Using administrative data.

Abstract

Child welfare agencies have always collected and reported some statistical data to their fenders and licensing agencies. The computerization of the records upon which such statistics are based, however, has had a major impact upon the application of such data in the management of agencies and the daily work of caseworkers. Equally important is the impact that the availability of such records is having upon the field of child welfare research. The nature of automated data-that they are maintained on a regular basis, that they address critical aspects of agency work, that they are client based-leads to using the data not only for management but also for research. This issue of Child Welfare presents a selection of studies by researchers who are using administrative data to better describe and understand social service programs. Administrators and workers who think of administrative databases as simply filing cabinets for data on individual cases will be interested to learn how these data can be of much greater use. Maza discusses the history of using the Adoption and Foster Care Analysis and Reporting System (AFCARS) data to reward states that achieve increased adoptions. Wulczyn et al, discuss the use of administrative data to measure contract agency performance. Moore et al. discuss the use of administrative data in developing outcome reports. These three articles are examples of measuring critical components of performance through such databases. English et al. provide a description of the progress from data collection to data utilization and its impact on improving agency performance. Graef and Hill's analysis of administrative data to determine costs of worker turnover provides an example of the use of such data for agency management. Percy et al.'s study of Northern Ireland describes a new way of allocating funds to small local areas. This set of articles clearly links the use of administrative data to agency culture and agency management. The remaining articles discuss additional ways of analyzing data and speak to administrators and researchers, who may already have some experience in using such data. Ernst combines census data with geographical mapping software to assess the risk of child abuse and neglect by neighborhood. Fluke et al. discuss strategies for overcoming problems caused by missing or noncomparable data in understanding the outcomes of an agency's services. Garnier and Poertner discuss measurement problems in using administrative data. Webster et al. discuss the use of a subset of a larger longitudinal dataset to discover information new to the field. Schoech et al. introduce data mining and give some glimpses of potential uses by agencies as well as funders. The term administrative data has its origins in the information used by agencies to manage and administer their programs and provide client services. For example, a report in the 1960s by the U.S. Children's Bureau discusses the universe from which a sample of data on children receiving child welfare services could be drawn. The report states that most public and voluntary agencies keep a frequent check, usually monthly, on the size of their caseloads for various administrative and reporting reasons [Jeter 1963: 160]. Administrative data, therefore, connotes data that are available because they are regularly collected to conduct the functions of the agency Among early child welfare agencies and indeed up to the present day, much of these data have been maintained in paper records. One hundred years ago, orphanage workers kept logs in large volumes of children who entered their care and their characteristics. Even now there are agencies and programs that keep their essential notes on children in their care on filing cards. In spite of the availability of these paper records, child welfare researchers have traditionally used other sources of data to gain information about the characteristics of children who are served, the nature of the services that they have received, and the results of these services. …