Preface

ANALYSIS OF STATE CHILD WELFARE DATA:
VCIS SURVEY DATA FROM 1990 THROUGH 1994

Preface

INTRODUCTION

This analysis of state child welfare data presents information on child substitute care and adoption that was gathered through the Voluntary Cooperative Information System (VCIS) for the years 1990 to 1994. Public child welfare agencies in the fifty states, the District of Columbia, and Puerto Rico were asked to provide aggregate state-level statistics to VCIS.

The on-line report is organized in five parts:

This on-line report is consistent with the entire original report which can be found here in PDF format. Additionally, the original report also contains graphs of national estimate data, which depict trends in substitute care and adoption populations.

This report presents state-specific data and summary statistics for the years 1990 to 1994.

1. BRIEF HISTORY OF THE VOLUNTARY COOPERATIVE INFORMATION SYSTEM

The American Public Human Services Association, formerly the American Public Welfare Association (APWA), operates VCIS with funding assistance from the Administration for Children and Families (ACF), U.S. Department of Health and Human Services (HHS). VCIS has been operational since 1982. The federal government had collected voluntary annual data on substitute care and adoption from the late 1940's until 1975. However, between 1975 and the early 1980's, virtually no state-specific data were reported. VCIS was developed as an APWA initiative as a way to fill the continuing need for national information on child welfare programs. APWA began publishing summaries of VCIS data in 1982. This report is concerned with the years 1990 to 1994.

In 1986, Congress amended title IV-E of the Social Security Act by adding section 479, which sets forth directives for establishing and implementing a mandatory substitute care and adoption data collection system. The Adoption and Foster Care Analysis and Reporting System (AFCARS) was implemented by a Federal Register announcement on December 22, 1993. AFCARS is designed to provide uniform, reliable information on children in the child welfare system. Each state is asked to provide 65 data elements for each child in foster care and 37 data elements for each child who has been adopted during a six-month reporting period. The data elements collected through AFCARS correspond closely to the data elements collected through VCIS.

The first AFCARS reporting period was from October 1, 1994, through March 31, 1995. As with any new, complex reporting system, the initial reporting was imperfect and incomplete and therefore limits the usefulness of the AFCARS data. The data collected through VCIS during the implementation phase of AFCARS can provide additional information on children in substitute care that is useful for policy development and program management issues. Some of the exhibits in this report will provide a comparison of 1994 VCIS data to the first AFCARS data collected in fiscal year 1995.

2. METHODOLOGY AND LIMITATIONS OF THE DATA

The substitute care and adoption data collected through VCIS provided the best available national statistics on these programs before the implementation of AFCARS. However, the VCIS data have some significant limitations that must be considered when interpreting the exhibits included in this report. These limitations, and the methodology used to deal with them, are discussed in detail in the sections that follow.

APWA submits the VCIS survey instrument to the primary state agencies administering public child welfare programs. The instrument is divided into two major sections:

Part I, Children in Substitute Care
Part II, Adoption Information.

For each section, states are asked to provide both point-in-time and cumulative information. For each question, states use their own programmatic definitions unless otherwise requested. All data reported by states are assumed to be correct. Anomalies in the responses or questionable responses by a state remain in the database.

2.1 States Use Different Reporting Periods

Reporting agencies are given three options for defining a reporting period:

Federal fiscal year (October 1 through September 30)
State fiscal year (typically July 1 through June 30)
Other twelve-month period (such as a calendar year).

For the 1994 VCIS survey, 7 states used a federal fiscal year, 9 states used a state fiscal year, 5 states used some other twelve-month period, and 31 jurisdictions did not indicate a reporting period. In addition, over the five years (1990 to 1994), 16 states identified their type of reporting period for all five years. Fifteen (94 percent) of these 16 states used the same type of reporting period over the five years.

VCIS data are therefore not completely comparable since states use different reporting periods. Further, the data collected from states that changed reporting periods may include the double counting or omission of some children for part of the five-year reporting period. Since the VCIS database does not contain sufficient information to correct for these problems, the data has not been altered.

2.2 States Use of Different Reporting Populations

Although many states use comparable reporting population definitions, some variations exist across the states. Two separate reporting populations are requested by VCIS:

Children residing in substitute care (Part I)
Children receiving adoption services (Part II).

The VCIS survey instrument specifies definitional criteria for each population.

Substitute Care

Substitute care is defined as a living arrangement where a child resides outside his/her own home, under the case management and planning responsibility of the primary state child welfare agency (or child placement agencies under contract to the state agency). These living arrangements are to include relative and non-relative foster homes, group homes, childcare facilities, emergency shelter care, supervised independent living, and non-finalized adoptive homes. State agencies are requested to count all children in substitute care regardless of the number of hours they remain in care, and to exclude children who have been reunified/placed with parents, relatives, or other caretakers, but who continue to receive post-placement supervision.

Only states that use the same population definitions can be considered comparable. Exhibits A-1 and A-2 in the Appendix illustrate the extent to which states use the specified population definitions and indicate the states' consistency in definitions across the five years. States do use different substitute care criteria and occasionally change definitions over time. Comparability across states is also limited by some states providing data on population totals without responding to the population definition questions. As with differing reporting periods, the VCIS database does not contain adequate information to correct for these inconsistencies.

Adoption

Part II of the VCIS instrument requests information on several adoption populations. Again, the survey specifies definitional guidelines for each section:

Finalized adoptions, defined as an adoptive home placement that has been approved by the courts as final (excluding non-finalized adoptive home placements)
Children legally free for adoption, with a permanency plan of adoption, but still residing in substitute care (excluding non-finalized adoptive home placements)
Non-finalized adoptive home placements, defined as an adoptive home placement in which the child's adoption is legally pending the results of a trial placement period
Children authorized to receive an adoption subsidy, defined as any federal, state, or local financial assistance that is authorized to support a child's adoption.

As with substitute care population definitions, states do deviate from the definitional categories suggested by the VCIS survey. Exhibits A-3 and A-4 in the Appendix indicate the comparability across states for Part II of the survey. Again, no attempts have been made to correct for the non-comparability.

2.3 Individual Question Limitations

In addition to the differences among reporting periods and definitions identified above, the VCIS database is incomplete since some states are not able to provide information for every data item. Generally, substitute care data are more complete than adoption data.

In interpreting responses to individual questions, it is important to consider that because states are using their own programmatic definitions, data may not be completely comparable across the states. Further, responses to similar questions in Part I and Part II of the survey are definitely not comparable. Substitute care and adoption data are often collected by different state agencies, using differing population definitions and different databases.

An analysis of the internal consistency of states' responses to individual questions has been conducted in two stages. The utility of individual questions was examined and flow data edits were conducted.

Internal Consistency

The examination of the utility of individual questions employed three steps. First, the consistency of totals edit was performed by comparing a state's total reported population to its numbers in responses to individual questions. All states passed this test, indicating that the data is reliable.

Second, an internal consistency edit qualitatively reviewed the reported sub-categories. The responses to each question were reviewed to determine whether the data available were sufficient to report that data reliably. Only those states responding within VCIS categories have been included in the analyses of a question. Omitting inconsistent states increases the comparability across states.

Third, the unknown data edit identified states responding to individual questions with more than 10 percent of the children categorized as "unknown or unreported." These states have been omitted from the analyses to improve the comparability across states. Exhibit A-5 in the Appendix shows which states passed each of these three criteria.

Flow Data Edits

During the second stage, three separate flow data edits were employed. The first edit required that the number of new entrants to care be less than or equal to the total number of children entering care during the period. The second flow edit required that the number in care at the end of the period be equal to the number in care at the beginning of the period plus those entering care and minus those leaving care during the period. States were considered to have usable flow data if their flow data variance was less than 5 percent. A third edit of year-to-year reporting consistency was also conducted. It aimed to determine if the number of children reported in substitute care on the last day of the prior reporting period corresponded to the number of children in care on the first day of the current reporting period.

2.4 National Estimates

For each substitute care and adoption question in the survey, national estimates for the years 1990 to 1994 have been calculated. To maximize accuracy within the limitations of the data, percentages for substitute care questions have been calculated using four different denominators as appropriate for the particular question, namely, the total reported populations for the four major categories of questions: Children in Care at Start, Children Entered Care, Children Left Care, Children in Care on Last Day. Similarly, percentages for adoption questions have been calculated using three different denominators as appropriate: total reported populations of Children with Adoptions Finalized, Children Free for Adoption, and Children in Non-Finalized Adoptive Home Placements.

Analyses incorporating data from a large number of states or states with large substitute care and adoption populations are likely to be more nationally representative than analyses including fewer and smaller states. In other words, the national estimates for questions including more and larger states are expected to be closer to true national totals than are estimates based on other states.

Five-year trend national estimates (1990 to 1994) have also been included whenever possible. These estimates have been calculated using only those states that responded using VCIS categories with less than 10 percent unknown for all five years. Estimates for all applicable questions have been calculated for all years. This maximizes the information available for analysis. It is important to recognize that the national estimates will differ from the trend national estimates presented in the five-year trend, because the two estimates are based on a different number of states. Exhibit A-5 in the Appendix identifies for each VCIS question the states that are included in the five-year trend national estimates.

The procedure for calculating the national estimates is based on the estimated number of children in care on the last day of the reporting period. Since some states did not provide this information in every year, the national estimates include the data from the APWA End-of-the-Year Survey. A comparison of the VCIS and APWA End-of-the-Year totals appears in Exhibit A-6 of the Appendix.

Based on the total number of children reported at the end of each year, a multiplier was identified. This multiplier represents the total number of children in care at the end of the year divided by the category’s total number of children in care at the end of the year for only the reporting states. This multiplier was then applied to reported states’ total for the category to generate the national estimates.