to examine the impact of welfare reform on child welfare services. In particular, whether these changes increase the likelihood of maltreatment has important consequences for both the TANF families as well as society. In addition to the immediate risk of physical harm and even death maltreated children face, consequences include deficits in emotional and physical health, cognitive development, and socialization difficulties (Ammerman, Cassissi, Hersen, & Van Hasselett, 1986; Couch and Milner, 1993). Beyond these immediate economic costs, observed relationships between childhood maltreatment and later criminal activity or abusive behavior also increase future consequences for both children and society (Gray, 1988).

Unfortunately current statewide evaluations will not provide detailed data regarding reform’s impact on child welfare. As we have illustrated, administrative data provides a comprehensive method for assessing the impact of welfare reform on child welfare. We have reviewed a variety of data configurations and designs that can be implemented to assess impact. Ultimately we conclude that those using linked longitudinal administrative data provide the most promise for this endeavor. In addition to answering critical questions regarding the immediate consequences of reform for child welfare, the linking of welfare and child welfare data sets builds a data infrastructure which can be updated annually providing a rich source of data for future studies. Additionally, with this foundation in place linkages to other data sources such as or employment and school data can be pursued to provide a more comprehensive understanding of the long-term consequences of welfare reform for children.

Education

Educational success is related to current and future economic and physical well-being (Barnett, 1998; Card & Krueger, 1998). This educational success is partly a result of parental work and educational histories. Since there is a strong relationship between education of parents and children, welfare programs that are successful in helping recipients to improve their educational skills (Boudett & Friedlander, 1997) can be expected to have an influence on the learning of their children.

Certainly, improved educational performance of children is one hope of welfare reformers. This process may take several forms. For instance, perhaps by hearing about their parents' success at the worksite, children would be inspired to have high achievement on their own daytime turf-the school. Yet, working parents know how difficult it is to be available to help meet the special educational needs of their children. So, the impact on academic success of parental transitions from welfare to work continue to need monitoring. Another possible mechanism for change might be through the educational attainment of parents who receive educational support through welfare. Since TANF does not emphasize additional education-in the way that JOBS did-participation in TANF is unlikely to have a result of mediating children’s higher educational gains via parent’s higher educational attainment.

A very limited set of Pre-TANF research studies indicates that there may not be a simple, sizeable effect of welfare participation on children’s educational attainment. Hill and O'Neill (1994) found that parental AFDC participation has no effect on children's scores on a standardized test of vocabulary, given income. Currie (1995) confirmed that their results hold up even when sibling comparisons are used to account for unobserved maternal background characteristics. Yet, a recent analysis of NLSY data that included access to other mother and child services found a relationship between program participation and children’s learning (Yoshikama, 1999). Although the evidence base for research on educational outcomes and welfare reform primarily comes from surveys, there is good reason to suggest the importance of using administrative records to study this relationship. This will be particularly fruitful as the availability and meaningfulness of educational records improves.

Measures of educational success include data elements that describe the child’s achievement as well as their receipt of services. Many of these data are now in electronic databases in the school districts, but the automation of educational records tends to begin with the high schools and trickle down to the elementary schools. So, elementary school grades are not as likely to be automated as middle school or high school grades. Standardized statewide test scores are now quite routinely required of all students as are periodic achievement test scores during certain sentinel years. The variety and repetition of tests is becoming quite extensive. (As an illustration, Exhibit 3 includes the testing schedule for students in the schools of North Carolina.)

Exhibit 3: Educational Tests Routinely Used In North Carolina

End-of-grade tests (grades 3 - 8) Writing assessment (grades 4, 7, 10) Norm-referenced testing (grades 5, 8; sampled) Open-ended assessment (grades 4, 8) Computer Skills Proficiency (grade 8) Reading and Mathematics competency testing (screen in grade 8; must pass for diploma by grade 12) End-of-course tests in Algebra, Biology, English, and U.S. History

Most, but not all, students take these tests. Exemptions may be given to students in special education, as determined by their IEP teams. Exemptions may also be given to students who are not following a standard course of study, for example those in alternative education or adolescent parenting programs.

Grade retention histories are usually available (or can be inferred from birthdates and grade levels). Educational reform is making grade retention data more valuable. Although widespread adherence to the principals of social promotion have dominated the nation’s public schools for many years, legislation in many states (e.g., California, New York, North Carolina) is now discouraging social promotion. In the future, grade retention may more truly indicate a child’s performance, not just a school’s educational strategy regarding social promotion.

School services data is also obtainable, although the lack of standardization makes it difficult to assess change when students also change schools during the period under study. Schools also have data about student’s disciplinary actions-almost always including suspensions or expulsions but also including a variety of other disciplinary actions that are less severe. School attendance data is also likely to be automated, although comparisons across schools and, especially, unified school districts must done with care because of different ways of administering the statewide definitions of attendance. This is particularly true of suspensions, as some schools use them routinely and some schools use them only after considerable effort to avoid that outcome by mediating the situation.

A major impediment to using educational data to estimate the well being of children is The Family Educational Rights and Privacy Act (FERPA). First enacted in 1974, FERPA gives parents the right to inspect and review their children’s education records, request amendment of the records, and have some control over the disclosure of information from the records. At age 18, this right is transferred to the student. The Act also restricts the release of school records or information from their records that could identify the student. Before releasing such records or information to a party outside the school system, the school must first obtain the consent of the student’s parent. FERPA offers a key exception to the prior consent requirement. Specifically, educators may disclose information without prior consent if the disclosure being made to organizations conducting studies for, or on behalf of , education agencies or institutions, in order to develop tests, administer student aid, or improve instruction (§99.31(a)(6) of the FERPA Regulations). To meet this requirement, the researcher must have an agreement with the educational institution that they are working on their behalf and that the study will create information that will improve instruction. Additionally, FERPA was amended in 1994 to permit nonconsensual disclosures of education records to officials in the State juvenile justice system as and under certain special circumstances. Despite some loosening of restrictions, FERPA remains a significant barrier to data access for researchers.

Perhaps because of the obstacles created by the FERPA, there has been little work matching administrative records to welfare data to assess possible relationships between educational progress and welfare program participation. Orthner and Randolph (1999) examined the impact of parental work and continuity of welfare receipt on the drop out rates of high school students in families in poverty. This work was accomplished by matching individual records of children who were enrolled in the JOBS program in Mecklenburg county with the administrative records from the Mecklenburg Unified School District. Some case record checking was also done of the paper school district records. Using event history analysis, they examined the risk of dropping out of school in light of potential effects on subsequent social and economic well-being. The data indicate consistency in parental employment (i.e. parents who worked in all quarters) and transitions off welfare are associated with lower rates of dropping out of high school. Longer spells on public assistance are associated with higher drop out rates.

Some community colleges have merged their data with the public welfare data to better understand the overlap between their student population and the welfare population (Community College Involvement in Welfare; www.aacc.nche.edu/research/welfare.htm/11/6/99). In a survey of 1124 community colleges conducted by the American Association of community Colleges, about 32% track students on public assistance. Among the primary reasons given for not tracking students was student confidentiality and privacy issues. Yet, the disruption in the educational careers of welfare recipients that may occur with the end of the JOBS program and the institution of TANF could be a very important outcome for young people (Furstenburg, 199x). A straightforward way to study this overlap and the changes in this population is to merge administrative data from community colleges and TANF programs.

The use of administrative data to understand the variety of paths by which welfare reform may influence educational attainment is fraught with challenges, from FERPA, to unautomated data, to unreliable measures. Yet, educational attainment represents one of the few unambiguous outcomes that can be assessed with administrative records. Substantial legislative efforts need to be taken to make this critical source of information about child well-being more available to researchers. FERPA should be amended to clearly indicate that parental signatures are not required for the routine use of administrative data in research. This is now left open to discretion in every school and school district. We have worked with educational agencies that have provided us with administrative data including children’s names, but have, then, cited FERPA to withhold the provision of social security numbers that would have eased, and probably improved, the matching of these data to those from other service systems. Interpretations of FERPA are, then, idiosyncratic at best. Even if we could increase the consistent application of FERPA it is likely that it would be increased by reducing the already small proportion of times when educational institutions share information about students.

More must be done to allow educational data to be used by researchers to better understand the educational implications of other social programs. This would certainly be a low cost way to try to assist our highest risk children-who are most likely to be involved in multiple systems of care. FERPA has been revised in order to make it possible for schools to share information with correctional agencies. Under FERPA, schools may disclose information from “law enforcement unit records (see §99.3 and §99.8 of FERPA regulations) without the consent of the parents or eligible students. This enables schools to give information to social services or juvenile justice agencies as long as the school district first creates and maintains a "law enforcement unit" that is officially authorized to (1) enforce federal, state, or local law, or (2) maintain the physical security and safety of schools in the district. Although this is a modest amendment of FERPA that may not have direct relevance to most researchers endeavoring to use administrative educational data (perhaps unless they are also studying law enforcement issues), it does suggest Congress’ willingness to modify FERPA for good cause. The "Solomon Amendment" of 1999 also limited the unintended implications of FERPA in order to deny aid to schools that either prohibit or prevent the Secretary of Defense from obtaining, for military recruiting purposes, access to directory information on students. The needs of researchers and policy makers to have good information about the educational outcomes of welfare reform (and other social programs) are also worthy of a FERPA amendment.

Juvenile Justice

There are several reasons why changes in a parent’s involvement with welfare could affect the likelihood of juvenile justice involvement by their children. Communities with high welfare participation also have high crime rates. In one study, these two factors were temporally linked, showing that mortality (particular intentional mortality) was far higher during the first week of the month, when welfare and other public assistance checks arrive. Some youth may be involved in such crimes. Households with parents who move off of welfare into self-sufficiency will have additional resources that they could use to purchase a variety of services and activities that would occupy their children and help them avoid the hazards of "hanging out." At the same time, with parents working away from the home there may be less supervision for those youth who do not become involved in other activities. Also, if families have their benefits cut, we know that they often rely on "other family members" to assist them. Although this typically means other adult relatives, it is possible that youth would feel pressured to bring new resources into the household or to, at least, find resources that would allow them to be less dependent on their families for food, clothing, and entertainment.

The most common approach to assessing criminal justice involvement is to study "arrest records." This is the device used in most studies of the transition from child welfare programs to juvenile justice involvement (e.g., Widom, 1996; English & Widom, 1999). The potential drawback of arrest records is that they reflect the combined behaviors of juveniles and criminal justice systems. This is counterbalanced by the fact that they are generally considered to be more useful than conviction records because convictions or incarcerations because these are determined by so many other factors-especially for less violent crimes. Still, convictions or incarcerations can be used if the theoretical relationship between welfare participation and crime suggests that there would be higher rates of major crimes. Incarcerations in state training programs have been shown to be sensitive enough to pick up differences between groups that did and did not obtain ongoing child welfare services following a child abuse investigation (Jonson-Reid & Barth, in press).

Juvenile justice data can also be obtained from a variety of settings, depending on the geographic locus of the study. At the local level, youth are often remanded to juvenile detention and county camps and ranches. At the state level, they may attend a training school or youth authority program. In larger counties, they generally have more capacity to hold more youth who commit more serious offenses at the local level-whereas more rural areas may use the statewide facilities to a greater extent. Statewide facilities often have their own databases which include substantial additional information that they collect about the child at intake. This makes such information particularly useful for trying to explain exit patterns and the path of services once in the training program.

Some juveniles are tried as adults and others may have their records sealed for a variety of offenses. Still, these remain the exception and unlikely to bias study results or affect interjurisdictional comparisons as long as reasonable sample sizes are maintained.

Although these authors were unable to identify any studies that have directly tested the relationship between parents’ welfare participation and children’s juvenile justice involvement, one important study matched juvenile justice data with survey data from the Moving to Opportunity (MTO) experiment. In the MTO, a total of 614 families living in high-poverty Baltimore neighborhoods were assigned into three different "treatment groups": experimental group families receive housing subsidies, counseling and search assistance to move to private-market housing in low-poverty census tracts (poverty rates under 10 percent); section 8-only group families receive private-market housing subsidies with no constraints on relocation choices; and a control group receives no special assistance under MTO. The impact of this "treatment" was then assessed on juvenile arrests (Ludwig, Duncan, and Hirschfield, 1999). (The authors also tested models that addressed the delinquency of males in the sample using convictions instead of arrests and found quite similar results.)

The Maryland Department of Juvenile Justice provided juvenile offender records containing arrest dates and charges, as well as the disposition of each arrest. The charges associated with these arrests range from shoplifting to attempted murder. Classification of crimes included assaults, property crimes, and a residual crime category (includes drug offenses, truancy, runaway and "ungovernable" behavior).

The study also cast light on the utility of a variety of indicators of juvenile justice involvement. Given the concern about false arrests-especially when there different housing patterns result in movement to communities that might have different threshholds for arresting youth-the authors tested the possibility that false arrests might distort their findings. Drawing on previous ethnographic work (e.g., Ogletree et al., 1995) suggesting that false arrests are likely to be crime-specific, and disproportionately involves charges such as disorderly conduct, resisting arrest, and assaulting a police officer, they focused on crime groups that exclude these charges. Second, they replicated their analysis using convictions instead of arrests. There should be less variation across neighborhoods in false convictions than arrests because juvenile prosecutions are handled at the county level, while arrests are made by local police. It also seems likely that the probability of detecting discriminatory behavior (and hence the expected cost of this behavior) is higher for prosecutors than for police.

Conclusions

Using administrative data for evaluating welfare reform presents challenges and opportunities within each of the domains of child well being. Child abuse and neglect data are generally quite available to the evaluation of welfare reform, because both child welfare and welfare/TANF data sets typically reside within the same governmental department at the local and state level. However, developing appropriate measures of child well being from administrative child welfare data is a challenge and requires programming of longitudinal data files, an understanding of the differences between the child’s experience and the system performance indicators; expertise with a range of sophisticated analysis methods; and understanding of many interpretations that administrative data might allow. In contrast, health measures of child well being--for example, birthweight or immunization completeness--are more uniformly defined and there is more agreement about their implications. However, these data are less available to study welfare reform because they typically reside within government entities separate from departments where welfare data reside. When these data sources differ, issues of compatibility of data formats and definitions, linking of data, confidentiality, and ownership of data files call for collaborative efforts to evaluate welfare reform. Evaluation of impacts within juvenile justice and education include particularly acute challenges of data availability, as well as the need to create valid and reliable measures.

The authors of this chapter have endeavored to increase reader’s familiarity with needed indicators of child well being and the administrative data sets that contain them. A secondary goal has been to alert readers to the ways that existing policies hamper getting access to data necessary to make informed decisions. Obtaining permission to use administrative data for evaluation purposes is harder than it needs to be. Without substantial convergence around the purposes of using administrative data, this emerging technology is going to be a partial, piecemeal, and ephemeral aid to government. The technical solutions for linking are increasing (storage is more affordable, processing times are shorter and matching software is better) but public support has not been built to encourage this linking. Issues of data access and confidentiality present the greatest barriers to full utilization of this resource.

Although the federal government is demanding more accountability from the states, and the states from the counties, there is little outcry from public officials to permit the broader use of administrative data to generate the information required to track the performance of human service agencies. The Scandinavian countries are generating invaluable research using linked data across generations that has been used, for example, to understand the transmission of schizophrenia across generations and the likelihood that children born with birth defects will give birth to children with birth defects. Similarly, program participation data has been combined with information from driving records, educational attainment, military service, and marriage certificates to understand lifetime outcomes of family re-composition and participation in service programs. Researcher access to administrative data is beneficial, and more open access will permit individuals to educate themselves about what is contained in such databases, use the information within those databases to conduct research for multiple purposes, and to reassure the public about the feasibility of using already gathered information for the public good. Concerns over confidentiality continue to present a major barrier to linking administrative data to evaluate the effects of welfare reform on child well being. Perhaps nowhere is there as much sensitivity concerning privacy and confidentiality as with records containing information about vulnerable children and parents who have been accused of violating social norms by abusing or neglecting their children. At the same time, electronic availability of information on individuals permits sophisticated research that were simply impossible in the past. How do we reconcile the need to provide privacy and confidentiality to individual patients while enabling public health researchers and policy makers to use available information to make the best decisions? While privacy and confidentiality of records about children’s well being are important, we suggest that there are already adequate protections, incentives and disincentives, policies and procedures, to preserve individual privacy. We already trust millions of individuals in our society to respect the confidentiality of information they encounter each day in the human services-child welfare, health care, law enforcement, juvenile justice, and education, to name a few. We trust the individuals conducting research within each of these systems to maintain the confidentiality of records. Most of these data are collected without any explicit discussion of whether or how they will be used for research that might inform administration of the program. Yet, we have generated the expectation that individuals not working for those institutions who obtain data from them in order to advance services research through data linking represent a risk to the confidentiality concerns of service recipients. This expectation that there is likely to be even a minimal risk of the mishandling of data lacks an evidentiary base. In our ten years of experiencw using administrative data of the most sensitive kinds (including child abuse reports and juvenile justice records) we know of no violations of individual rights of persons in those data sets. Nor do we have any stories to tell about exceptional procedures we instituted to prevent such misuse. The handling of that information was simply very routine. Perhaps we need a more systematic effort to determine what real and imagined threats to confidentiality exist in data linking efforts. Until we have evidence to the contrary, we should continue to maintain databases with adequate identifying information to support future research projects, and should advocate for change in unwisely broad legislative or regulatory language that adverse effects inter-organizational research. We believe it is appropriate and indeed necessary to maintain personal identifying information on public health and child well-being databases, and that those identifiers should be available to facilitate linkage of electronic health databases to support research to improve the health of our population as well as enhance the health of individuals. At the same time, we emphasize that availability of such identifiers is quite different from license to invade the privacy of individuals or disregard the need for strict confidentiality of the information held within medical records. We believe that it is possible to reconcile all these goals.

We need to encourage constant conversation between investigators specializing in administrative data and those designing surveys so that the surveys can be used to help inform the interpretation of the administrative data. Survey researchers are generally not familiar with the needs of administrative researchers to be provided with data that has adequate variables for matching. For example, data that tell us about the reasons why clients change service use patterns can be combined with information from administrative data about how often and when these service use patterns change. Further, we must develop better strategies for making survey data available for linking with administrative data. A serious threat to this possibility is the assumption that if it is possible for the confidentiality of a dataset to be compromised, it will. This leads to such counterproductive strategies as making it impossible to accurately match samples to their communities or counties of origin (thus obviating the possibility of exploring neighborhood or county effects).

Whereas linked administrative data can provide important information on the impact of welfare reform on child well being, it is not a panacea and will not provide us with all the information we need to monitor welfare reform We must be wary of the conclusions we draw from linked data because we often cannot determine whether an individual did not experience the outcome, was recorded as experiencing the outcome but could not be matched across data systems (e.g., if they moved across jurisdictional lines), or experienced the outcome but was not recorded as such. Even when the data are accurate, they at best help us monitor who appears to affected by welfare reform, when those impacts occur, and where the impact is greatest or least. Sometimes we do not even know the direction of that change—for example, if more children per capita are reported for abuse and neglect under TANF than were reported under JOBS, this could mean that the smaller TANF case loads have resulted in more opportunities for home visiting and better early identification of child abuse and neglect. As to why welfare reform affects children and families differentially, administrative data can only guide us as to the best places to looks for those answers. Carefully designed representative samples can be drawn and subjected to other, more qualitative methods (e.g., surveys) that can build on the framework that a comprehensive administrative data analysis provides.

References

Aber, J. L., Brooks-Gunn, J., & Maynard, R. (1995). The effects of welfare reform on teenage parents and their children. The Future of Children, 5(2), 53-71.

American Humane Association. (1984). Highlights of official child neglect and abuse reporting: 1982 . Denver, CO: American Humane Association.

Ammerman, R. J., Cassissi, J. E., Hersen, M. , & Van Hasselt, V. R. (1986). Consequences of physical abuse and neglect in children. Clinical Psychology Review, 6, 291-310.

Andrulis, D. P. (1998). Access to care is the centerpiece in the elimination of socioeconomic disparities in health. Annuals of Internal Medicine, 129(5), 412-416.

Barnett, W. S. (1998). Long-term cognitive and academic effects of early childhood education on children in poverty. Preventive Medicine, 27, 204-207.

Berrick, J. D., Needell, B., & Minkler, M. (1999). The policy implications of welfare reform for older caregivers, kinship care, and family configuration. Children and Youth Services Review, 21(9-10), 843-864.

Bilaver, L. A., Jaudes, P. K., Koepke, D., & Goerge, R. M. (1999). The health of children in foster care. Social Service Review, 73, 401-417.

Boudett, K. P., & Friedlander, D. (1997). Does mandatory basic education improve achievement test scores of AFDC recipients? A reanalysis of data from California's GAIN program. Evaluation Review, 21, 568-588.

Brady, et al. (1999) CITE CONFERENCE PAPER TITLEAFTER WE GET FINAL LIST

Brauner, S., & Loprest, P. (1999). Where are they now? What states’ studies of people who left welfare tell us. New Federalism: Issues and options for states. Series A. No. A-32. Washington, DC: The Urban Institute.

Card, D., & Krueger, A. B. (1998). School resources and student outcomes. Annals of The American Academy Of Political And Social Science, 559, 39-53.

Children’s Defense Fund. (1998). Welfare to What? Early findings on family hardship and well-being. Washington, DC: Author.

Cohen, M. A. (1998). The monetary value of saving a high-risk youth. Journal of Quantitative Criminology, 14(1), 5-33.

Collins, A., & Aber, J. L. (1997a). Children and welfare reform, Issue brief I: How welfare reform can help or hurt children . New York: National Center for Children in Poverty.

Collins, A., & Aber, J. L. (1997b). Working paper: State welfare waiver evaluations: Will they increase our understanding of the impact of welfare reform on children? . New York: National Center for Children in Poverty.

Couch, J. L., & Milner, J. S. (1993). Effects of child neglect on children. Criminal Justice and Behavior, 20(1), 49-65.

Coulton, C. J., Korbin, J. E., Su, M., & Chow, J. (1995). Community level factors and child maltreatment rates. Child Development, 66, 1262-1276.

Culp, R. E., Culp, A. M., Soulis, J., & Lets, D., (), . (1989). Self esteem and depression in abusive, neglecting, and non-maltreating mothers. Infant Mental Health Journal, 10(4), 243-251.

Currie, J., & Duncan, T. Could Subsequent School Quality Affect the Long Term Gains from Head Start . Cambridge MA: National Bureau of Economic Research.

Darnell, J., & Rosenbaum, S. (1997). Welfare reform: Unanticipated but inevitable consequences for health insurance coverage for the poor. Nutrition, 13 (5), 490-491.

Dean, J. M., & Olson, L. (1999). Protecting Confidentiality of Linked Datasets:

Don't Throw the Baby Out with the Bathwater: Workshop on Confidentiality of and Access to Data Research Files . Washington, DC: National Academy of Science.

Department of Education. (1996). Federal Register, [online]. Department of Education. Available: [wais.access.gpo.gov].

Donahue, J. J., & Siegelman, P. (1998). Allocating resources among prisons and social programs in the battle against crime. Journal of Legal Studies, 27, 1-43.

Doolittle, D. (1998). Welfare reform: loss of supplemental security income (SSI) for children with disabilities. Journal of the Society of Pediatric Nurses, 3(1), 33-44.

Families USA. (1999). One step forward, one step back, [online]. Families USA. Available: GOTOBUTTON BM_1_ http://www.familiesusa.org.

Farel, A.M. (1997). Children with special health care needs. Chapter 11 in Jonathan B. Kotch (ed.). Maternal and Child Health: Programs, problems, and policy in public health. Gaithersburg, MD: Aspen Publishers, Inc.

Farel, A. M., Locklin, E. I., & Peoples-Sheps, M. D. (1999). Sharing information and experiences about maternal and child health data sets through the world wide web. Journal of Public Health Management Practice, 5(3), 93-100.

Furstenburg, (199x). Reference is AWOL, but will be located!

Gabor, V., Ben-Avi, A., & Hill, I. (1997). Developing performance measures and performance-based contracts for state and local MCH activities . Washington, DC: Health Systems Research, Inc: Under contract to Health Resources and Services Administration, Maternal and Child Health Bureau.

Garfinkel, I., & McLanahan, S. S. (1986). Single mothers and their children: Summary and Recommendations. In I. Garfinkel & S. S. McLanahan (Eds.), Single Mothers and Their Children (pp. 165-188). Washington, D.C: The Urban Institute.

Geen, R., & Boots, S. W. (1999). Family Care or Foster Care? How State Policies Affect Kinship Caregivers . Washington, D.C.: The Urban Institute.

Geen, R., & Waters, S. (1997). The Impact of Welfare Reform on Child Welfare Financing . Washington, D.C.: The Urban Institute.

Gelles, R. J. (1992). Poverty and violence toward children. American Behavioral Scientist, 35((3), 258-274.

Geltman, P. L., Meyers, A. F., Greenberg, J., & Zuckerman, B. (1996). Welfare reform and children's health. Archives of pediatric and adolescent medicine, 150, 384-389.

Gil, D. G. (1971). Violence against children. Journal of Marriage and the Family, 33(4), 637-648.

Giovannoni, J. M., & Billingsley, A. (1970). Child neglect among the poor: A study of parental adequacy in families of three ethnic groups. Child Welfare, XLIX(4), 196-204.

Goerge, R. M., & Joo Lee, B. (1999). CITE CONFERENCE PAPER TITLE. I DO NOT HAVE FINAL LIST

Gortmaker, S. L., & Walker, D. K. (1984). Monitoring child health in communities. In D. K.. Walker. J. B. Richmond (Ed.), Monitoring child health in the United States . Cambridge, MA: Harvard University Press.

Grason, H., & Nachbar, N. E. (1997). Developing MCH systems measures: A process and strategy in Kansas . Washington, DC: Health Systems Research, Inc. Under contract with the Health Research and Services Administration, Maternal and Child Health Bureau.

Gray, E. (1988). The link between child abuse and juvenile delinquency: What we know and recommendations for policy and research. In G. T. Hotaling, D. Finkelhor, J. T. Kirpatrick, & M. A. Straus (Eds.), Family Abuse and its Consequences: New Directions in Research (pp. 109-123). Newbury Park. CA: Sage.

Haskins, R., In Eds). (pp. ). : . (1995). Losing ground or moving ahead? Welfare reform and children. In P. Chase-Lansdale & J. Brooks-Gunn (Eds.), Escape from Poverty: What Makes a Difference for Children (Vol. Escape from Poverty: What Makes a Difference for Children, pp. 63-86). New York: Cambridge University Press.

Institute of Medicine. (1988). The Future of Public Health . Washington, DC: National Academy Press.

King, A. G. (1997). Welfare reform: Unanticipated but inevitable consequences for health insurance coverage for the poor. Nutrition, 13(5), 490-491.

Knitzer, J., & Bernard, S. (1997). Children and welfare reform, Issue brief I: The new welfare law and vulnerable families: Implications for child welfare and child protection systems . New York: National Center for Children in Poverty.

Koppelman, J., & Lear, J. G. (1998). The new child health insurance expansions: How will school-based health centers fit in? Journal of School Health, 68(10), 441-446.

Ludwig, J., Duncan, G. J., & Hirschfield, P. (1999). Urban poverty and juvenile crime: Evidence from a randomized housing-mobility experiment. American Journal Of Epidemiology, 49(11), 203-218.

Magura, S. & Moses, B.S. (1986). Outcome Measures for Child Welfare Services: Theory and Applications. Washington, D.C.: Child Welfare League of America, Inc.

Margolis, L.H., Cole, G.P., & Kotch J.B. (1997). Historical Foundations in Maternal and Childn Health. Chapter 2 in Jonathan B. Kotch (ed.) Maternal and Child Health: Programs, problems and policy in public health. Gaithersburg, MD: Aspen Publishers, Inc.

Meezan, W., & Giovannoni, J. (1995). The current threat to protective services and the child welfare system. Children and Youth Services Review, 17(4), 567-574.

Meriwether, R.A. (1996). Blueprint for a national public health surveillance system for the 21st century. Journal of Public Health Management Practice 2, 16-23.

Moffitt, R.A., & Slade, E.P. (1997) Health care coverage for children who are on and off welfare. Future of Children, 7 (1), 87-98.

Moore, K. A., & Driscoll, A. K. (1997). Low wage maternal employment and outcomes for children: A study. The Future of Children, 7(1), 122-127.

National Center for Health Statistics, (1996). Healthy People 200 Review, 1995-96. Hyattsville, MD: Public Health Service.

National Conference of State Legislators. (1999). Tracking welfare reform: Designing follow-up studies of recipients who leave welfare, [online]. National Conference of State Legislators. Available: http://www.ncsl.org/statefed/welfare/trackbrf.html.

Needell, B., Cuccaro-Alamin, S., Brookhart , A., & Lee, S. (1999). Transitions from AFDC to child welfare in California. Children and Youth Services Review, 21(9/10), 815-841.

Needell, B., Webster, D., Barth, R. P., Armijo, M., & Fox, A. (1997). Performance Indicators for child welfare in California: 1996 . Berkeley, CA: Child Welfare Research Center.

North Carolina State Center for Health Statistics. Personal communication, 1997.

O’Campo, P., & Rojas-Smith, L. (1998). Welfare reform and women’s health: Review of the literature and implications for state policy. Journal of Public Health Policy, 19, (4), 420-446.

Orthner, D. K., & Randolph, K. A. (1999). Welfare reform and high school dropout patterns for children. Children And Youth Services Review, 21, 881-900.

Pappas, G. (1998). Monitoring the health consequences of welfare reform. International Journal of Health Services, 23(4), 703-713.

Perrin, J. M. (1997). The implications of welfare reform for developmental and behavioral pediatrics. Journal of Developmental and Behavioral Pediatrics, 18(4), 264-266.

Philips, D. P., Christenfeld, N., & Ryan, N. M. (1999). An increase in the number of deaths in the United States in the first week of the month: An association with substance abuse and other causes of death. New England Journal of Medicine, 341(2), 93-98.

Reiss, J. (1999). Does your state's Title XXI SCHIP plan promote the development and maintenance of quality systems of care for children with special health needs, issues and criteria for SCHIP plan review and analysis. Available: GOTOBUTTON BM_2_ http://www.ichp.edu/schip/materials/893253826.html.

Research Forum on Children, Families, and the New Federalism (July, 1999). Evaluating the effects of state welfare policies on children: The project on state level outcomes. The Forum, 2(2), 4-5.

Schorr, L.B. (1997). Common Purpose: Strengthening families and neighborhoods to rebuild America. New York: Doubleday. Pg. 181.

Shenkman, E. (1999). Analysis of pending accounts in the Florida Healthy Kids Program: Executive summary prepared for the Healthy Kids Corporation. Available: http://www.ichp.edu/rmanaged/materials/896969477.html.

Shook, K. (1998). Assessing the consequences of welfare reform for child welfare. Poverty Research News, 11(1), 8-12.

Sondik, E. (1996). Healthy People 2000: Meshing national and local health objectives. Public Health Reports, 111, 518-526.

Willis, E., & Kleigman, R.M. (1997) Wisconsin’s welfare reform and its potential effects on the health of children. Journal of Health Care for the Poor and Underserved, 8 (1), 25-35.

Wilson, J. B., Ellwood, D. T., & Brooks-Gunn, J. (1995). Welfare-to-work through the eyes of children. In P. Chase-Lansdale & J. Brooks-Gunn (Eds.), Escape from Poverty: What Makes a Difference for Children (pp. 63-86). New York: Cambridge University Press.

Wolock, I., & Magura, S. (1996). Parental substance abuse as a predictor of child maltreatment re-reports. Child Abuse & Neglect, 20(12), 1183-1193.

Wulczyn, F. H., Brunner, K., & Goerge, R. M. (1999). A report from the multistate foster care data archive. Foster care dynamics, 1983-1997 . Chicago: Chapin Hall Center for Children at the University of Chicago.

Yaffe, A. (1998). Examining the effects of welfare reform on children: The project on state level outcomes. Poverty Research News, 2(1).

Yoshikawa, H. (1999). Welfare dynamics, support services, mothers' earnings, and child cognitive development: Implications for contemporary welfare reform. Child Development, 70(3), 779-801.

Zaslow, M. J., Moore, K. A., Morrison, D. R., & Coiro, M. J. (1995). The family support act and children: Potential pathways of influence. Children and Youth Services Review, 17(1,2), 231-249.

Zaslow, M., Tout, K. Botsko, C. & Moore, K. (1998). Welfare reform and children: Potential implications. New Federalism: Issues and Options for States, Series A, No A-23. Washington, D.C.: The Urban Institute.

Zeller, D. E. (1999). Welfare Reform Impacts on Child Wlefare Caseloads: A Research Agenda. Troy, NY: Hornby Zeller Associates, Inc.

Zuravin, S. J., & DiBlasio, F. A. (1996). The correlates of child physical abuse and neglect by adolescent mothers. Journal of Family Violence, 11(2), 149-166.