The CHILD System is composed of linked administrative records of children in Cuyahoga County, Ohio, beginning with the 1989 birth cohort to the present (see the Data Sets link below). The linkage of records across time and systems is performed via deterministic and probabilistic matching techniques. The records contain geographic information that enables aggregation to the neighborhood, city, county or other jurisdiction level and also allows linking with other data systems at various levels of geographies such as parcel, address, or census tracts. The end result is a longitudinal data system in which children are observed if and when they are served by one of the 35 administrative systems that compose CHILD.
Incoming data are geocoded and standardized in preparation for linkage. A third-party SAS macro, LinkPro performs deterministic and probabilistic matching to determine whether the new records match those already in the system. Matching variables include the child’s and mother’s names, birth dates, social security numbers, family address, child’s race and gender. Non-matched records are manually reviewed and either matched or appended to CHILD. Continuous evaluation of these methods guarantees that the linkages are at acceptable levels of reliability and completeness.
Get more more information on the structure of the CHILD System, a technical paper, An IDS Resource for Cuyahoga County: Childhood Integrated Longitudinal Data System, written by Dr. Claudia Coulton, Dr. Robert Fischer, Nina Lalich, Dr. Elizabeth Anthony, Dr. Francisca Richter, and Dr. Meghan Salas Atwell.